HANDBOOK OF

COMPOSITES
HANDBOOKOF
COMPOSITES

SECOND EDITION

Edited by

S.T. Pe ters
Process Research, Mountain View, California, USA

la ni SPRINGER-SCIENCE+BUSINESS MEDIA, B.V.

First edition 1982

Second edition 1998

© 1998 Springer Science+Business Media Donlrecht

Originally published by Chapman & Hall in 1998
Softcover reprint ofthe hardcover 2nd edition 1998
Thomson Science is a division of International Thomson Publishing

Typeset in 10/12 pt Palatino by GreenGate Publishing Services, Tonbridge, Engiand

ISBN 978-0-412-54020-2 ISBN 978-1-4615-6389-1 (eBook)

DOI 10.1007/978-1-4615-6389-1

All rights reserved. No part of this publication may be reproduced, stored in a

retrieval system or transmitted in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without the prior written permission of the
publishers. Applications for permission should be addressed to the rights manager
at the London address of the publisher.
The publisher makes no representation, express or implied, with regard to the
accuracy of the information contained in this book and cannat accept any legal
responsibility or liability for any errors or omissions that may be made.

A catalogue record for this book is available from the British Library
CONTENTS

Contributors ix

Preface xv

About the editor xvi

Foreword xvii

Acknowledgements xviii

Introduction, composite basics and road map 1

S.T. Peters
1 Overview of composite materials 21
Theodore]. Reinhart

PART ONE: BASIC MATERIALS

Polymeric matrix systems
2 Polyester and vinyl ester resin 34
Frank A. Cassis and Robert C. Talbot
3 Epoxy resins 48
L.S. Penn and H. Wang
4 High temperature resins 75
Hugh H. Gibbs
5 Speciality matrix resins 99
David A. Shimp
6 Thermoplastic resins 115
Lars A. Berglund

Reinforcements and composites

7 Fiberglass reinforcement 131
Dennis J. Vaughan
vi Handbook of composites

8 Boron, high silica, quartz and ceramic fibers 156

Anthony Marzullo
9 Carbon fibers 169
Khalid Lafdi and Maurice A. Wright
10 Organic fibers 202
Linda L. Clements
11 Particulate fillers 242
Harry S. Katz
12 Sandwich construction 254
Andrew C. Marshall
13 Metal matrix composites 291
VJ. Kostikov and V.S. Kilin
14 Ceramic composites 307
M.F. Amateau
15 Carbon-carbon composites 333
John D. Buckley

PART TWO: PROCESSING METHODS

General composites and reinforced plastics
16 Hand lay-up and bag molding 352
D.R. Sidwell
17 Matched metal compression molding of polymer composites 378
Enamul Haque and Burr (Bud) L. Leach
18 Textile preforming 397
Frank K. Ko and Ceorge W. Du
19 Table rolling of composite tubes 425
John T. Kanne and Jerome S. Berg
20 Resin transfer molding 433
Lihwa Fong and S.C. Advani
21 Filament winding 456
Yu.M. Tarnopol'skii, S.T. Peters, A.I. Bei!,
22 Fiber placement 476
Don O. Evans
23 Pultrusion 488
Brian A. Wilson
24 Processing thermoplastic composites 525
James L. Throne
 Contents vii

Advanced composites
25 Tooling for composites 556
Jerry L. Cadden and Paul F. Sadesky
26 Consolidation techniques and cure control 576
Zhong Cai and Timothy Gutowski
27 Composite machining 596
Kent E. Kohkonen and Nitin Potdar
28 Mechanical fastening and adhesive bonding 610
D. W. Oplinger
29 Surface preparations for ensuring that the glue will stick in bonded
composite structures 667
L.J. Hart-Smith, D. Brown and S. Wong

PART THREE: DESIGN AND ANALYSIS

30 Laminate design 686
Jocelyn M. Seng
31 Design of structure with composites 709
F.J. Schwan
32 Analysis methods 736
v.v. Vasiliev
33 Design allowables substantiation 758
Christy Kirchner Lapp
34 Mechanical tests 778
Yu.M. Tamopol'skii and v.L. Kulakov

PART FOUR: ENVIRONMENTAL EFFECTS

35 Durability and damage tolerance of fibrous composite systems 794
Ken Reifsnider
36 Environmental effects on composites 810
Ann F. Whitaker, Miria M. Finckenor, Harry W. Dursch, R.c. Tennyson
and Philip R. Young
37 Safety and health issues 822
Jennifer A. Heth
38 Nondestructive evaluation methods for composites 838
Thomas S. Jones
39 Repair aspects of composite and adhesively bonded aircraft structures 857
Anton L. Seidl
40 Reuse and disposal 883
Harry E. Pebly
viii Handbook of composites

PART FIVE: APPLICATIONS

41 Land transportation applications 905
Douglas L. Denton
42 Marine applications 916
Wayne C. Tucker and Thomas Juska
43 Commercial and industrial applications of composites 931
Stewart N. Loud
44 Composite biomaterials 957
Shalaby W. Shalaby and Robert A. Latour
45 Scientific applications of composites 967
Vicki P. McConnell
46 Construction 982
Ever J. Barbero
47 Aerospace equipment and instrument structure 1004
Gary C. Krumweide and Eddy A. Derby
48 Aircraft applications 1022
Richard N. Hadcock
49 Composites in the sporting goods industry 1044
Brian E. Spencer

APPENDICES
Appendix A
Typical properties for advanced composites 1053
Kenneth R. Berg
Appendix B
Specifications and standards for polymer composites 1059
Frank T. Traceski
Index 1069
CONTRIBUTORS

SURESH G. ADVANI KENNETH R. BERG

Department of Mechanical Engineering, Riggs Corporation,
University of Delaware, 837 Agate Street,
Spencer Laboratory, Medford, OR 97501,
Newark, DE 19716, USA
USA
LARS A. BERGLUND
MAURICE F. AMATEAU Lulea University of Technology,
Applied Research Laboratory, SE-971 87 Lulea,
Pennsylvania State University, Sweden
PO Box 30,
State College, PA 16804, D.BROWN
USA Boeing Commercial Airplane Group,
Douglas Products Division,
EVER J. BARBERO Mail Stop DOOI-0018,
315 Engineering Science Building, Long Beach, CA 90846,
West Vll'ginia University, USA
Morgantown, WV 26506-6106,
USA JOHN D. BUCKLEY
23 East Governor Drive,
A.I. BElL' Newport News, VA 23602,
Institute of Polymer Mechanics, USA
Latvian Academy of Sciences,
23 Aizkraukles Street, JERRY L. CADDEN
Riga LV-I006, C & S Technologies,
Latvia 42759 Mountain Shadow,
Murrieta, CA 92562,
JEROME S. BERG USA
True Temper Sports,
5421 Avenida Encinas, ZHONG CAl (deceased)
Suite G, 4180 Berkeley Creek Drive,
Carlsbad, CA 92008, Duluth, GA 30136,
USA USA
x Handbook of composites

FRANK A. CASSIS MIRIA M. FINCKENOR

FAC Associates, EH12 Bldg 4711,
1150 N. Mountain, Suite 102B, Marshall Space Flight Center,
Upland, CA 91786, AL35812,
USA USA

LINDA L. CLEMENTS LIHWAFONG

C & C Technologies, BLKG5,
PO Box 1089, Nanyang Avenue,
Dayton, NY 89403, Singapore 63616
USA
HUGH H. GIBBS
DOUGLAS L. DENTON Polycomp Consulting, Inc.,
Chrysler Corporation, 25 Crestfield Road,
CIMS 482-00-13, WIlmington, DE 19810,
800 Chrysler Drive, USA
Auburn Hills, MI 48326-2757,
USA TIMOTHY GUTOWSKI
Department of Mechanical Engineering,
EDDY A. DERBY Massachusetts Institute of Technology,
Composite Optics, Bldg 35-234,
9617 Distribution Ave, Cambridge, MA 02139,
San Diego, CA 92121, USA
USA
RICHARD N. HADCOCK
GEORGEW.DU 6 Sue Circle,
Principal Engineer, Huntington, NY 11743,
16331 Bay Vista Drive USA
Clearwater, FL 34620,
USA ENAMUL HAQUE
Azdel, Inc.,
HARRY W. DURSCH Technology Center,
Boeing Defense and Space Group, 658 Washburn Switch Road,
PO Box 3999, Mail Stop 73-09, Shelby, NC 28151-2284,
Seattle, WA 98124-2846, USA
USA
L.J. HART-SMITH
DONO. EVANS Boeing Commercial Airplane Group,
Cincinnati Milacron, Douglas Products Division,
4701 Marburg Avenue, Mail Stop 0800-0019,
Cincinnati, Ohio 45209, 4000 Lakewood Boulevard,
USA Long Beach, CA 90846,
USA
 Contributors xi

JENNIFER HETH VALERY I. KOSTIKOV

Cytec Fiberite, Niigrafit Institute,
501 W. Third Street, 2 Electrodonaya Street,
Winona, MN 55987-2854, Moscow, 111524,
USA Russia

THOMAS S. JONES GARY C. KRUMWEIDE

Industrial Quality, Inc., Composite Optics,
640 E. Diamond Ave., Suite C, 9617 Distribution Avenue,
Gaithersburg, MD 20877, San Diego, CA 92121,
USA USA

THOMAS JUSKA v.1. KULAKOV

Naval Surface Warfare Center, Institute of Polymer Mechanics,
Carderock Division, Latvian Academy of Sciences,
Structures and Composites Department, 23 Aizkraukles Street,
Bethesda, MD 20084-5000, Riga LV-1006,
USA Latvia

JOHN T. KANNE KHAUD LAFDI

2201 Johnson Road, Center for Advanced Friction Studies,
Memphis, TN 38139, Southern lllinois University at Carbondale,
USA Carbondale, IL 62901-4343,
USA
HARRY S. KATZ
Utility Development Corporation, CHRISTY KIRCHNER LAPP
112 Naylon Avenue, 1412 Bellingham Way,
Livingston, NJ 07039, Sunnyvale, CA 94087,
USA USA

V.S. KILIN ROBERT A. LATOUR

Niigrafit Institute, Clemson University,
2 Electrodonaya Street, Clemson, SC 29634,
Moscow, 111524, USA
Russia
BURR 1. LEACH
FRANKK.KO Cambridge Industries,
Drexel University, 1700 Factory Avenue,
Fibrous Materials Research Laboratory 27-439, Marion, IN 46952,
Philadelphia, PA 19104, USA
USA
STEWART N. LOUD
KENT E. KaHKONEN Composites Worldwide Inc.,
Brigham Young University, 991 Lomas Santa Fe Drive, C469,
435 CTB Technology Department, Solana Beach, CA 92075-2125,
Provo, ur 84602, USA
USA
xii Handbook of composites

VICKI P. MCCONNELL NITIN POTDAR

Ray Publishing, Brigham Young University,
Independence Street, 435 CTB Technology Department,
Suite 270, Wheat Ridge, Provo, UT 84602,
CO 80033, USA
USA
KENNETH REIFSNIDER
ANDREW C. MARSHALL Virginia Polytechnic Institute and State
Marshall Consulting, University,
720 Appaloosa Drive, Patton Hall 120,
Walnut Creek, CA 94596, Blacksburg, VA 24061-0219,
USA USA

ANTHONY MARZULLO THEOOORE J. REINHART

39 Harold Street, 345 Forrer Boulevard,
Cos Cob, CT 06807-2132, Dayton, OH 45419-3238,
USA USA

OONALD W. OPLINGER PAUL F. SADESKY

Federal Aviation Administration, C & S Technologies,
Wm. J. Hughes Technical Center AAR-431, 23547 Mountain Court,
Atlantic City, International Airport, Murrieta, CA 92562,
NJ 08405, USA
USA
FRANK J. SCHWAN
HARRY E. PEBLY 36671 Montecito Drive,
198 Center Grove Road, Fremont, CA 94536,
Randolph, NJ 07869, USA
USA
ANTON L. SEIDL
LYNNS.PENN 18941 Mellon Drive,
Department of Chemical and Materials Saratoga, CA 95070,
Engineering, USA
177 Anderson Hall,
University of Kentucky, JOCELYN M. SENG
Lexington, KY 40506-0046, Owens Corning Science and Technology
USA Center,
2790 Columbus Road,
S.T. PETERS Granville, OH 43023-1200,
Process Research, USA
925 Sladky Avenue,
Mountain View, CA 94040-3625, SHALABY W. SHALABY
USA Clemson University,
301 Rhodes Res.,
Clemson, SC 29634,
USA
 Contributors xiii

DAVIDA. SHIMP WAYNE C. TUCKER

PO Box 974, Naval Undersea Warfare Center,
Prospect, KY 40059, PO Box 86,
USA Exeter, RI 02822,
USA
DONALD R. SIDWELL
44609 Grove Lane, V. V. VASILIEV
Lancaster, CA 93534-2833, Moscow State University,
USA 14-1-110 Podolskih Kursantov Street,
Moscow 113545,
BRIAN E. SPENCER Russia
Spencer Composite Corporation,
3220 Superior Street, PO Box 4377, DENNIS J. VAUGHAN
Lincoln, NE 68504-0377, 146 Longview Drive,
USA Anderson, SC 29621,
USA
ROBERT C. TALBOT
7199 Lorine Court, H.WANG
Columbus, OH 4323~5125, Department of Chemical and Materials
USA Engineering,
177 Anderson Hall,
YU.M. TARNOPOL'SKll University of Kentucky,
Institute of Polymer Mechanics, Lexington, KY 40506-0046,
Latvian Academy of Sciences, USA
23 Aizkraukles Street,
Riga LV-1006, ANN F. WHITAKER
Latvia EH01 Bldg 4612,
Marshall Space Flight Center,
R.c. TENNYSON AL35812,
University of Toronto, USA
Institute for Aerospace Studies,
4925 Dufferin Street, BRIAN A. WILSON
Downsview, Ontario, Wilson Composite Group,
Canada M3H 5T6 6611 Folsom-Auburn Road, Suite C,
Folsom, CA 95630,
JAMES L. THRONE USA
Sherwood Technologies, Inc.,
158 Brookside Boulevard, S.WONG
Hinckley, OH 44233-9676, Boeing Commercial Airplane Group,
USA Douglas Products Division,
Mail Stop DOOl-0018,
FRANK TRACESKI Long Beach, CA 90846,
Department of Defense, USA
5203 Leesburg Pike Suite 1403,
Falls Church, VA 22041,
USA
xiv Handbook of composites

MAURICE A. WRIGHT PHILIP R. YOUNG

Center for Advanced Friction Studies, Emory & Henry College,
Southern lllinois University at Carbondale, Department of Chemistry,
Carbondale, IL 62901-4343, Emory, VA 24327,
USA USA
PREFACE

Today, fiber reinforced composites are in use • properties of different component (fiber,
in a variety of structures, ranging from space- matrix, filler) materials;
craft and aircraft to buildings and bridges. • manufacturing techniques;
This wide use of composites has been facili- • analysis and design;
tated by the introduction of new materials, • testing;
improvements in manufacturing processes • mechanically fastened and bonded joints;
and developments of new analytical and test- • repair;
ing methods. Unfortunately, information on • damage tolerance;
these topics is scattered in journal articles, in • environmental effects;
conference and symposium proceedings, in • health, safety, reuse, and disposal;
workshop notes, and in government and com- • applications in:
pany reports. This proliferation of the source - aircraft and spacecraft;
material, coupled with the fact that some of - land transportation;
the relevant publications are hard to find or - marine environments;
are restricted, makes it difficult to identify and - biotechnology;
obtain the up-to-date knowledge needed to - construction and infrastructure;
utilize composites to their full advantage. - sporting goods.
This book intends to overcome these diffi-
Each chapter, written by a recognized expert,
culties by presenting, in a single volume,
is self-contained, and contains many of the
many of the recent advances in the field of
'state-of-the-art' techniques reqUired for prac-
composite materials. The main focus of this
tical applications of composites. Thus, this
book is on polymeric matrix, metal matrix,
book should serve as a useful source of infor-
and ceramic matrix composites. The book
mation for practicing engineers and
treats a wide range of subjects. The topics, pre-
specialists, as well as for workers new to this
sented in 49 chapters and two appendices
field.
include:
George S. Springer
• overview of composite material systems
and products;
ABOUT THE EDITOR

S.T. Peters was previously a fellow engineer He is a private consultant with worldwide
with Westinghouse Electric Corporation, clients and has presented tutorials on compos-
Marine Division prior to devoting full time to ites to many audiences, including the US
composite and materials and processing con- Navy and NASA, several technical societies
sulting for his own company, Process and two universities. He is a licensed profes-
Research, in Mountain View, CA. He has writ- sional engineer in the state of California, a
ten many articles on composites and filament member of ASM, and the composites division
winding, a book on filament winding, edited of SME and has been elected a fellow of
one previous book and holds several patents SAMPE.
on winding techniques and composite joints.
FOREWORD

Composites have emerged as a valuable class and is one of the prime drivers for' concurrent
of engineering materials because they offer engineering'. The needs of concurrent engi-
many attributes not attainable with other neering require all contributors to a successful
materials. Light weight, coupled with high composite structure design to have some
stiffness, and selectable properties have fos- knowledge of composites. It is hoped that this
tered their use for many years in satellites, handbook will facilitate the acquisition of that
high performance aircraft and world class sail- knowledge.
boats as well as submarines. Now, these There is little change in scope or objectives
materials demonstrate their worth in the mun- from the first edition, but many changes in
dane, but equally as demanding, consumer, content. Organic matrix composites are still
infrastructure, and sporting goods arenas. highlighted and a limited treatment of metal
However, the transition to many of these con- matrix, carbon and ceramic matrix compos-
sumer and infrastructure-relat¢d arenas has ites, as well as treatment of the new polymers,
been slow, primarily due to the l,ack of under- including the thermoplastic materials have
standing, the lack of validated data, absence of been added.
design allowables, lack of clear guidelines, The treatment of processing methods has
limited hands-on experience, low confidence been expanded for reader convenience, dou-
and less than optimum supplier data sheets. It bling the previous number of chapters. The
is hoped that this book, crafted by well known structural efficiency of a composite structure is
contributors, pre-eminent in their field, mem- often determined by its joints, not by its basic
bers of international universities, industrial structure. Thus, an extensive treatment of
research organizations, and government labs composite surface treatment, mechanical fas-
will remedy some of these weak areas and that tening, and adhesive bonding, which continue
this edition will continue as the primary refer- to be important, sometimes vexing areas of
ence in the composites field and as basic a text concern in composite structure fabrication,
for educational uses as was the first edition. have been added.
The editor of the first edition, George The analysis section has been expanded
Lubin, was a valued mentor to many, includ- with chapters dealing with laminate design,
ing the present editor. This book carries on the composite structure design and a new, impor-
legacy he started: that of a compact but thor- tant subject, design allowables substantiation.
ough sourcebook, with validated, useful data There are new chapters on repair, safety and
to increase understanding, incorporating reuse of composites as well as chapters on the
application guidelines, fabrication methods applications of composites to medical, con-
and tradeoffs, basic polymer information and struction and sporting goods uses that reflect
design recommendations. The successful, current concerns and important new uses for
cost-effective fabrication of composite struc- composite materials.
tures requires utmost cooperation between the
S.T. Peters
various design and manufacturing specialties
ACKNOWLEDGEMENTS

As with any large undertaking there is a sup- approach to resolving problems. Thanks also
porting group of people without whose help go to Mr Frank Hell and Dr Alvin Nakagawa
the objective would not be met. I wish to of Westinghouse Electric, Marine Division
acknowledge my wife, Lynn, for her help in (now Norton Grumman) for their editorial
deciphering and rewriting some of the articles and review help. I also wish to thank Dr Linda
and for enduing my sometimes uncivil Clements for her advice and support.
INTRODUCTION, COMPOSITE BASICS AND
ROADMAP*
S.T. Peters

This is an introduction to composites and will metallic, ceramic or carbon. When the fiber
encourage the reader to obtain more informa- and the matrix are joined to form a composite
tion. Only the basic concepts will be covered they retain their individual identities and both
here; reference will be made to the chapters in directly influence the composite's final proper-
the book that expand or follow up and elabo- ties. The resulting composite will generally be
rate on these basics. The reader will see that composed of layers (laminae) of the fibers and
the subjects of this book cover the spectrum of matrix stacked to achieve the desired proper-
composites and range from the basic and sim- ties in one or more directions.
ple to the complex. Thus, there are The high strength or stiffness to weight
complicated equations because they are the ratios of advanced composites are well known,
tools that are used every day to describe real but there are other advantages also (Table I.1).
structures; and there will also be the more gen- These advantages translate not only into air-
eral, less complicated approaches that are craft, but into everyday activities, such as
limited in analysis power. These chapters have longer drives with a graphite-shafted golf club
been developed by the most knowledgeable (because more of the mass is concentrated at
composite professionals in the world; a blend the clubhead) or less fatigue and pain because
of academicians and the engineers who fabri- a graphite composite tennis racquet has inher-
cate real composite structures. ent damping. Generally, the advantages accrue
Modem structural composites, frequently for any fiber I composite combination and dis-
referred to as 'Advanced Composites', are a advantages are more obvious with some. These
blend of two or more components, one of advantages have now resulted in many more
which is made up of stiff, long fibers, and the reasons for composite use as shown in Table
other, a binder or 'matrix' which holds the I.2. Proper design and material selection can
fibers in place. The fibers are strong and stiff circumvent many of the disadvantages.
relative to the matrix and are generally
orthotropic (having different properties in two 1.1 MATERIAL SYSTEMS
different directions). The fiber, for advanced
structural composites, is long, with length to An advanced composite laminate can be tai-
diameter ratios of over 100. The fiber's lored so that the directional dependence of
strength and stiffness are usually much strength and stiffness matches that of the load-
greater, perhaps several times more, than the ing environment. To do that, layers of
matrix material. The matrix material can by unidirectional material called laminae are ori-
polymeric (e.g. polyester resins, epoxies),
* This chapter has been adapted from S.T. Peters, in
Handbook of Plastics Elastomers and Composites, 3rd edn, (ed.
Handbook of Composites. Edited by S.T. Peters. Published c.A. Harper). McGraw-Hill, New York, 1996, and is used
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 with permission of the McGraw-Hill companies.
2 Introduction, composite basics and road map

Table 1.1 Advantages/disadvantages of advanced composites

Advantages Disadvantages
Weight reduction Cost of raw materials and fabrication
High strength or stiffness to weight ratio
Tailorable properties Transverse properties may be weak
Can tailor strength or stiffness to be in the
load direction
Redundant load paths (fiber to fiber) Matrix is weak, low toughness
Longer life (no corrosion) Reuse and disposal may be difficult
Lower manufacturing costs because of Difficult to attach
less part count
Inherent damping Analysis is difficult
Increased (or decreased) thermal or Matrix subject to environmental degradation
electrical conductivity

ented to satisfy the loading requirements. These Carboni graphite fibers (Chapter 9) have
laminae contain fibers and a matrix. Because of demonstrated the widest variety of strengths
the use of directional laminae, the tensile, flex- and modulii and have the greatest number of
ural and torsional shear properties of a suppliers. The fibers begin as an organic fiber,
structure can be disassociated from one another rayon, polyacrylonitrile or pitch which is
to some extent and a golf shaft, for example, can called the precursor. The precursor is then
be changed in torsional stiffness without chang- stretched, oxidized, carbonized and graphi-
ing the flexural or tensile stiffness. tized. There are many ways to produce these
Fibers can be of the same material within a fibers, but the relative amount of exposure at
lamina or several fibers mixed (hybrid). The temperatures from 2S00-3000°C results in
common commercially available fibers are as greater or less graphitization of the fiber.
follows: Higher degrees of graphitization usually
result in a stiffer fiber (higher modulus) with
• fiberglass; greater electrical and thermal conductivities
• graphite; and usually higher cost.
• aramid; The organic fiber Kevlar 49, (Chapter 10)
• polyethylene; also called aramid, essentially revolutionized
• boron; pressure vessel technology because of its great
• silicon carbide; tensile strength and consistency coupled with
• silicon nitride, silica, alumina, alumina silica.
low densi~ resulting in much more weight
The advantages of fiberglass (Chapter 7) are its effective designs for rocket motors. Aramid
high tensile strength and strain to failure, but composites are still widely used for pressure
heat and fire resistance, chemical resistance, vessels but have been largely supplanted by
moisture resistance and thermal and electrical the very high strength graphite fibers. Aramid
properties are also cited as reasons for its use. composites have relatively poor shear and
It is by far the most widely used fiber, primar- compression properties; careful design is
ily because of its low cost; but its mechanical requires for their use in structural applications
properties are not comparable with other that involve bending or compression.
structural fibers.
 Material systems 3

Table 1.2 The reasons for using composites

Reason for use Material selected Applicationldriver

Lighter, stiffer stronger Boron, all carboni graphites, Military aircraft, better
somearamid performance
Commercial aircraft, operating costs
Controlled or zero Very high modulus Spacecraft with high positional
thermal expansion carboni graphite accuracy requirements for optical
sensors
Environmental resistance Fiberglass, vinyl esters, Tanks and piping, corrosion
bisphenol A fumarates, resistance to industrial chemicals,
chIorendic resins crude oil, gasoline at elevated
temperatures
Lower inertia, faster startups, High strength Industrial rolls, for paper, films
less deflection carboni graphite, epoxy
Lightweight, damage tolerance High strength carboni graphite, CNG tanks for 'green' cars, trucks
fiberglass, (hybrids), epoxy and busses to reduce environmental
pollution
More reproducible complex High strength or high High-speed aircraft. Metal skins
surfaces modulus carbon graphite I cannot be formed accurately
epoxy
Less pain and fatigue Carboni graphite I epoxy Tennis, squash and racquetball
racquets. Metallic racquets are no
longer available
Reduces logging in Aramid, carboni graphite Laminated 'new' growth wooden
'old growth' forests support beams with high modulus
fibers incorporated
Reduces need for intermediate High strength Cooling tower driveshafts
support and resists constant carboni graphite-epoxy
100% humidity atmosphere
Tailorability of bending and Carboni graphite-epoxy Golf shafts, fishing rods
twisting response
Transparency to radiation Carboni graphite-epoxy X-ray tables
Crashworthiness Carboni graphite-epoxy Racing cars
Higher natural frequency, lighter Carboni graphite-epoxy Automotive and industrial
driveshafts
Water resistance Fiberglass (woven fabric), Commercial boats
polyester or isopolyester
Ease of field application Carboni graphite, fiberglass Freeway support structure repair
- epoxy, tape and fabric after earthquake

The polyethylene fibers have the same their use to composites that cure or operate
property drawbacks as aramids, but also suf- below 149°C (300°F) and a susceptibility to
fer from low melting temperature which limits degradation by ultraviolet light exposure.
4 Introduction, composite basics and road map

Both of these types of fibers have wide usage 1.2 MATRIX SYSTEMS
in personal protective armor. In spite of the If parallel and continuous fibers are combined
drawbacks, production of both of these fibers
with a suitable matrix and cured properly, uni-
is enjoying strong worldwide growth. directional composite properties such as those
Boron fibers (Chapter 8), the first advanced shown on Table 1.3 are the result.
composite fibers to be used on production
The functions and requirements of the
aircraft, are produced as individual mono- matrix are to:
filaments upon a tungsten or carbon substrate
by pyrolytic reduction of boron trichloride • keep the fibers in place in the structure;
(BCI3) in a sealed glass chamber. The relatively • help to distribute or transfer loads;
large cross section fiber is used today primar- • protect the filaments, both in the structure
ily in metal matrix composites which are and before and during fabrication;
processed at temperatures which would attack • control the electrical and chemical proper-
carboni graphite fibers. ties of the composite;
• carry interlaminar shear.

Table 1.3 Properties of typical unidirectional graphite/ epoxy composites (Fiber volume fraction, VI = 0.62)

High strength High modulus

Elastic constants CPa (psi x 106) CPa (psi x 106)

Longitudinal modulus, EL 145 (21) 220 (32)
Transverse modulus, ET 9.6 (1.4) 6.9 (1.0)
Shear modulus, C LT 5.8 (0.85) 4.8 (0.7)
Poisson's ratio (dimensionless) Uu 0.30 0.30
-----------------------------------------------------
Strength properties MPa (1(J3 psi) MPa (10 3 psi)
Longitudinal tension, puL 2139 (310) 760 (110)
Transverse tension, FtuT 54 (7.8) 28 (4)
Longitudinal compression, FeuL 1724 (250) 690 (100)
Transverse compression, pm} 76 (11) 170 (25)
Inplane shear, P\T 87 (12.6) 70 (10)
Interlaminar shear, FLsu 128 (18.5) 70 (10)

Ultimate strains % %

Longitudinal tension, EtuL 1.4 0.3

Transverse tension, EtuT 0.67 0.4
Longitudinal compression, ECUL 0.9 0.3
Transverse compression, EcuT 1.6 2.8
Inplane shear 2.0

Physical properties
Density, kg/m3 (lb/in3) 1600 (0.056) 1700 (0.058)
Longitudinal CTE,
IlE/K (llE/oF) -0.079 (-0.044) -0.54(-0.3)
Transverse CTE
IlE/K (llEjOF) 21.6 (12) 58 (32)
From References 1, 2 and 3; CTE = coefficient of thermal expansion
 Matrix systems 5

The needs, or desired properties of the composite. The common thermoset matrices
matrix, that depend on the purpose of the struc- for composites include the following:
ture are:
• polyester and vinylesters (Chapter 2);
• minimize moisture absorption; • epoxy (Chapter 3);
• have low shrinkage; • bismaleimide (BMI) (Chapter 4);
• Must wet and bond to fiber; • polyimide (Chapter 4);
• low coefficient of thermal expansion; • cyanate ester and phenolic triazine
• must flow to penetrate the fiber bundles (Chapter 5).
completely and eliminate voids during the
Each of the resin systems has some drawbacks,
compacting/ curing process;
which must be accounted for in design and
• have reasonable strength, modulus and
manufacturing plans.
elongation (elongation should be greater
Polyester matrices have been in use for the
than fiber);
longest period, and are used in the widest
• must be elastic to transfer load to fibers;
range and greatest number of structures. The
• have strength at elevated temperature
usable polymers may contain up to 50% by
(depending on application);
weight of unsaturated monomers and solvents
• have low temperature capability (depend-
such as styrene. Polyesters cure via a catalyst
ing on application);
(usually a peroxide) resulting in an exothermic
• have excellent chemical resistance (depend-
reaction, which can be initiated at room tem-
ing on application);
perature.
• be easily processable into the final compos-
The most widely used matrices for
ite shape;
advanced composites have been the epoxy
• have dimensional stability (maintain its
resins. These resins cost more than polyesters
shape).
and do not have the high temperature capabil-
There are many matrix choices available; each ity of the bismaleimides or polyimides, but
type has impact on the processing technique, because of the advantages shown in Table 1.4
physical and mechanical properties and envi- they are widely used.
ronmental resistance of the finished

Table 1.4 Selection criteria for epoxy resin systems

Advantages Disadvantages
Adhesion to fibers and to resin Resins and curatives somewhat toxic in uncured form
No by-products formed during cure Absorb moisture
Low shrinkage during cure Heat distortion point lowered by moisture absorption
High or low strength and flexibility Change in dimensions and physical properties
due to moisture absorption
Solvent and chemical resistance Limited to about 200°C upper temperature use (dry)
Resistance to creep and fatigue Difficult to combine toughness and high
temperature resistance
Solid or liquid resins in uncured state High thermal coefficient of expansion
Wide range of curative options High degree of smoke liberation in a fire
Adjustable curing rate May be sensitive to ultraviolet light degradation
Good electrical properties Slow curing
6 Introduction, composite basics and road map

There are two resin systems in common use radiation under cryogenic conditions was a
for higher temperatures, bismaleimides and prime cause for their choice for use in the
polyimides. New designs for aircraft demand superconducting supercollider, subsequently
a 177°C (350°F) operating temperature not met canceled by the US Congress.
by the other common structural resin systems. Polyimides are the highest temperature
The primary bismaleimide in use is based on polymer in general advanced composite use
the reaction product from methylene dianiline with a long term upper temperature limit of
(MDA) and maleic anhydride: bis (4-maleimi- 232°C (450°F) or 316°C (600°F). Two general
dophenyl) methane (MDA BMI). types are: condensation polyimides, that
Two newer resin systems have been devel- release water during the curing reaction, and
oped and have found applications in widely addition type polyimides with somewhat eas-
diverse areas. The cyanate ester resins, mar- ier process requirements.
keted by Ciba-Geigy, have shown superior
dielectric properties and much lower moisture
1.3 FIBER MATRIX SYSTEMS
absorption than any other structural resin for
composites. The dielectric properties have The end user sees a composite structure.
enabled their use as adhesives in multilayer Someone else, probably a prepregger com-
J

microwave printed circuit boards, and the low bined the fiber and the resin system and
moisture absorbance have caused them to be someone else caused the cure and compaction
the resin of universal choice for structurally- to result in a laminated structure. A schematic
stable spacecraft components. of the steps to arrive at a finished composite
The phenolic triazine (PT) resins also have from the initial fiber is shown in Fig. 1.1.
superior elevated temperature properties, In many cases, the end user of the structure
along with excellent properties at cryogenic has fabricated the composite from prepregJ
temperatures. Their resistance to proton which is a low-temperature-stable combination

rl
YES
WOUEN
FABAIC
NO------,

NO

>-----YES----IlI

Fig. I.1 Manufacturing steps in composite structure.

 Fiber matrix systems 7

of the resin, its curing agents and the fiber. The The prepreg process for thermoset matrices
three types of continuous fibers, roving, tape can be accomplished by feeding the fiber con-
and woven fabric available as prepregs give tinuous tape, woven fabric or roving through a
the end user many options in terms of design resin-rich solvent solution and then removing
and manufacture of a composite structure. the solvent by hot tower drying. The excess
Although the use of dry fibers and impregna- resin is removed via a doctor blade or meter-
tion at the work (i.e. filament winding, ing rolls and then the product is staged to the
pultrusion or hand lay-up) is very advanta- cold-stable prepreg form, (B stage) (Fig. 1.2).
geous in terms of costs; there are many The newer hot melt procedure for prepregs
advantages to the use of prepregs as shown in is gradually replacing the solvent method
Table 15, particularly for the manufacture of because of environmental concerns. A film of
modern composites. resin that has been cast hot onto release paper
Table 1.5 Advantages of prepregs over wet impregnation

Prepregs reduce the handling damage to dry fibers

Improve laminate properties by better dispersion of short fibers
Prepregs allow the use of hard-to-mix or proprietary resin systems
Allow more consistency because there is a chance for inspection before use
Heat curing provides more time for the proper laydown of fibers and for the resin to move and degas
before cure
Increased curing pressure reduces voids and improves fiber wetting
Most prepregs have been optimized as individual systems to improve processing

Oven

Release
Poly
Unwind

Prepreg
Wind

Pump and
Reservoir

Fig. 1.2 Schematic of typical solvent prepregging process. (Adapted from Reference 2.)
8 Introduction, composite basics and road map

o 0
D
Resin Paper
o Bath Take-up

Impregnation
Zone
Creel Prepreg
Windup

Fig. 1.3 Schematic of typical film impregnating process. (Adapted from Reference 2.)

is fed, along with the reinforcement, through a Wf = weight fraction of fiber

series of heaters and rollers to force the resin wf = weight of fiber
into the reinforcement. Two layers of resin are we = weight of composite
commonly used so that a resin film is on both Pf = density of fiber
sides of the reinforcement; one of the release Pe = density of composite
papers is removed and the prepreg is then vf = volume of fiber
trimmed, rolled and frozen (Fig. 1.3)2. The sol- ve = volume of composite
vent technique has been largely replaced for V f = volume fraction of fiber
advanced fibers because of environmental pol- V m = volume fraction of matrix
lution concerns and a need to exert better Pm = density of matrix.
control over the amount of resin on the fiber. A percentage fiber that is easily achievable
and repeatable in a composite and convenient
1.3.1 UNIDIRECTIONAL PLY PROPERTIES for reporting mechanical and physical proper-
ties for several fibers is 60%. The properties of
The manufacturer of the prepreg reports an
unidirectional fiber laminates are shown in
areal weight for the prepreg and a resin per-
centage, by weight. Each of the different fibers Table 1.3 for carbon/graphite/epoxy. Values
for the other fibers can be seen in their respec-
has a different density, resulting in a composite
tive chapters. These values are for individual
of different density at the same fiber volume
lamina or for a unidirectional composite, and
percentage. Since fiber volume is used to relate
they represent the theoretical maximum (for
the properties of the manufactured composites,
that fiber volume) for longitudinal in plane
the following equations can be used to convert
between weight fraction and fiber volume. properties. Transverse, shear and compression
properties will show maxima at different fiber
(1.1) volumes and for different fibers, depending on
W=
f
how the matrix and fiber interact. These val-
ues can be used to calculate the properties of a
(1.2) laminate which has fibers oriented in several
directions. To do that, the methods of descrip-
where: tion for ply orientation must be introduced.
 Quasi-isotropic laminate 9

1.4 PLY ORIENTATIONS, SYMMETRY AND lined to indicate that half of it lies on either
BALANCE side of the plane of symmetry (Fig.1.4(f).
1.4.1 PLY ORIENTATIONS
I.4.2 SYMMETRY
One of the advantages of using a modem com-
posite is the potential to orient the fibers to The geometric midplane is the reference sur-
respond the load requirements. This means face for determining if a laminate is
that the composite designer must show the symmetrical. In general, to reduce out-of-
material, the fiber orientations in each ply, and plane strains, coupled bending and stretching
how the plies are arranged (ply stackup). of the laminate and complexity of analysis,
A 'shorthand' code for ply fiber orientations symmetric laminates should be used.
has been adapted for use in layouts and studies. However, some composite structures (e.g. fila-
Each ply (lamina) is shown by a number rep- ment wound pressure vessels) can achieve
resenting the direction of the fibers in degrees, geometric symmetry so that symmetry
with respect to a reference (x) axis. 0° fibers of through a single laminate wall is not neces-
both tape and fabric are normally aligned with sary, if it constrains manufacture. To construct
the largest axial load (axis) (Fig. I.4(a». a midplane symmetric laminate, for each layer
Individual adjacent plies are separated by a above the midplane there must exist an identi-
slash in the code if their angles are different cal layer (same thickness, material properties,
(Fig. I.4(b». and angular orientation) below the midplane
The plies are listed in sequence, from one (Fig. 1.4(e».
laminate face to the other, starting with the ply
first on the tool and indicated by the code
1.4.3 BALANCE
arrow with brackets indicating the beginning
and end of the code. All laminates should be balanced to achieve in-
Adjacent plies of the same angle of orienta- plane orthotropic behavior. To achieve balance,
tion are shown by a numerical subscript (Fig. for every layer centered at some positive angle
I.4(c». +8 there must exist an identical layer oriented
When tape plies are oriented at angles equal at - 8 with the same thickness and material
in magnitude but opposite in sign, (+) and (-) properties. If the laminate contains only 0°
are used. Each (+) or (-) sign represents one and/ or 90° layers it satisfies the requirements
ply. A numerical subscript is used only when for balance. Laminates may be midplane sym-
there are repeating angles of the same sign. metric but not balanced and vice versa. Figure
Positive and negative angles should be consis- 1.4(e) is symmetric and balanced whereas Fig.
tent with the coordinate system chosen. An 1.4(g) is balanced but unsymmetric .
orientation shown as positive in one right
handed coordinate system may be negative in
1.5 QUASI-ISOTROPIC LAMINATE
another. If the y and z axis directions are
reversed, the ± 45 plies are reversed (Fig. I.4(d». The goal of composite design is to achieve the
Symmetric laminates with an even number lightest, most efficient structure by aligning
of plies are listed in sequence, stating at one most of the fibers in the direction of the load.
face and stopping at the midpoint. A subscript Many times there is a need, however, to pro-
'5' following the bracket indicates only one duce a composite which has some isotropic
half of the code is shown (Fig.1.4(e». properties, similar to metal, because of multi-
Symmetric laminates with an odd number ple or undefined load paths. A 'quasi-isotropic'
of plies are coded as a symmetric laminate laminate lay-up accomplishes this for the x and
except that the center ply, listed last, is over- y planes only; the z or through-the-Iaminate-
10 Introduction, composite basics and road map

•
"- 1'\"-1'\ "-1'\ "-1'\ /1/ /1/ /V'"

"" v
"""- V
V

"- "",,"- """ ""

"-" ""
/V /V /V /V
"- ,,"- ""-"
"lXX
/V /V VV/
"- ,,"- "
"-
"-" "-" "-
"-"
"-
"
"-1'\ "-"
) Reference Axis

"-",,"'-
900

tE:.:
"-
"- "-
"- lXX rxx a)
)(
"-
vx XIX "- "- "I' 1'\"- I" "-
~o

Tape Laminate
/r---- Tool side

~...
[O/90)s
Typical Callout

if +~" 90" -45" o· Z

b)

0"
90·
90·
90·
I
[0/903/0)
Typical Callout
~.o· -4/5" 9t:f +45· o·
d)
0" I c)

[O/90/±45]s
0" I Typical Callout
90· I
+45" I
-45· I
-45· I
+45" I
90· I e)
0" I
Tape and Fabric Laminate
(0/±45/90]s,
0" I Typical Callout
±45" I
90·
±45·
0" I
I
~ Line of Symmetry
f)

Fig. 1.4 Ply orientations, symmetry and balance. (Continued on next page)
 Methods of analysis 11

Tape Laminate
(]'
I
90· 1
+45 0 I [0/901±451 +45219010]
Typical Callout
-45 0 J
-45 0 I
+45 0 I
-45 0 J
+45 0 I g)
90 0 I
(]'
I
Fabric Laminate [(0 ,90}/(±45 )/(0,90)]
(],,9U' I Typical Callout
±45° I
±45° I
OO,9(]' I h)

Fig. 1.4 Ply orientations, symmetry and balance. (Continued)

thickness plane is quite different and lower. 1. arrive at quick values to determine if a com-
Most laminates produced for aircraft applica- posite is feasible;
tions have been, with few exceptions, 2. arrive at values for insertion into computer
'quasi-isotropic'. As designers become more programs for laminate analysis or finite ele-
confident and have access to a greater database ment analysis;
with fiber-based structures, more applications 3. check on the results of computer analysis.
will evolve. For a quasi-isotropic laminate, the
The rule of mixtures holds for composites. The
following are requirements:
micromechanics formula to arrive at the
• It must have three layers or more. Young's modulus for a given composite is:
• Individual layers must have identical stiff-
ness matrices and thicknesses. Ec = VfEf + VmEm
• The layers must be oriented at equal angles. and V f + Vm "" 1
For example, if the total number of layers is
n, the angle between two adjacent layers = Vf Ef + Em (1 - Vf) (I.3)
should be 360° In. If a laminate is con- where
structed from identical sets of three or more
layers each, the condition on orientation
Ec = compositeor ply Young's modulus in
tension for fibers oriented in direction of
must be satisfied by the layers in each set,
applied load
for example: (0°1 ± 60°). or (0° I ± 45° 190)•.
V = volume fraction of fiber (f) or matrix (m)
E = Young's modulus of fiber (f) or matrix
1.6 METHODS OF ANALYSIS (m).
There are a number of methods in common But, since the fiber has much higher
use for the analysis of composite laminates. Young's modulus than the matrix, the second
The use of micromechanics, i.e. the application part of the equation can be ignored.
of the properties of the constituents to arrive at
Ef » Em
the properties of the composite ply can be
used to: Ec "" EfVf (I.4)
12 Introduction, composite basics and road map

This is the basic rule of mixture and repre- appropriate for a particular application.
sents the highest Young's modulus composite, Figure 1.5 shows the progression of physical
where all fibers are aligned in the direction of properties for Young's modulus in tension, Ef
load. The minimum Young's modulus for a (fiber), EL (lamina) and Exy (laminate), longi-
reasonable design (other than a preponder- tudinal tensile strength, 'and coefficient of
ance of fibers being· orientated transverse to thermal expansion ex where the subscripts L
the load direction) is the quasi-isotropic com- and X stand for in-plane in the principal fiber
posite and can be approximated by: direction and t and Y stand for the transverse
direction for a theoretical high strength (from
(I.5)
Table 13) carboni graphite fiber composite
The quasi-isotropic modulus, E, of a composite from the fiber to the laminate. The values
laminate is (3/8)En+(5/8)E22 where En is the decrease or are 'translated' in a logical fashion
modulus of the lamina in the fiber direction and and reflect the law of mixtures. The analysis is
E22 is the transverse modulus of the lamina3• relatively simple for modulus dominated
The transverse modulus for polymeric-based properties but strength-dominated values
composites is a small fraction of the longitudinal must be treated in light of one of several fail-
modulus (see Et in Table 1.3) and can be ignored, ure theories and changes in the thermal
for preliminary estimates, resulting in a slightly coefficient of expansion are not predictable
lower-than-theoretical value for Ec for a quasi- from laws of mixtures. Other factors which
isotropic laminate. This approximate value for enter into the translation efficiency are: com-
the quasi-isotropic modulus represents the patibility of the resin system with the fiber and
lower limit of composite modulus. It is useful in the fiber finish, strain-to-failure of the resin
comparing of composite properties to those of system and the damage the fiber undergoes
metals and in establishing if a composite is during impregnation, laydown and cure.

811fTl

O.13mm*'
I ===~!a
=~

Ilz > Ily

Fig. 1.5 The anatomy of a composite laminate.

 Composite fabrication techniques 13

Table 1.6 High-strength carbon/ graphite laminate competitor, so vendor values in a generic
properties class may differ widely.
4. Most tables of values are presented as 'typi-
Laminate Longitudinal Bending Shear cal values'. Those values and the values that
modulus modulus, modulus, are part of the menu of many computer
En (GPa) EB (GPa) GXY (GPa) analysis programs should be used with care.
(0/9°2 /0) 76.5 126.8 5.24 Each user must find their own set of values
(90/0/90) 76.5 26.3 5.24 for design, develop useful design allow-
(0/9%2) 98.5 137.8 5.24 ables, and apply appropriate 'knock down'
(0/±45/02) 81.3 127.5 21.0 factors, based on the operating environ-
(0/ ±45/90)s 55.0 89.6 21.0 ments expected in service. (Chapter 33 and
Aluminum 41.34 41.34 27.56
Appendix A give guidelines.)

Table 1.6 shows mechanical values for sev- 1.7 COMPOSITE FABRICATION TECHNIQUES
eral composite laminates with the fiber of
The goals of the composite manufacturing
Table 1.3 and a typical resin system. The first
process are to:
and second entries are for simple 0/90 lami-
nates and show the effect of changing the • achieve a consistent product by controlling
position of the plies. The effect of increasing - fiber thickness;
the number of 0 plies is shown next and the - fiber volume;
final two laminates demonstrate the effect of - fiber directions;
±4S plies on mechanical properties, particu- • minimize voids;
larly the shear modulus. The last entry is a • reduce internal residual stresses;
quasi-isotropic laminate. These laminates are • process in the least costly manner.
then compared to a typical aluminum alloy. The procedures to reach these goals involve
When employing the data extracted from iterative processes to select the three key com-
tables, some caution should be observed by ponents:
the reader. The values seen in many tables of
data may not always be consistent for the • composite material and its configuration;
same materials or the same group of materials • tooling;
from several sources for the following reasons: • process.
1. Manufacturers have been refining their pro- Once material selection has been completed,
duction processes so that newer fibers may the first step leading to the acceptable com-
have greater strength or stiffness. These posite structure is the selection of tooling,
new data may not be reflected in the com- which is intimately tied to process and mater-
piled data. ial. For all curing techniques the tool must be:
2. The manufacturer may not be able to • strong and stiff enough to resist the pres-
change the value quoted for the fiber sure exerted during cure;
because of government or commercial • dimensionally stable through repeated
restrictions imposed by the specification heating and cooling cycles;
process of his customers. • light enough to respond reasonably quickly
3. There are many different high-strength to the changes in cure cycle temperature
fibers commercially available. Each manu- and to be moved in the shop;
facturer has optimized their process to • leakproof so that the vacuum and pressure
maximize their mechanical properties and cycles are consistent.
each process may differ from that of the
14 Introduction, composite basics and road map

The tool face is commonly the surface the introduction of elastomeric tooling
imparted to the outer surface of the composite wherein the thermal expansion of an elas-
and must be smooth, particularly for aerody- tomer provides some or all of the pressure
namic surfaces. The other surface frequently curing cure, or a rubber blanket is used as a
may be of lower finish quality and is imparted reusable vacuum bag. The volumetric expan-
by the disposable or reusable vacuum bag. This sion of an elastomer can be used to fill a cavity
surface can be improved by the use of a sup- between the uncured composite and an outer
plemental metal tool known as a caul plate. mold. The use of elastomeric tooling can pro-
(Press curing, resin transfer molding, injection vide the means for fabricating complex
molding and pultrusion require a fully closed box-like structures such as integrally stiffened
or two sided mold). Figure 1. 6 shows the basic skins with a co-cured substructure in a single
components of the tooling for vacuum bag or curing operation.
autoclave processed components and Table 1.7 Tooling (Chapter 25) and the configuration
shows the function of each part of the system. of the reinforcement have a great influence on
Tooling options have been augmented by the curing process selected and vice-versa. The

Fig. 1.6 Typical vacuum bag lay-up components.

Table 1.7 Functions of vacuum bag components

Component .. Functions
1 Bag sealant Temporarily bonds vacuum bag to tool
2 Vacuum fitting and hardware Exhausts air, provides convenient connection to vacuum pump
3 Bagging film Encloses part, allows for vacuum and pressure
4 Open weave breather mat Allows air or vacuum transfer to all of part
5 Polyester tape (wide) Holds other components of bag in place
6 Polyester tape (narrow) Holds components in place
7 Caul sheet Imparts desired contour and surface finish to composite
8 Perforated release film Allows flow of resin or air without adhesion
9 Non-perforated release film Prevents adhesion of laminate resin to tool surface
10 Peel ply Imparts a bondable surface to cured laminate
11 Laminate
12 1581-style glass breather manifold Allows transfer of air or vacuum
13 1581 style glass bleeder ply Soaks up excess resin
14 Stacked silicone edge dam Forces excess resin to flow vertically, increasing fluid pressure
" numbers refer to Fig. 1.6
 Composite fabrication techniques 15

probable reinforcement configuration that Table 1.8 Common reinforcement configuration for
facilitates the completion of the finished com- the manufacturing process
posite is shown on Table 1.8.
The choice between unidirectional tape and Reinforcement Prepreg Prepreg Prepreg Other,
configuration tape or (dry) or (dry) woven
woven fabric has frequently been made on the tow woven preforms,
basis of the greater strength and modulus or non- chapped
attainable with the tape particularly in appli- woven fibers
cations which compression strength is fabric
important. There are other factors that should Hand lay-up X X, (X) X
be included in the trade, as shown in Table 1.9. Automatic tape X
laydown
1.7.1 LAY-UP TECHNIQUE Filament X, (X) X,(X) X,(X)
winding
Lay-up techniques along with composite cure
control have received the greatest attention for Resin (X) (X) X
processing. In efforts to reduce labor costs of transfer
molding
composite fabrication, to which lay-up
(Chapter 16) has traditionally been the largest Pultrusion (X) X
contributor, mechanically assisted, controlled Fiber X X
tape laying and automated integrated manu- placement

Table 1.9 Fabric compared with tape reinforcement

Tape advantages Tape disadvantages

Best modulus and strength efficiency Poor drape on complex shapes
High fiber volume achievable Cured composite more difficult to machine
Low scrap rate Lower impact resistance
No discontinuities Multiple plies required for balance and symmetry
Automated lay-up possible Higher labor costs for hand lay-up
Available in thin plies
Lowest cost prepreg form
Less tendency to trap volatiles
Fabric advantages Fabric disadvantages
Better drape for complex shapes Fiber discontinuities (splices)
Single ply is balanced and may be Less strength and modulus efficient
essentially symmetric
Can be laid up without resin Lower fiber volume than tape
Plys stay in line better during cure More costly than tape
Cured parts easier to machine Greater scrap rates
Better impact resistance Warp and fill properties differ
Many forms available Fabric distortion can cause part warping
16 Introduction, composite basics and road map

facturing systems have been developed. Table Generally, the percent matrix weight is higher
1.10 shows some of the considerations for before cure initiation; the matrix flows out of
choosing a lay-up technique. the laminate and takes the excess resin with
In addition to any cost savings by the use of the potential voids. An arbitrary 1% void limit
an automated technique for long production has been adopted for most autoclaved com-
runs, there are two key quality assurance fac- posites; filament wound and pultruded
tors which validate the automated techniques. composites will have higher void volumes
They are: greatly reduced chance that release depending upon the application.
paper or film could be retained, which would An autoclave is essentially a dosed, pres-
destroy shear and compressive strength if surized oven; many common epoxy laminates
undetected, and reduced probability of the are cured at an upper temperature of 177°C
addition or loss of an angle ply which would (350 0 P) and 6 MPa (100 psi). Autoclaves are
cause warping due to the laminate's lack of still the primary tool in advanced composite
symmetry and balance. processing and have been built up to 16 m (55
All curing techniques use heat and pressure feet) long at 6.1 m (20 feet) diameter. Since
to cause the matrix to flow and wet out all the autoclaves are expensive to build and operate,
fibers before the matrix solidifies (Chapter 26). many other methods of curing, compacting

Table 1.10 Considerations in composite lay-up technique

Consideration Manual Flat tape Contoured tape

Orientation accuracy Least accurate Automatic Somewhat dependent on
tape accuracy and
computer program
Ply count Dependent on operator, Dependent on Program records
count Mylars operator
Release film retention Up to operator Automatic Automatic removal
Labor costs High 86% improvement Additional improvement
quoted
Machine costs N/A Some costs Approximately IM$ or
greater
Production rate Low (l.5lb/h) 10 lb/h Approximately same as flat
tape
Machine 'up' time N/A Not a consideration Complex program and
machine make this a
consideration
Varying tape widths Notaconcem Easily changed Difficulty in changing
Tape lengths Longer tapes more Longer is more Longer tape is more
difficult economical economical
Cutting waste Scrap on cutting Less scrap Least scrap due to back and
forth laydown
Compaction pressure No pressure Less voids Least voids
Programming N/A N/A Necessary
 Composite fabrication techniques 17

composites have been developed. The two a wind eye at speeds synchronized with the
newest and most attractive methods are fiber mandrel rotation, control winding angle of the
placement and resin transfer molding. reinforcement and the fiber lay-down rate.
The reinforcement may be wrapped in adja-
cent bands or in repeating bands that are
1.7.2 RESIN TRANSFER MOLDING
stepped the width of the band and that even-
Previous discussions have centered on moving tually cover the mandrel surface. Local
resin out of the laminate to reduce voids. Resin reinforcement can be added to the structure
transfer involves the placement of dry fiber using circumferential windings, local helical
reinforcement into a closed mold and then bands, or by the use of woven or unidirec-
injecting a catalyzed resin into the mold to tional cloth. The wrap angle can be varied
encapsulate the reinforcement and form a from low angle helical to high angle circum-
composite (Chapter 20). The impetus for the ferential or 'hoop', which allows winding from
use of this process comes from the large cost about 4°-90° relative to the mandrel axis;
reductions that can be realized in raw materi- newer machines can 'place' fiber at 0°.
als and lay-up. The process can utilize low
injection pressures i.e. 55 MPa (80 psi), there-
1.7.4 FIBER PLACEMENT
fore, the tooling can be lower cost plastic or a
vacuum bag rather than metal. Fiber placement, initially developed by
Hercules Aerospace Co., is a cross between fil-
ament winding and automatic tape laydown,
1.7.3 FILAMENT WINDING
retaining many of the advantages of both. The
Filament winding is a process by which con- natural outgrowth of adding multiple axes of
tinuous reinforcements in the form of rovings control to filament winding machines results
or tows (gathered, untwisted strands of fiber) in control of the fiber laydown so that non axi-
are wound over a rotating mandrel. The man- symmetric surfaces can be wound. This
drel can be cylindrical, round or any other involves the addition of a modified tape lay-
shape as long as it does not have re-entrant down head to the filament winding machine
curvature. Special machines (Fig. 1.7) traversing and much more. The Cincinnati-Milacron
machine additions include in-process com-
paction, individual tow cut/ start capabilities,
a resin tack control system, differential tow
payout, low tension on fiber and enhanced off-
line programming (Chapter 22).

1.7.5 PULTRUSION
Pultrusion is an automated process for the
manufacture of constant volume/shape pro-
files from composite materials (Chapter 23).
The composite reinforcements are continu-
ously pulled through a heated die and shaped
and cured simultaneously. If the cross-sec-
tional shape is conducive to the process, it is
the fastest and most economical method of
Fig. 1.7 The helical filament wound ply. (Courtesy composite production. Straight and cured con-
of Westinghouse Electric Co., Marine Division.) figurations can be fabricated with square,
18 Introduction, composite basics and road map

round, hat-shaped, angled 'I' or 'T' -shaped

cross-sections from vinylester, polyester, or
epoxy matrices with E and 5-glass, Kevlar and
graphite reinforcements.. The curing is
effected by combinations of dielectric preheat-
ing and microwave or induction (with
conductive reinforcements like carbon
graphite) while the shape traverses the die.

3
1.7.6 BRAIDING, WEAVING AND OTHER
PREFORM TECHNIQUES Fig. 1.8 The unidirectional ply.
Braiding, weaving, knitting and stitching rep-
resent methods of forming a shape, generally be the same in any transverse direction. This is
referred to as preforming, with the composite the transverse isotropy assumption; it is
fibers before impregnation (Chapter 18). The approximately satisfied for most unidirec-
shape may be the final product or some inter- tional composite plies.
mediate form such as a woven fabric. The These properties are typically modified by
braiding process is continuous and is transformation relative to the laminate axis
amenable to round or rectangular shapes or where these may not be the same as the ply
smooth curved surfaces and can transition axes.
easily from one shape to another. In a multidirectional laminate there can be
The other fabric preforming techniques are as many as 21 stiffness constants. Strength pre-
weaving, knitting and the non-structural dictions are equally as complicated because of
stitching of unidirectional tapes. Stitching sim- directional differences, i.e. compression is not
ply uses a non-structural thread, such as nylon always equal to tension, and because the sev-
or Dacron, to hold dry tapes at selected fiber eral failure theories are complex. As the
angles. Preforming in this manner results in a complexity of the matrix calculations increase,
higher-cost raw material but saves labor costs it becomes evident that errorless mathematical
for orientation of individual lamina. The manipulations are impossible without the aid
stitched preform has known, stable fiber ori- of computers. Chapters 30 and 32 elaborate on
entations similar to woven fabric, without the the techniques of laminate analysis and the
crossovers which could reduce compressive applications of laminates to structures
strength.
1.9 DESIGN OF COMPOSITES
1.8 MECHANICS OF COMPOSITE MATERIALS
The design process for composites involves
The 1,2,3 axes in Fig. 1.8 are special and are both laminate design and component design
called the ply axes, or material axes. The 1 axis and must also include considerations of man-
is in the direction of the fibers, and is called the ufacturing process and eventual
longitudinal axis or the fiber axis. environmental exposure. These steps are all
The longitudinal axis is typically the high- interdependent with composites and the most
est stiffness and strength direction. Any efficient design must involve true concurrent
direction perpendicular to the fibers (in the 2,3 engineering. Figure 1.9 shows the various con-
plane) is called a transverse direction. cerns that should be a part of the composite
Sometimes, to simplify analysis and test design process at the initiation of the design
requirements, ply properties are assumed to process, and continuously from there on.
 Design of composites 19

1.9.1 LAMINATE DESIGN loads. This improves handling character-

RECOMMENDATIONS istics, and serves to prevent large matrix
cracks from forming.
1. Take advantage of the orthotropic nature of
the fiber composite ply. • Locally reinforce with fabric or mat in
areas of concentrated loading. (This tech-
• To carry in-plane tensile or compressive
nique is used to locally reinforce pressure
loads align the fibers in the directions of
vessel domes).
these loads.
• Use fabric, particularly fiberglass or
• F~r. in-plane shear loads, align most
Kevlar, as a surface ply to restrict surface
fibers at ± 45° to these shear loads.
(handling) damage.
• For combined normal and shear in-plane
loading provide multiple or intermediate • Ensure that the laminate has sufficient
fiber orientations to avoid dependence
ply angles for a combined load capabil-
on the matrix for stability. A minimum
ity.
coverage of 6 to 10% of total thickness in
2. Intersperse the ply orientations. 0, ±45°, 90° directions is recommended.
• If a design requires a laminate with 16
3. Select the lay-up to avoid mismatch of
plies at ±45°, 16 plies at 0°, and 16 plies at
properties of the laminate with those of the
90°, use the interspersed design (90/ ±
adjoining structures - or provide a
452 /02)40 rather than (90s/ ± 45 /10 ) .
Concentrating. I 8 8. shear / separator ply.
pies at nearly the same
• Poisson's -ratio: if the transverse strain of
angle (0° and 90° in the above example)
a laminate greatly differs from that of
provides the opportunity for large
adjoining-structure, large interlaminar
matrix cracks to form. These produce
stresses are produced under load.
lower laminate allowables, probably
• Coefficient of thermal expansion: tem-
because large cracks are more injuriou,s
perature change can produce large
to the fibers, and more readily form
interlaminar stresses if coefficient of ther-
delaminations than the finer cracks
mal expansion of the laminate differs
occurring in interspersed laminates.
greatly from that of adjoining structure. '
• If a design requires all 0° plies, some 90°
plies (and perhaps some off-angle plies) • The ply layer adjacent to most bonded
joints should not be perpendicular to the
should be interspersed in the laminate to
direction of loading. Thicken the com-
provide some biaxial strength and stabil-
posite in the joint area, soften the
ity and to accommodate unplanned
co~posite by adding fiberglass or angle
plies and select the highest strain-capa-
CompoeJta bility adhesive.
Mat.'a' 4. Use multiple ply angles.
Typical composite laminates are constructed
from multiple unidirectional or fabric layers
which are positioned at angular orientations
in a specified stacking sequence. From many
choices, experience suggests a rather narrow
range of practical construction from which
the final laminate configuration is usually
selected. The multiple layers are usually ori-
ented in at least two different angles, and
Fig. 1.9 Design considerations for composites. possibly three or four; (±80 , 0° /±80 , or
20 Introduction, composite basics and road map

0° / ±f}0/90° cover most applications, with e attempt to standardize the raw materials and
between 30 and 60 degrees). Unidirectional their test methods by publication of specifica-
laminates are rarely used except when the tions (Appendix A). However, these standards
basic composite material is only mildly have not reached the level of use to allow com-
orthotropic (e.g. certain metal matrix appli- plete dependence upon them without
cations) or when the load path is absolutely supplier-user interaction and user testing.
known or carefully oriented parallel to the The fabricators of composites will rely on
reinforcement (e.g. stiffener caps). specifications for control of fiber, resin and/ or
the prepreg. Many prepreg resin and fiber
Further suggestions can be seen in Chapter 31. vendors will certify only to their own specifi-
cations which may differ from those shown;
1.10 COMPOSITE TESTING users should consult the vendors to determine
what certification limits exist before commit-
To ensure consistent, reproducible compo- ting to specification control.
nents, three levels of testing are employed: As part of raw materials verification, com-
incoming materials testing, in-process testing posite design effort and final product
and control and final structure verification. verification mechanical testing of composite
test specimens will be performed. The testing
1.10.1 INCOMING MATERIALS TESTING of composite materials offers unique chal-
lenges because of the special characteristics of
Incoming materials testing seeks to verify the composites. Factors not considered important
conformance of the raw materials to specifica- in metals testing are very important in testing
tions and to insure processibility. The levels of composites (Chapters 34, 39).
knowledge of composite raw materials do not
approach those for metals, which can be
bought to several consensus specifications and REFERENCES
will appear generally identical although pur- 1. Foral, RF. and Peters, S.T., Composite
chased from many manufacturers. Although Structures and Technology Seminar Notes, 1989
there are fewer suppliers for composite raw 2. Hercules Data Sheet for AS-4/3901-6 prepreg
materials, the numbers of permutations of HOSO-377 /GF Prod Hdbk (4)/jc/2
3. Agarwal, B.D. and Broutman, L.J., Analysis and
resins, fibers and manufacturers prevents the
Perfor11Ulnce of Fiber Composites 2nd edn, John
kind of standardization necessary to be able to Wtley and Sons, New York, 1990 p. 103
buy composite raw materials as if they were 4. Mayorga, G.D. in International Encyclopedia of
alloys. ASTM (American Society for Testing Composites, (ed. S.M. Lee) Vol 4, VCH
and Materials), SAE/ AMS/NOMETCOM Publishers, N.Y., N.Y., 1991
(Society of Automotive Engineers, Aeronautical 5. Tsai, S.w. and Pagano, N.J. in Composite
Materials Standards/ Nonmetallic Materials Materials Workshop, (eds. S.w. Tsai, I.e. Halpin
Committee) and SACMA (Suppliers of and N.J. Pagano), Technomic Publishing Co.,
Lancaster, PA, 1978, p. 249
Advanced Composite Materials Association)
OVERVIEW OF COMPOSITE MATERIALS 1
Theodore ]. Reinhart

1.1 INTRODUCTION fraction results in increases in the levels of

mechanical properties up to the point where
Composite materials are macroscopic combi-
there is insufficient matrix material to support
nations of two or more distinct materials
having a discrete and recognizable interface the fibers and to transfer load within the com-
posite. The matrix is the adhesive binder that
separating them. Thus composites are hetero-
supports the fibers under compressive loads,
geneous materials, and many are naturally
provides shear capabilities in two dimensional
occurring; the most common is wood. The
fiber lay-ups, and transfers loads internally in
composites of interest in this discussion, how-
the composite among the myriad fibers and
ever, are the synthetic, man-made materials,
fiber bundles that comprise the load-carrying
possessing high strength and/ or stiffness rela-
portions of the composite material. In two-
tive to weight which are used in high
dimensional composites the matrix provides
performance structural applications because
the basic resistance to impact damage and
of these properties. Figure 1.1 shows the
delamination.
strength- and stiffness-to-weight relationships
for several fibers when arrayed in unidirec-
0
tionallaminates.

~j
The narrower definition of composites
becomes more specific and can be restricted to 2.5
those combinations of materials that contain
high strength/stiffness fiber reinforcements
0
supported by a high performance matrix e Corban (1MB,

material. Fibers and matrix materials may be I

organic or inorganic in chemical make up.
Normally the fibrous reinforcement material is
.51--,id i' J e
(Kovllr
e Carbon (~43'

referred to as the discontinuous phase and the e $.gIuI e ,rbon (T300{

matrix material as the continuous phase. .0

The primary engineering properties of the

eE-glia
_f.BarDn Ion ~""' C.rbon (P1201
ejiC I
composite are derived predominantly from 0.5
I
..
the mechanical and physical properties of the
I I
Alumi...m (2024-Til

discontinuous phase, the fiber reinforcement. e AIUmj (FP'

o eSIMI (mild'
These are the fiber-dominated properties of 100 150 1GO
the composite. Increasing the fiber volume Specific 1ImIi1e madUf, ~
......
Fig•.1.1 Plot of specific tensile strength to specific
Handbook of Composites. Edited by S.T. Peters. Published tensile modulus for commercially available com-
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 posites.
22 Overview of composite materials

Matrix materials may be selected from composite ~ngineering properties. At some

metallic, ceramic and organic resin materials. point in this fiber length decrease the compos-
With perhaps the exception of ceramic matrix ite properties will approach those of a system
composites, the matrix material is not as filled with particulate materials.
strong or as stiff as the discontinuous phase or Continuous fiber-reinforced composites are
fiber material. There are many exceptions or normally fabricated via a ply by ply lay-up
perhaps special cases to these statements. technique resulting in a laminar (layered) con-
Many organic resin matrix materials are mod- struction, which may be converted, before or
ified to increase toughness by the addition of after impregnation with the matrix, but prior
small amounts of rubber base or thermoplastic to consolidation and cure, to a three-dimen-
modifiers, thus forming a discontinuous phase sional composite. Figure 1.2 shows the wide
of particles within the continuous matrix range of composite constituent materials, and
phase of the composite. The classes of fiber- manufacturing and fabrication options.
reinforced composites are usually related to High performance composites offer combi-
the form of the fibrous reinforcing material nations of engineering properties which
and include continuous, long discontinuous cannot be achieved using homogeneous
and short discontinuous fiber-reinforced com- metallic alloy structural materials such as alu-
posites. Complicating the straightforward minum, titanium and steel, and may provide a
classification are the forms of the reinforce- higher level of these properties. The largest
ment, such as woven, unwoven, braided, volume usage of structural composites in
knitted and orthogonal arrays of reinforce- indus,trial, aerospace, commercial and military
ments. All of the above are possible and each applications is based upon resin matrix fiber-
provides a unique set of engineering proper- reWorced composites. Thus these composites
ties. The available levels of engineering will be emphasized throughout this volume.
properties are governed by fiber length, fiber High performance composite materials were
orientation and fiber volume fraction. developed by the US Air Force Materials
Continuous and long discontinuous fiber-rein- Laboratory in a systematic search for structural
forced composites offer maximum levels of materials that could offer weight savings over
translation of fiber properties into useful com- the conventional materials. This large and
posite engineering properties. Decreases in comprehensive R&D program to develop com-
fiber length are accompanied by a decline in posites was initiated in response to a general

RESIN REINFORCEMENT FIBER PROCESS

TYPE TYPE PLACEMENT

,.,.
,. LONG CONTINUOUS

THERMOSET .~=
,. __
--
,.,."

,'",'" -
"', , ,
-SHORT ALIGNED'"
...

,
, : CHOPPED ~RANDOM"':=:::;;""------!od--INJECTION MOLDING
THERMOPLASTIC
PARTICULATE~ "TRANSFER MOLDING

THERMOSTAMPING

Fig. 1.2 Display of composite constituent materials and manufacturing options.

 Introduction 23

call from the using communities for structural presently decreasing, and the commercial air-
materials that could reduce weight and reduce plane producers, specifically Boeing and
corrosion and fatigue problems in aircraft com- Aerospaciale, will continue to be a market for
ponents. Fiberglass reinforced resin base advanced composites. Large scale applications
composites have been used successfully in fila- in transportation, both air and ground, and in
ment wound rocket motor cases as well as in the civil engineering infrastructure arena are
various other aircraft and missile applications. and will continue to be an unrealized goal for
Fiberglass reinforced composites in general advanced composite materials at least in the
lacked the stiffness that is desired in high per- near term.
formance structural applications in aircraft, Fibers are inherently stronger than the bulk
missiles and spacecraft. The search for higher form of a material for a number of reasons.
modulus glass fibers met with only limited suc- Two major reasons are (i) the fiber's small
cess, however the use of boron, carbon, diameter with the probability of fewer flaws
graphite and other compositions resulted in a and (ii) the opportunity to align or preferen-
technological breakthrough and provided a tially orient molecular or crystal structure. All
new class· of structural materials called structural materials fail mechanically because
IAdvanced Composites'. More recent develop- of the formation, propagation and growth of
ments have led to the commercial availability of flaws. Since composite construction provides
a wide range of high strength, high stiffness very large numbers of load-carrying fibers,
fiber materials, both organic and inorganic in catastrophic failure requires the failure of
nature, for use in high performance composite many fibers. This situation then provides
applications. many early warning signs when significant
Advanced composite materials, while basi- damage is accumulating in a fiber-reinforced
cally more expensive as a raw material, more composite structure.
expensive and complicated to design with, Fibrous materials such S-glass, (Kevlar 49)
more difficult to characterize, more expensive aramid, Spectra, boron and the many types of
to manufacture and assemble and more diffi- carbon fibers produced commercially possess
cult to inspect when compared to conventional specific properties (strength/density) and
structural metallic materials, have found sig- (modulus/ density) many times greater than
nificant commercial, recreational, and military structural alloys of aluminum, titanium or steel.
applications. However, when the fibers are combined with a
Inexpensive fiberglass reinforced plastic matrix into a near quasi-isotropic lay-up, a
composites, initially put to broad use in the highly useful engineering form of the material,
1950s, are found in a variety of industrial and the specific properties are greatly reduced but
consumer products from automotive and are still superior compared to conventional
truck applications, to corrosion resistant pip- homogeneous metallic materials. Figure 1.1
ing and construction, to large and small water plots the specific tensile strength versus the
craft. The service performance of these materi- specific tensile modulus for a number of unidi-
als in the above and many other applications rectional composites and compares them to
has been highly satisfactory. steel, aluminum and titanium materials. These
Advanced composites have found exten- are calculated values based upon literature
sive applications in many areas where cost is fiber values and 65 vol.% fiber content. It can
secondary to performance. Sporting and recre- readily be seen that these high performance
ational applications including golf club shafts, fiber materials form the basis for the advanced
fishing rods, and skis, have provided a signifi- composites technology. The wide variety of
cant market and will continue to do so for the materials that can be combined to form com-
foreseeable future. Military applications, while posites having highly acceptable levels of
24 Overview of composite materials

engineering properties can make the selection certain threshold stress levels. The phenome-
of specific materials a challenging task. non known as stress-rupture failure has been
This introduction will present the basic thoroughly investigated by researchers in the
materials, design factors, material forms and industry and government.
fabrication methods used in the manufacture The glass fiber that is the industry standard
and assembly of advanced composite compo- is E-glass, which is a calcium aluminoborosili-
nents. cate formulation having very good mechanical
and electrical characteristics at very reason-
able cost. Average mechanical property levels
1.2 GLASS FIBERS
for individual filaments are 3450 MPa (500 ksi)
The development of glass fibers on a commer- for tensile strength and 72.4 GPa (10.5 x 106 psi)
cial scale was the enabling technology that for Young's modulus. Extensive research has
allowed the glass fiber-reinforced organic been conducted to develop glass fibers pos-
resin matrix composite industry to begin. The sessing higher strength and stiffness
initial understanding of organic matrix com- characteristics.
posites was based on glass fiber reinforcement. Glass formulations producing filaments of
Glass fiber composites of all descriptions increased strength and stiffness have been
have found extensive and successful applica- found to be toxic (beryllium glasses) or very
tions including low performance non-structural high melting and difficult to handle in com-
applications as well as high performance struc- mercial scale equipment. S-glass fibers contain
tural applications. The applications range from a higher percentage of alumina compared to
the building construction trades, to auto, truck E-glass. Filament strength, modulus and melt-
and rail transportation, seagoing applications ing point are higher than E-glass. Typical
including high performance racing craft and filament strength and stiffness are close to
commercial and military aerospace. Specific 4600 MPa (670 ksi) and 85.5 GPa (12.4 x 106 psi).
applications involve, decorative panels, appli- The technology developed using glass
ances, ship and boat hulls, light aircraft and fiber-reinforced resin matrix composites
glider construction, nearly all forms of recre- resulted in many high performance applica-
ational equipment, high pressure gas tions, and was the start of the effort to develop
containers and rocket motor casings. This wide stronger and stiffer fibers to meet high perfor-
spread use of glass fiber-reinforced organic mance structural needs in commercial and
composites and their continued future growth military aerospace applications. Figure 1.3
is due to many factors, including: cost, avail- shows the installation of a large commercial
ability, handling and processing ability, useful fiberglass composite pipe.
properties and characteristics and past good Closely related to the E- and S-glasses are the
experience in service. Many improvements to high silica and quartz fibers produced commer-
glass fiber composites have been made over the cially for highly specialized applications in
years including the development of highly ben- rocket motors and electrical windows. High sil-
eficial silane coupling agents. These agents, ica fibers are produced by chemically leaching
which can be tailored for compatibility with the highly soluble components from E-glass
almost any resin matrix, provide a higher level fibers and then consolidating the fibers by a
of engineering properties and environmental heat treatment. These fibers contain upwards of
durability far exceeding earlier composites. 95% silica.
Glass fiber composites provide high levels of Quartz fibers are drawn from mineral
engineering properties, however the glass fiber quartz rods using oxy-hydrogen or laser heat-
surface is sensitive to moisture attack under ing sources. Commercially available fibers are
certain conditions of exposure and above 99.95% silica. Quartz fibers typically have
 Carbon/graphite fibers 25

Fig. 1.3 Underwater installation of a large fiberglass composite pipe.

similar mechanical properties as the E-glass organic precursor fibers such as rayon or poly-
materials but have a much higher melting point. acrylonitrile (PAN). The process, shown in
Fig. 1.4, involves highly controlled steps
including heat treatment and tension, to trans-
1.3 CARBON/GRAPHITE FIBERS form the organic precursor into a highly
The search for advanced fibers led to the ordered carbon or graphitic structure. Carbon
development of carbon and graphite fibers. and graphite fibers are also produced from
These fibers are currently the best known and pitch base precursor materials. Although the
most widely utilized in high performance potential for low cost exists, complex process-
resin base composites. Primarily developed ing steps involved in the pitch treatment have
for military aerospace applications, these prevented the production of really low cost
materials have found wide-spread commer- fibers. US manufactured pitch base carbon and
cial and industrial applications. This situation graphite fibers while having a very high mod-
has resulted in the availability of a wide vari- ulus, up to 830 GPa (120 x 106 psi), have
ety of fibers having various levels of
engineering properties at costs once thought
unachievable. These materials are now avail-
able on a world wide basis at competitive
prices and are now experiencing increasing
levels of interest for applications such as com-
mercial aerospace, ground transportation and
the infrastructure.
The production of carboni graphite fibers is
well suited to large scale continuous operation
where economies of scale operate effectively
and the stability of operating conditions pro-
vides a narrow band of fiber mechanical
properties. Carbon fibers are produced com- Fig. 1.4 Conceptualized production process for car-
mercially by the thermal decomposition of bon fiber manufacture.
26 Overview of composite materials

demonstrated low tensile properties and thus applications in high performance structural
have not been seriously considered for high applications.
performance structural applications. The high tensile strength and low density of
Advances in PAN precursor carbon fiber aramid fibers allowed applications in filament
technology have resulted in cost reductions wound pressure vessels and rocket motor
and properties increases which have improved cases replacing S-glass. Combined with its
the status of these materials in high perfor- lower density and high levels of properties,
mance structural applications. PAN base very lightweight structures resulted. Aramid
carbon fibers are now commercially available fibers are highly damage- and cut-resistant
having modulus values of 280-450 GPa and provide high levels of damage tolerance
(40-65 x 1()6 psi) and tensile strengths of to composite components. Other applications
4140-5170 MPa (600-750 ksi) and some versions involved aramid usage in lightly loaded sec-
up to 6890 MPa (1000 ksi) in tensile strength. ondary structures on commercial aircraft and
High fiber elongation has long been considered helicopters, reducing structural weight com-
a very desirable trait. Elongations of 2% and pared to glass composites. Issues in the use of
above are highly desirable and provide tough- these fibers have involved moisture absorp-
ness in the composite that low elongation tion by the fiber, relatively low adhesion to
fibers «1.8%) cannot provide. Since high most resin matrix materials, low compression
strain fibers must be coupled with a tough properties and difficult machining compared
matrix resin the availability of these fibers has to other composites.
stimulated the introduction of new formula-
tions of high elongation resin matrix materials.
1.4.2 OTHER ORGANIC FIBERS
In many instances in aerospace applications,
compression is the design controlling factor. Continuing research into the various other
The new high strength, high modulus carbon rigid rod molecule fiber forming polymers
fibers have smaller diameters thus requiring including the spinning of polyethylene fibers
higher levels of support from the resin under by Allied Chemical Co. and others has
compression loading. Thus, the limiting factor resulted in high performance organic fibers
becomes resin shear modulus and the newer that are finding specialized and general com-
fibers have not resulted in increased levels of mercial applications. Spectra fiber, a high
compression strength in the composite. performance product spun from polyethylene
offers high abrasion resistance, high strength,
and high modulus at a very low density. This
1.4 ORGANIC FIBERS
material has found use in combat equipment
for both ground troops and aviators.
1.4.1 ARAMID
Recreational uses of this material include ship
Aramid is a generic term for the class of aro- sails and lines. Applications of Spectra are lim-
matic-polyamide fibers produced using ited to temperatures of about 120°C (250°F)
para-phenylene terephthalamide. These mate- due to its low crystallite melting point.
rials were introduced by E.I. DuPont de
Nemours & Co., Inc. in the 196Os. Competing
1.4.3 PBO AND PBr FffiERS
fibers are now produced in Europe and Asia.
Although organic polymer fibers can be pro- High modulus high strength organic fibers
duced that have fiber tensile strength and have been prepared by the Air Force Materials
modulus values approaching the mid-range Laboratory from the rigid rod molecules of
carbon fibers, low fiber compression capabili- poly (p-phenylene benzobisoxazole) (PBO)
ties have limited organic fiber composites and poly (p-phenylene benzobisthiazole)
 Other inorganic fibers 27

(PBT). These rigid rod polymers yield high technology, carbon-epoxy composites still
modulus fibers due to their high degree of cannot compare in compression properties to
aromaticity, high planarity, and linearity in those of boron-epoxy.
the chain. These polymers in general are diffi- Boron-epoxy composites have been used in
cult to process into fiber due to their the sporting goods industry and boron fibers
infusability and insolubility in anything have been used in metal matrix composites
except the most powerful organic solvents (MMC) because of their excellent mechanical
such as methanesulfonic acid or concentrated properties, thermal stability, and reduced reac-
phosphoric acid. tivity with the matrix (compared to carbon
The high axial molecular orientation which fibers). In order to prevent reaction between
leads to extreme anisotropy of mechanical the boron fiber and the molten metal matrix
properties is a distinctive feature of rigid rod during MMC production, boron fibers were,
molecule fibers. While tensile strength and until recently, available with silicon carbide or
modulus may in some instances be competitive boron carbide coatings that acted as diffusion
with carbon fibers the transverse properties of barriers. Boron fibers with these coatings are
the fiber are less than 20% of the axial proper- no longer commercially available because sili-
ties. In tension the axial characteristics con carbide fibers are now replacing boron in
dominate and excellent properties are achieved. metal matrix composites.
In compression and flexure the low transverse Boron fibers are produced as a rather large
properties result in low mechanical properties monofilament fiber or 'wire' (100-200 J.1ffi or
in composites of PBO and PBT fibers. 4-8 mils) by chemical vapor deposition (CVD)
High modulus organic fibers have not of boron onto a tungsten or pyrolyzed carbon
found serious applications in structural com- substrate. The resulting fibers have excellent
posites due to their poor compressive strength (3450 MPa or 500 ksi) and stiffness
performance and relatively high cost. (400 CPa or 58 x 106 psi». However, because
both the precursor gases and the manufactur-
ing process are inherently expensive, boron
1.5 BORON FIBERS fibers cannot compete with carbon fibers on
Boron fiber in organic resin matrices was ini- the basis of cost.
tially developed by the Air Force Materials
Laboratory in the early 1960s and became the
1.6 OTHER INORGANIC FIBERS
first high performance reinforcement for use
in structural composite materials. Early appli- Continuous filament forms of other inorganic
cation of these fibers was found on the US Air fibers are also now commercially available.
Force F-15 and the US Navy F-14 fighter air- Silicon carbide (SiC) fiber is produced from
craft. Boron composites in honeycomb CVD of SiC by a method similar to that used in
structure were utilized in the empennage boron fiber production. SiC fibers are also pro-
structure of these aircraft. Subsequent studies duced from an organic borosilane precursor
have found that carbon-epoxy composites that is spun and pyrolyzed to produce silicon
could be utilized in these aircraft to reduce carbide. Aluminum oxide fibers are produced
costs of the composite structure. Considering in both the mono and polycrystalline forms.
the enormous changeover costs involved in Advanced forms of ceramic fibers are now
such a project and the limited production under intensive development for use in metal
quantities of these aircraft remaining to be matrix composites for high temperature tur-
produced it has been determined that return bine engine applications.
on investment costs would not be achieved. In
spite of the advances made in carbon fiber
28 Overview of composite materials

1.7 MATRICES and bodies, appliances, bathtubs and showers

and automobile hoods, decks, and doors.
As stated above, the purpose of the composite The very large number of resin formula-
matrix is to bind the fibers together by virtue of tions, curing agents, fillers, and other
its cohesive and adhesive characteristics, to components provide a tremendous range of
transfer load to and between fibers, and to pro- possible properties. The resulting resin systems
tect them from environments and handling. include such types as general purpose, chemi-
In continuous multifilament composites the cally resistant and heat resistant, with
fiber dominated properties form the basis for general-purpose polyester and vinyl ester resin
useful engineering properties of the material. composites being the most widely used by far.
In such composites the matrix material is sub- The development of highly effective silane
jected to high levels of strain magnification coupling agents for glass fibers allowed the
caused by the close proximity of the fiber and fabrication of glass fiber reinforced polyester
fiber bundles and becomes subject to failure and vinyl ester composites having excellent
by cracking. mechanical properties and acceptable
Thus the matrix is the 'weak link' in the environmental durability. These enhanced
composite, especially because resins do not characteristics have been the major factors in
presently exist that allow utilization of the the widespread use of these composites today.
stresses that the fibers are able to withstand. The problems of attaining adequate adhe-
Thus, when the composite is under load, sion to carbon and aramid fibers have
resins may microcrack and craze, form larger discouraged the development of applications
cracks through coalescence of microcracks, for polyester or vinyl ester composites that use
debond from the fiber surface, and generally these fibers. Although there are applications of
break down at composite strains far lower high performance fiberglass composites in
than desired. military and aerospace structures, the rela-
Thermoplastic and toughened thermoset- tively poor properties of advanced composites
ting matrices have been developed that allow of polyester and vinyl ester resins when used
increased composite strain levels but still do with other fibers, combined with the compar-
not provide for full fiber property translation. atively large cure shrinkage of these resins,
Nevertheless, the matrix resin provides have generally restricted such composites to
many essential functions; in addition to those lower-performance applications.
mentioned above, the matrix keeps the rein-
forcing fibers in the proper orientation and 1.8 EPOXY RESINS
position so that they can carry the intended
loads, distributes the loads more or less Epoxy resin R&D has provided a rich and high
evenly among the fibers, provides resistance performing family of resin materials that have
to crack propagation and damage, and pro- found extensive commercial and aerospace
vides all of the interlaminar shear strength of applications. Epoxy resins have demonstrated
the composite. Furthermore, the matrix gener- a tendency to absorb moisture both in the
ally determines the overall service green or uncured and the cured stages. This
temperature limitations of the composite, and property poses engineering limitations and
may also control its environmental resistance. resulted in the development of the cyanate,
Polyester and vinyl ester resins are the most bismaleimide, and thermoplastic materials to
widely used of all matrix materials. They are be covered later in this introduction.
utilized mainly in commercial, industrial, and Although epoxies are sensitive to moisture
transportation applications, including chemi- they are generally superior to polyesters in
cally resistant piping and reactors, truck cabs resisting other environmental influences, and
 Cyanate resins 29

offer better mechanical properties. Even though No degradation of the composite materials
the elongation to failure of most cured epoxies has been observed even after over 20 years of
is relatively low, epoxies provide an attractive service exposure. The moisture absorption of
combination of handling characteristics, pro- epoxy resins causes a decrease in the glass tran-
cessing flexibility, composite mechanical sition temperature (Tg) of the material and
properties, ease of manufacture, and acceptable limits (for conservative engineering practice)
cost. Recent developments include modified epoxy resins to use at temperatures not exceed-
epoxy resin formulations that have improved ing the wet Tg of the cured material. While this
elongation capabilities. In addition, a substan- use limitation is conservative, its imposition has
tial data base exists for epoxy resins since both generally avoided unwanted service experience
the US Air Force and the US Navy have been problems. Figure 1.7 shows the range of com-
flying aircraft with epoxy matrix structural posite applications on the C-5 cargo aircraft.
components since 1972, and the in-service expe-
rience with these components has been very
1.9 CYANATE RESINS
satisfactory. Figures. 1.5 and 1.6 show epoxy
composite applications on the AV-SB horizontal The more recent commercial availability of
stabilizer and forward fuselage. cyanate resins from a number of commercial

_ G' phlt l' f.JlO~V Til'"

G, phlll' EIl<l~1' Cloth
Alu I Inurn
- Tit. um

TO ROUE BOX UPPER SKIN \

c ]__....[~,[_.[. . ___] ~[
ALUMINUM LE

J TE STRUCTURI;

TOROUE BOX LOWER SKIN WITH INTEGRAL SPARS

Fig. 1.5 Application of epoxy composites to the horizontal stabilizer of the AV-8B aircraft.
30 Overview of composite materials

C:=J Graphite{epoxy

~Aluminum

C]Titanium

•
61 /:-.
~CLOSURE
WHEEL WELL

RIGHT
SHELL
MOLOING
LEFT SHELL
MHOING
FWO PRESSURE
BULKHEAO

~ EQUIP ACCESS
FO!MARO FRAME ~PANEL
LOWER SKIN PANEL

Fig. 1.6 Application of epoxy composites to the forward fuselage of the AV-8B aircraft.

sources has provided a matrix material gener- epoxies in maximum hot/wet use tempera-
ally tougher and far more moisture resistant ture, extending the safe in-service temperature
compared to epoxies. Thus they possess better to 205-220°C (400-430°F) or higher. They are
electrical characteristics and are not subjected to available from a number of suppliers.
the so called hot/wet T service temperature The BM! materials have been selected for
reduction. The lack of a d~ta base and manufac- structural applications on the US Air Force
turing experience base tend to inhibit the F-22 fighter in combination with carbon fiber
introduction of the cyanate materials into new reinforcements.
systems.
1.11 OTHER THERMOSETTING RESINS
1.10 BISMALEIMIDE RESINS
Polyimide resins including the PMR
BMIs have a very competitive set of features (Polymerization of Monomeric Reactants) and
including; toughness, moisture resistance, acetylene terminated versions are available
ease of handling and manufacture and good and possess service temperature capabilities,
engineering properties. They are superior to in air, close to 260°C (500°F). Applications in
 Thermoplastic resins 31

Nonstructural parts,
liners, troop seats

CarboniEpoxy

CarbonlKevlar/ Epoxy

_ GFRP/Nomex core

_ Kevlar/foam core

KevlarINomex

D CarboniNomex

Fig. 1.7 Composite applications on the C-5 cargo aircraft; total weight of composites: 7167 kg (15 7871b);
weight saving: 1997 kg (43981b).

military turbine engine casings (liners) and sulfide (PPS), polyether ketone ketone (PEKK)
electronics form the major present day appli- and several others. Most thermoplastic matri-
cations for these materials. Lack of toughness ces do not absorb any significant amount of
and difficulty in handling and in manufacture water, but organic solvent resistance is an area
characterize this family of heat resistant of concern for the noncrystalline thermoplas-
organic matrix materials. tics. Because thermoplastic matrices have an
unlimited shelf life before molding (unlike
thermosetting matrices), and because they can
1.12 THERMOPLASTIC RESINS
potentially be remolded by the application of
The families of amorphous and semicrys- heat and pressure, thermoplastic matrix com-
talline thermoplastic materials have been posites also offer the possibility of lower-cost
thoroughly evaluated for use as matrix resins. fabrication. However, cross-linking or thermal
Thermoplastic matrix materials are generally degradation with repeated temperature
tougher than most thermoset resins and offer cycling is a concern with most of the systems.
the potential of improved hot/wet resistance. The fabrication procedures necessary for the
Because of their high strains to failure, they low-cost manufacture of thermoplastic matrix
also are the only matrices currently available composites have been extensively studied,
that allow thermoforming and other forms of including effort to determine and understand
rapid manufacture to be utilized. the mechanical properties obtained from the
The thermoplastic resin materials include various techniques currently available for fab-
polyether ether ketone (PEEK), polyphenylene ricating thermoplastic matrix composites.
32 Overview of composite materials

Thermoplastic composites are deemed to be craft (Fig. 1.8) that successfully circumnavi-
a mature technology and will compete with gated the earth non-stop.
other composites on a properties and cost
basis.
1.14 MATERIAL FORMS
Reinforcing fibers are commercially available
1.13 DESIGN PHILOSOPHY
in many product forms including monofila-
The design of a high performance composite ments for fibers such as boron on silicon
structural component should be based upon a carbide to multifilament fiber bundles.
thorough understanding of the service condi- Product forms include unidirectional ribbons
tions to be encountered. The 'building block' and broadgoods up to 152 cm (60 ins) wide to
approach has been found successful in the single and multiple layer fabrics and mats that
design of both aircraft and spacecraft structural mayor may not be stitched in the Z direction.
components. A comprehensive properties data These same forms are available combined with
base is required for this process. The database the matrix resin of choice for a particular com-
is obtained by a thorough program of coupon, ponent. Using prepregs rather than in-line
subelement and element fabrication test and impregnation of the fibers during the final
failure analysis. The building block program composite fabrication process can offer signif-
must account for properties in the as-manufac- icant advantages. Prepregs can have very
tured article including variables induced by precisely controlled fiber/resin ratios, highly
worker skill levels, tooling and suppliers, etc. controlled tack and drape (in the case of ther-
An example of combining these diverse build- moset matrices), controlled resin flow during
ing blocks was the rapid design and the cure process, and, in some processes, better
manufacture of the all composite Voyager air- control of fiber angle and placement. Prepreg

Fig. 1.8 All composite Voyager aircraft.

 Further reading 33

materials can be produced and stored for process. High prepreg lay-down rates and
future use, normally under refrigeration for fairly low scrap rates are possible. However,
thermosetting matrices, and then used in most of the existing automatic lay-up
processes ranging from hand lay-up to highly processes are quite limited in their ability to
automated filament winding or machine tape produce satisfactory components having com-
laying. Processes such as pultrusion and plex curvatures. This is more an artifact of the
braiding can also use prepreg forms instead of tape raw material than it is an equipment lim-
in-line resin impregnation. While the latter itation.
may be lower in initial cost, it may be prohibi- To select the best composite fabrication
tive for some resin systems (such as process, the designer generally chooses the
thermoplastics), and parameters such as process that will provide an acceptable quality
fiber/resin ratio may not be as easily con- component for the lowest cost. In evaluating
trolled as is the case with a prepreg. cost and quality, however, tooling cost, pro-
duction rate, materials cost, desired part finish
and many other factors must be considered.
1.15 MANUFACTURINGIFABRICATION
Only after all the relevant factors have been
The most common composite lay-up processes weighed can the fabrication method (or the
include: hand lay-up, fiber placement and material) be selected.
machine tape lay-up. In lay-up, material that is This overview chapter was intended as a
usually in prepreg form is cut and laid up, primer to introduce the reader to composite
layer by layer, to produce a laminate of the technology. Fiber and matrix materials, design
desired thickness, number of plies, and ply concerns, material forms and fabrication
orientations. In filament winding, a fiber bun- processes have been described briefly. For
dle or ribbon is impregnated with resin and more specific and detailed data and guidelines
wound upon a mandrel to produce a simple on these topics, the reader should consult the
shape, such as a tube or pressure vessel, or a technical sections in this volume. However,
more complex shape, such as a wind turbine this introduction should have clarified that
or helicopter rotor blade. composites offer tremendous opportunities -
Hand lay-up processes are usually accom- as well as impressive challenges - for the
panied by a compaction step for each one or designer and manufacturer of high perfor-
more plies that are applied. Lay-up of thermo- mance components.
plastic matrix material is always accompanied
by heat and pressure.
FURTHER READING
For most components produced for aero-
space applications, except filament wound 1. Weeton, J.W. and Peters, D.M., eds, Engineers'
components, the autoclave is the curing tool of Guide to Composite Materials, American Society
choice. The autoclave cure process is chosen for Materials, 1987.
2. SPI Composites Institute, Introduction to
for many reasons, these include: reproducible, Composites, SPI, Washington, 1992.
high quality, low void content, parts at reason- 3. Beukers, A. and de Jon, Th., Materials Delft
able production rates and costs. There are also University Press, Delft, Netherlands, 1992.
a number of other fabrication processes that 4. Corish, P.J., ed, Concise Encyclopedia of Polymer
are used including pultrusion, resin transfer Processing and Applications, Pergamon Press,
molding and press curing. These processes Oxford, 1992.
5. Schwartz, M., ed Composite Materials Handbook
have found application in the fabrication of
2nd edn, McGraw-Hill, New York, 1992.
various components for aerospace usage.
Very sophisticated, numerically controlled
equipment is presently available for the lay-up
POLYESTER AND VINYL ESTER RESINS 2
Frank A. Cassis and Robert C. Talbot

2.1 INTRODUCTION AND HISTORY sored production of styrene-butadiene rubber.

At the same time, scientists found that styre-
Organic polymers are divided into two types,
nated polyesters could yield high strength,
reinforced-thermoplastic and thermoset. With
light weight structures when reinforced with
thermoset polymers such as unsaturated poly-
gl~ss fibers. They also learned that fiberglass-
esters and vinyl esters, a chemical reaction
reinforced polyesters had excellent electrical
cross links the material so that it cannot be
properties and that large structures could be
returned to liquid form. Other common ther-
molded at low pressures with low cost tooling.
mosetting polymers include epoxy and
As a result, commercial development pro-
phenolic resins. Thermoset plastics made with
ceeded rapidly after World War II with
polyester and vinyl ester resins represent the
materials and molding research moving in
major portion of the reinforced plastic com-
many directions. In the 20 years that followed,
posites industry today.
polyester and vinyl ester resins matured
Early workers on unsaturated polyesters
rapidly and by the mid-1970s, the composites
soon learned that despite the possession of
fabricator and end user had numerous options
reactive double bonds, these resins were slug-
with these matrix systems to achieve the
gish in reacting with themselves. Even with
desired properties in the finished part.
effective catalysts, they still required high tem-
peratures and lengthy cure times to complete
the cross linking reaction. The key to modern 2.2 POLYESTER RESINS
day application of unsaturated polyesters was
The reaction of an organic acid with an alcohol
the discovery by Carlton Ellis in 19371 that the
addition of reactive monomers, such as styrene, results in the formation of an ester. By using a
difunctional acid and a difunctional alcohol
~ave mixtures that would copolymerize many
(glycol) a linear polyester is produced (Fig. 2.1).
times faster than homopolymerization. The
styrene addition produced the added benefit of
an easily handled liquid material that could be o 0
pumped, transported and fabricated into a fin- II II
H-(-O - C - R - C - 0 - R -)n -OH
ished plastic by a myriad of molding processes.
Developments during the 1940s accelerated
the commercial applicability of unsaturated
polyesters to the position they hold today. Fig. 2.1
Styrene became readily available and lower in
cost as a result of the US Government's spon- P~operties of polyesters can be varied by using
different combinations of diacids and glycols.
Handbook of Composites. Edited by S.T. Peters. Published These products are thermoplastic polyesters
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 and they are made with various acids and
 Polyester resins 35

glycols such as the following: In the esterification reaction with maleic anhy-
dride, the unsaturated acid isomerizes to the
Acids Glycols
fumarate structure which copolymerizes with
Phthalic anhydride Ethylene glycol
styrene much faster than the maleate form. A
Isophthalic acid Propylene glycol
high degree of isomerization to the fttn1arate
Terephthalic acid Neopentyl glycol
structure is essential to produce an unsaturated
Adipic acid Diethylene glycol
polyester with high reactivity. Although the iso-
The reaction product of terephthalic acid and merization of maleic anhydride is usually from
ethylene glycol is the well known polyethyl- 65-95% in the esterification reaction, some com-
ene terephthalate (pEn which is used to make mercial resins are deliberately formulated with
polyester fibers and polyester plastics such as the more expensive fumaric acid to obtain max-
clear plastic bottles for soft drinks. imum reactivity with the monomer employed.
Unsaturated polyesters are produced by
replacing part of the saturated diacid with an
2.2.1 UNSATURATED POLYESTER CLASSES
unsaturated diacid such as maleic anhydride
or fumaric acid (Fig. 2.2). The former is vastly Unsaturated polyesters are divided into types
preferred since it is lower in cost, easily han- or classes depending on the structure of the
dled and produces only half the water that basic building block. These are orthophthalic,
would be generated in the reaction when isophthalic, terephthalic, bisphenol-fumarate,
fumaric acid is used. chIorendic and dicyclopentadiene.

Orthophthalic resins
CH=CH
I ,
O=C-O-C=O These are commonly referred to as ortho or
general purpose resins and are usually based
Maleic
anhydride
on phthalic anhydride, maleic anhydride and
propylene glycol. Since the acid groups in
H H phthalic anhydride are on adjacent carbons of
II II the benzene ring, it is very difficult to produce
HOOC- C = C - COOH
resin molecular weights as high as those
Fumaric acid
achievable with isophthalic and terephthalic
acid. Accordingly, resins made from phthalic
anhydride have poorer thermal stability and
Fig. 2.2 chemical resistance than their iso/tere coun-
The resultant polyester contains reactive dou- terparts.
ble bonds (unsaturation) along the entire
polyester chain, which becomes the site for the Isophthalic resins
eventual cross linking to produce the cured
plastic (Fig. 2.3). These resins are produced from isophthalic
acid and are characterized by greater strength,
heat resistance, toughness and flexibility than
o 0 0 HO their ortho cousins. In isophthalic acid, the
II II II I II
acid groups are separated by one carbon of the
HO (C-R-C-O-R'-O-C-C=C-C-O-R'-O }nH
I benzene ring which increases the opportunity
H to produce polymers with greater linearity
and higher molecular weight in the esterifica-
Fig. 2.3 tion reaction (Fig. 2.4).
36 Polyester and viny ester resins

ester can be three times longer than for an iso

~-C'~o ~ resin. As a result of this, researchers have

turned to polyethylene terephthalate scrap

V<
Phthalic
~
Isophthalic
from the previously mentioned fiber and plas-
tic operations to develop an economical source
of terephthalic polyesters. TIUs scrap can be
anhydride acid effectively depolymerized by using different

(oH amounts of propylene glycol at elevated tem-

o
peratures. The glycolyzed product is then
reacted with maleic anhydride and diluted
with styrene monomer to produce a cost effec-
tive terephthalic polyester.
c=o
'OH
Bisphenol A fumarate resins
Terephthalic
acid These resins are unsaturated rigid polyesters
made by reacting bisphenol A with propylene
Fig. 2.4 oxide to produce the glycol shown in Fig. 2.5.
Therephthalic resins This propoxylated bisphenol A is then
reacted with fumaric acid to form an unsatu-
Unsaturated polyesters can be produced from rated polyester. The bisphenol structure
terephthalic acid with the expectation that the illustrated above imparts a high degree of
resin property improvement obtained in going hardness, rigidity and thermal stability to this
from phthalic anhydride to isophthalic acid particular resin.
will be matched in going from isophthalic to
terephthalic acid. This, however, is not the case
and terephthalic resins appear to offer only a Chlorendic resins
slight advantage in heat distortion tempera- These unique polyester resins are based on
ture over their isophthalic counterparts. Other HET acid (hexachlorocyc1opentadiene) or the
important resin properties such as modulus, anhydride shown in Fig. 2.6.
hardness and overall chemical resistance favor When reacted with an unsaturated acid and
the iso resins. a stable glycol such as neopentyl, an extremely
Because of its lower solubility and poorer rigid unsaturated polyester results with out-
reactivity, therephthalic acid requires the use standing thermal stability and resistance to
of esterification catalysts or pressure process- oxidizing environments. The inherent chlorine
ing to produce a resin economically. Without in the resin chain imparts some fire retardancy
these, processing time for a terephthalic poly- as well.

CI~~O
CI-C-CI 'OH
CI
H-c=o
,
OH

Fig. 2.5 Fig. 2.6

 Polyester resins 37

o 0
HC-COOH II II
II + HO-C c-o
HC-COOH I I
HC=CH

Maleic acid DCPD Half-ester

Fig. 2.7

Dicyc10pentadiene resins viscosity change until gellation occurs.

Furthermore, this gel time can be adjusted for
The basic building block for this class of resins
process and other fabrication conditions by
results from the addition of maleic acid to one
balancing the effects of inhibitor, promoter
of the double bonds of dicyc10pentadiene
and catalyst.
(OCPD) are shown in Fig. 2.7 above.
The additional 'bonus' offered by unsatu-
The OCPD maleate half ester can then be
rated polyesters is embodied in the simplicity
esterified with various glycols such as propy-
of modifying mechanical and environmental
lene, neopentyl and diethylene glycol to
properties to fit the requirements of the fabri-
produce an interesting series of resins. DCPD
cation process or the end use intended. For
resins are very useful in general purpose
example, simply by changing the glycol used
applications because of the low cost of DCPD.
from propylene to diethylene one can
In addition, the alicyclic rings raise the heat
markedly increase the elasticity (flexibility) of
distortion temperature of the resin and
the resultant resin. By changing the saturated
markedly enhance thermal and ultraviolet
diacid from phthalic to isophthalic, one can
resistance.
markedly increase heat and chemical resis-
tance of the resultant resin. By simply
2.2.2 PROPERTIES/APPLICATIONS changing the ratio of unsaturated acid to satu-
rated acid one can markedly influence the heat
Unsaturated polyester resins are very versatile
distortion temperature and stiffness of the
materials. At ambient temperatures, the liqUid
resultant resin. Obviously then the chemistry
polymers which are available at varying vis-
of unsaturated polyesters offers almost end-
cosities are stable for many months. In spite of
less possibilities for property changes which
this great stability, they can be cured in a mat-
enables one to 'tailor' the final resin to the end
ter of minutes by adding a peroxide catalyst.
use application by varying the building blocks
This curing takes place by a cross linking reac-
or their percentage in the resin formulation:
tion between the reactive diluent (most often
styrene) and the double bonds in the polyester • saturated acids - phthalic, isophthalic,
chain to form a strong three-dimensional terephthalic, adipic, het acid;
nework. A normal cure will most often convert • unsaturated acids - maleic, fumaric;
90% of the reactive double bonds into single • glycols - ethylene, propylene, diethylene,
bonds in the polymer network. Many ther- dipropylene, neopentyl, bisphenol A,
mosetting polymers such as epoxies and trimethyl pentanediol;
phenolics begin to increase in viscosity when • hydrocarbons - dicyclopentadiene;
catalyzed and continue to increase until they • reactive monomers - styrene, vinyl toluene,
are cured. Polyesters, on the other hand, pro- p-methyl styrene, t-butyl styrene,
vide a specific working or 'gel' time without chlorostyrene, a-methyl styrene.
38 Polyester and viny ester resins

In addition to tailoring the resin for specific cations which do not involve reinforcing mate-
applications by varying the building blocks, rials. These so-called casting applications
the properties of unsaturated polyesters can include buttons, bowling balls, putties, cul-
often be altered significantly by selection of tured marble, gel coats and decorative
the esterification process. This is particularly products. The marble industry and the more
true with isophthalic/terephthalic polyesters recently developed polymer concrete industry
which are slower reacting than phthalic anhy- represent outstanding applications for highly
dride. By using a two stage or modified two filled unsaturated polyesters which offer very
stage reaction with these aromatic diacids, the economical materials to the building and con-
molecular structure of the resultant polyester struction industry. Fiberglass reinforced
can be changed to markedly improve heat dis- polyesters (FRP) are used in the manufacturing
tortion temperature, hydrolytic stability and of boats, automobile and truck parts, building
chemical resistance2. In the two stage process panels, corrosion resistant equipment such as
the aromatic acid and glycol are fully or par- pipes, tanks, ducts, scrubbers, etc., appliances
tially reacted before the faster reacting and business equipment, electrical equipment,
unsaturated acid is added to the cook. This construction products such as grating and rail-
processing method, compared to charging all ing, sporting equipment and consumer
ingredients at once (one stage method), also products that are almost endless. According to
leads to a more random distribution of the the Composites Institute of the Society of
unsaturation in the polymer chain which Plastics Industry (SPI), automotive, construc-
changes the character of the final cross linked tion, marine and corrosion resistant
network in the cured resin. equipment are the four largest FRP markets, in
Cure plays one of the most important roles that order, in the United States which pro-
in the chemical resistance developed by unsat- duces 2.5 billion pounds of FRP annually.
urated polyester resins. Theoretically, the Mechanical properties are most often the
curing reaction should go to completion at critical factor in the selection of a polyester resin
room temperature with all the double bonds for a specific application. Testing of mechanical
converted to single bonds in the three-dimen- properties for both resin castings and fiberglass
sional network. However, complete cross reinforced composites is carried out using stan-
linking is rarely achieved at ambient tempera- dardized ASTM (American Society for Testing
tures. This then will result in reduced chemical and Materials) tests for all plastics.
resistance and, quite often, poorer than
ASTM 0-638 Standard Test Method for
expected mechanical properties. In addition,
Tensile Properties of Plastics
unreacted diluent (styrene) can remain in the
ASTM 0-790 Standard Test Method for
not-so-well cured polymer leading to major
Flexural Properties of Plastics
problems when the polyester is used for food
ASTM 0-695 Standard Test Method for
grade applications. Accordingly, maximum
Compressive Properties of
chemical resistance and certain other property
Rigid Plastics
improvements can most often be achieved by
ASTM 0-256 Standard Test Method for
utilizing elevated temperatures for 'post cure'
Impact Strength (IZOO) of
of the polyester resin finished product.
Plastics
Unsaturated polyester resins are used in the
ASTM 0-648 Standard Test Method for Heat
manufacture of a broad range of plastic prod-
Oistortion Temperature of
ucts. A high percentage of these products
Plastics
utilize reinforcing materials, particularly fiber-
ASTM 0-2583 Standard Test Method for
glass. It is estimated that less than 20% of the
Barcol Hardness of Plastics
polyester resins produced are utilized in appli-
 Polyester resins 39

As mentioned earlier, glycol selection has a produce a rigid polyester which tends to be
significant effect on the properties of poly- hard, brittle and lower in tensile elongation.
esters. Ether glycols are of great value in Higher unsaturation also leads to higher heat
increasing tensile elongation and impact distortion temperature resins. The latter is also
strength which is of great importance in auto- achieved by formulating higher molecular
motive, casting and liner applications. A weight resins with the chlorendic, bisphenol A
principal deficiency of polyester resins is lack and dicyclopentadiene building blocks. As
of alkali resistance because the ester linkages expected, all of these resin classes are more
are subject to hydrolysis in the presence of brittle and have low tensile elongation. The
caustics. Accordingly, increasing the size of the major exception in this scenario are the
glycol has the same effect as reducing the con- iso/terepolyesters. Using the multi-stage pro-
centration of attackable ester linkages. Thus, a cessing methods described earlier, these resins
resin containing neopentyl glycol, propxylated can be formulated with reasonably high mole-
bisphenol A, or trimethyl pentanediol will cular weights (more linearity) to give very
exhibit improved water and chemical resis- tough resins having a good balance of ten-
tance which is highly important in gel coats, sile/flexural properties plus higher tensile
corrosion resistant equipment, construction elongation and heat distortion temperatures.
products and many consumer products. Obviously then, when the end use criteria
The major effect on polyester physical prop- requires the 'something more' than is offered
erties is, however, provided by the by general purpose polyesters (orthophthalics
unsaturation content in the polyester polymer. and dicyclopentadienes), the formulator turns
Higher unsaturation makes for more cross to iso/terepolyesters which have no disadvan-
linking and a stiffer cured composite. tages compared to general purpose resins
Accordingly, the formulators' selection of other than slightly higher cost.
unsaturated acid to saturated acid ratio which Table 2.1 summarizes the property and
determines cross linking density can move the application status for the various classes of
resin flexural modulus from rigid to resilient unsaturated polyesters.
to very flexible. In most cases, a 1/1 ratio will
Table 2.1 Properties and applications of unsaturated polyesters

Class Characteristics Uses

Orthophthalics, Rigid, resistant to Boats, tub/shower, spas, marble,
dicyclopentadiene crazing, light in color consumer products, buttons, corrugated
sheet, building panels, seating,
decorative products
Isophthalics / terephthalics Tough, good impact and Automotive parts, gel coats,
overall mechanical properties, electrical, bowling balls, trays,
resistant to environmental gasoline, tanks, septic tanks, swimming
elements and moderate pools, tooling, aerospace products,
chemical attack. Hlighly corrosion, construction products
resistant to aromatics
Chlorendic Rigid, high heat distortion, Corrosion resistant tanks,
highly resistant to oxidizing ducting, stacks, industrial vessels
chemical environments
Bisphenol A fumarates Rigid, high heat distortion, Corrosion resistant tanks,
highly resistant to most piping, stacks, industrial vessels
chemical environments
particularly caustics
40 Polyester and viny ester resins

2.3 VINYL ESTER RESINS Although vinyl esters have often been clas-
Vinyl ester resins are the most recent addition sified as polyesters, they should be designated
to the family of thermosetting polymers. separately because they are typically diesters
Although several types of these resins were with a recurring ether linkage provided by the
synthesized in small quantities during the late epoxy resin backbone.
1950s, it was not until the mid-1960s that com-
mercialization, principally by Shell and Dow 2.3.1 VINYL ESTER RESIN TYPES
Chemical led the push to establish an
extremely important segment of today's com- Aside from the fire retardant versions of vinyl
posite industry. Vinyl esters are unsaturated ester resins which are discussed in the next
resins made from the reaction of unsaturated section, there are two basic types of vinyl
carboxylic acids (principally methacrylic acid) esters having commercial significance. These
with an epoxy such as a bisphenol A epoxy are the general purpose lower molecular
resin. The typical structure of a vinyl ester weight vinyl esters and the higher heat resis-
resin is shown in Fig. 2.8. tant vinyl esters with greater cross link
The structure of vinyl ester resins shows density.
several important features which account for
the resultant exceptional properties of vinyl General purpose vinyl esters
ester resins. There is an epoxy resin backbone
with a high molecular weight that provides These are principally methacrylated epoxies
excellent mechanical properties combined made by the reaction of methacrylic acid with
with toughness and resilience. Secondly, vinyl a bisphenol A epoxy resin. When dissolved in
esters display terminal unsaturation which styrene monomer they provide a thermosetting
makes them very reactive. They can be dis- resin with good heat resistance, excellent
solved in styrene and cured like a mechanical properties (particularly high ten-
conventional unsaturated polyester to give sile elongation) and outstanding chemical
rapid green strength. Obviously, the vinyl resistance to acids, bases, hypochlorites and
ester structure also enables convenient many solvents.
homopolymerization which could lead to high
heat distortion products. Finally, vinyl esters
Heat resistant vinyl esters
have much fewer ester linkages per molecular
weight which combined with the acid resistant These vinyl esters have higher density cross
epoxy backbone, give outstanding chemical linking sites available which leads to a more
resistance (acids, caustics and solvents) to this heat resistant polymer network. They are pro-
class of resins. duced from novolac modified epoxy resins

I
I OH
OH -oC1HD-
H-y-CH2-O h ~h O-CH2-?-H ~
CH2
I '" II CH 3 '" II CH2
I
o 0
I I
f=o 9=0
C- CH 3 C- CH3
"
CH 2 "2
CH

Fig. 2.8
 Vinyl ester resins 41

and methacrylic acid which provides more cal properties can be 'tailored' to meet the
unsaturation sites and higher molecular requirements of specific applications. Another
weight due to the epoxy backbone. These unique property of vinyl ester is the bondabil-
vinyl esters increase the heat resistance by ity of these resins to other surfaces. They are
17-27°C (3O-S0°F) over the general purpose not as good as epoxy resins in this character-
types. This often translates to higher useful istic, but obviously the epoxy resin
operating temperatures for vinyl ester based component gives them a boost over other
reinforced plastics even in corrosive environ- unsaturated polyesters in this area. A case can
ments. The higher-density cross linked vinyl also be made for vinyl esters providing better
esters are less resilient (lower tensile elonga- fiberglass wet out in FRP composites due to
tion) but still have excellent mechanical the backbone hydroxyl groups and their inter-
properties. Cure of the higher cross linked action with these groups on the fiber surface.
vinyl esters may require the use of different Some fabricators have reported that observ-
peroxide catalysts to reduce the peak able resin savings can be achieved with vinyl
exotherm and thereby prevent cracking/ craz- esters because of this characteristic.
ing in resin rich areas. In other words, resins of However, vinyl esters such as bisphenol A
this type are more reactive and more caution is polyesters and chIorendic polyesters are made
required in the fabrication of FRP laminates. from higher cost materials and often require
extended process times which leads to higher
finished cost. Accordingly, the specifier / fabri-
2.3.2 PROPERTIES/APPLICATIONS
cator turns to commercial applications where
The development of vinyl esters has led to the the improved performance of vinyl esters can
fastest growing segment of the thermosetting justify the premium price of the finished com-
resin industry today. This is not surprising, posite.
since vinyl esters combine inherent toughness The foremost application for vinyl esters is
with outstanding heat and chemical resis- in glass reinforced laminates for corrosion
tance. In all other thermosetting resin types resistant equipment. Because of outstanding
one has to sacrifice some heat resistance and chemical resistance combined with excellent
often chemical resistance to increase resiliency mechanical properties, vinyl ester based FRP
and toughness. Unlike polyesters, vinyl ester tanks, piping, scrubbers, fans and ductwork
resins possess low ester content and low are being specified for waste water treatment
unsaturation which results in greater resis- plants, mining facilities, chemical processing
tance to hydrolysis, lower peak exotherms and storage units, semi-conductor chip opera-
during cure and less shrinkage during cure. tions, pulp and paper manufacturing and odor
They are easily dissolved in reactive control facilities. Since FRP corrosion resistant
monomers such as styrene which provides equipment is the fastest growing segment of
easy handling and transportation to the fabri- the US composites industry, the future for
cation site. As with polyesters, other reactive vinyl esters looks extremely strong. They are
monomers such as vinyl toluene, chIorostyrene comparable to other premium resins for chem-
and t-butyl styrene can be employed with few ical resistance and secondary bonding
problems. combined with a good balance of chemical
The toughness of vinyl esters comes from resistance (acids, bases, solvents) at the same
the epoxy resin backbone. Since the molecular or lower cost. As a result, chlorendics and
weight and structure of the epoxy resin can be bisphenol A polyesters have been reduced to
varied like the polyester resin building blocks, 'niche' applications where their specific prop-
physical properties such as tensile elongation, erty advantages such as heat resistance and
heat distortion temperature and key mechani- resistance to oxidizing environments demand
42 Polyester and viny ester resins

their use. Since iso/terepolyesters also give an Table 2.2 summarizes the resin casting
excellent balance of properties in corrosion properties of the various resins used in corro-
applications, these unsaturated polyesters and sion resistant applications today.
vinyl esters now dominate the corrosion mar- The outstanding balance of properties pro-
ket. The bonus provided by vinyl esters is of vided by vinyl ester resins is obvious and
course higher heat resistance and extended life bodes well for continued strong growth in US
at higher operating temperatures, but at sig- corrosion markets. Other significant markets
nificant additional cost compared with the for vinyl esters includes pultruded construc-
iso / terepolyesters. tion and electrical components, automotive
The next major market area for vinyl esters structural applications, polymer concrete ves-
utilizes the high tensile elongation characteris- sels for mining and chemical operations,
tics of these resins to produce linings and grating, high performance marine applications
coating with outstanding adhesion to other and sporting goods.
types of plastics and conventional materials
such as steel and concrete. For example, vinyl
2.4 FLAME RETARDANT VERSIONS
esters are an excellent barrier coat for fiberglass
boats and acrylic spas. Vinyl ester corrosion The need for flame retardant polymers is
coatings are used everywhere today for steel essential in many plastics applications today.
tank linings and industrial flooring. In dual The combustibility of plastics has drawn so
laminate structures, a vinyl ester is often the much attention to the safety aspects of these
back up for exotic thermoplastics or the supe- materials in construction applications, that
rior corrosion barrier for lower cost polyesters designers and specifiers have been pressured
in many FRP tank and pipe applications. by fire officials to provide fiberglass-rein-
The growth of vinyl esters has also been forced construction materials that exhibit low
boosted by their excellent handling character- flame/low smoke characteristics. Since all
istics and ease of cure. For example, vinyl plastics are based on organic constituents, they
esters are much preferred by FRP fabricators are inherently flammable and once ignited will
in filament winding operations because of bum until they are completely consumed.
excellent glass wet out and in fabrication of There are, however several methods available
large structures because the resins are forgiv- for making thermosetting resin flame retar-
ing and provide predictable curing over a dant and these provide the capability to
wide range of temperatures. The latter situa- supply fire retardant FRP and corrosion resis-
tion has resulted in a virtual exclusive use for tant/fire retardant FRP for the numerous
vinyl esters in field fabrication of large FRP applications that have a need for some degree
structures. of fire retardancy.

Table 2.2 Resins for corrosion resistant applications

Tensile, Flexural, Elongation HDT,

psi psi break, % OF
General purpose vinyl ester 12500 20500 6.7 221
Heat resistant vinyl ester 13000 20000 5.6 248
Chlorendic polyesters 5500 10000 1.4 284
Bisphenol A polyester 10000 16500 3.2 288
Rigid isopolyester 8500 19000 1.9 234
Resilient isopolyester 12500 20000 4.4 201
 Flame retardant versions 43

Flame retardancy can be achieved by using range of chemical environments, both acid and
numerous additives, both organic and inor- alkali, at operation temperatures similar to the
ganic. However, most of these have a negative general purpose vinyl esters. Brominated high
effect on mechanical and/or chemical resis- molecular weight isopolyesters offer economic
tance properties. Accordingly, the most widely advantages and are suitable for moderate cor-
used system for achieving optimum fire retar- rosion applications. These two resin types
dancy will be covered here, namely, have become the workhorses for the waste
halogenated thermosetting resin systems com- water/odor control FRP market and the chem-
bined with inorganic synergists. ical and pulp/paper industries because they
exhibit excellent impact properties combined
with good overall corrosion resistance.
2.4.1 CHEMISTRY AND APPUCATIONS
Variations of these resins are used to meet
Flame retardancy of unsaturated polyester MIL-R-21607 or MIL-R-7575 requirements.
and vinyl ester resin is an extension of the non- Dibromoneopentyl glycol formulated with
flame retardant systems (as discussed above). carefully selected chemical building blocks
Almost all of these resins can be reformulated provides resins for exposure to severe weather-
to include a halogen in the chemical composi- ing conditions. The construction industry uses
tion by either blending or by an in situ cook of these resin systems, which are specially formu-
the resin. There is an advantage to locking in lated to meet optimum fire retardance for the
the halogen in the original resin cook, in order continuous line products of corrugated and flat
to chemically tie in the halogen (C~ or Br) to sheet panels. Such systems are formulated with
prevent migration of the halogen when sub- ultraviolet (UV) stabilizers and acrylates to
jected to thermal degradation. While flame achieve excellent color stability with acceptable
retardancy can be achieved with additives low smoke and flame spread (FS) properties. In
(Dekabrom or Dechlorine), these additives most cases, these formulations offer good
have not been used for high performance chemical resistance for splash and spill on the
applications in either the corrosion or con- exposed surfaces. Highly filled halogenated
struction industries (corrugated FRP panels). resin systems are designed to accept high filler
The chlorendic resins were developed in the loading with aluminum trihydrate (ATH) and
1950s and were based on HET acid (hexa- other synergists to meet DOT requirements for
chloroxyclopentadiene). Other formulations low smoke, low flame spread properties
followed, based on either tetrabromo bisphe- (ASTM-E-662 and E-162 respectively). Values
nol A (TBBPA) or dibromo neopentyl glycol of <150 smoke and 10 flame spread are
(DBNPG). These components react with achieved. Highly filled resins are specified for
maleic anhydride or fumaric acid in the pres- applications where people could be exposed to
ence of a difunctional glycol, to produce flame indoor fires, such as underground transporta-
retardant unsaturated polyesters that can be tion. Low smoke allows visual capability to
combined with styrene monomer, or other exit an entrapped area.
monomers used for smoke control, such as Electrical applications often require the
methyl methacrylate. The use of bromine as addition of halogenated base resins to achieve
the halogen in the resin building block has flame snuffing properties resulting from high
proved to be the most efficient way of achiev- voltage shorting or sparking. Such fire retar-
ing optimum flame retardant thermosetting dant systems are used for compression
resins. Certainly a lower percentage of molding of complex electrical shapes, using
bromine than chlorine is required for satisfac- BMC or SMC molding materials. Wet mat
tory reduction of flame spread. molding is also used to produce flat sheet for
Brominated vinyl esters handle a wide electrical insulation components.
44 Polyester and viny ester resins

When compared to the non-flame retardant and distance of the flame front advancing
versions, the addition of a halogen to the resin down the test tunnel during a ten-minute
formulation has very little, if any, effect on the duration test and compared to those values
chemical or mechanical properties of a FRP established for asbestos-cement board (at
laminate, whereas the use of additives such as O-FS) and red oak material (at 100 FS). This
aluminum hydrates, clays, carbonates or test method establishes the rating at 0-25FS as
fumed silica will have a direct and adverse Class I, 25-75FS as Class IT and 76-plus FS as
effect on the chemical resistance of FRP lami- Class ITr. During the test procedure, the
nates. In critical corrosion applications such smoke emission is measured and can range
filler additions could result in early failure of from 450 to 1000 or more for unfilled lami-
FRP laminates. nates. When additives such as ATH are used,
smoke emissions of less than 450 can be
achieved.
2.4.2 TESTING AND CLASSIFICATION
Corrosion resistant FRP ducting exhibiting
Most halogenated resin systems require a syn- low flame/low smoke characteristics is
ergist such as antimony usually 3-5% Sb20 3 by required for waste water/odor control and
weight of resin in order to achieve a class I semiconductor applications. This can be
flame spread rating. Other proprietary syner- achieved with brominated vinyl esters or
gists can be substituted, especially in brominated isopolyesters as the base for the
translucent laminates as used for siding and FRP duct which is then coated with an intu-
roofing materials. Antimony oxide enhances mescent paint to reduce smoke emissions.
the flame spread rating by forming a char on Such systems are currently qualified by the
the burning surface of a laminate and effec- International Conference of Building Officials
tively subdues the rate of flame progression by (ICBO) with tunnel test ratings of <25FS and
snuffing the flame out when the source of igni- <50 smoke development.
tion is removed or extinguished. Such flame Unfortunately, many specifying engineer-
resistant· FRP laminates will burn when sub- ing companies will request and specify
jected to a high temperature flame source, but values that are not readily achievable for
the rate of burning is substantially less than for most flame retardant resin systems. Ideally, a
non-flame retardant systems. The use of Sb20 3 Class I system with 25FS (max) to 450 (max)
will turn laminate opaque, which restricts smoke is acceptable for corrosion service,
visual inspections of a laminate in production. when the use of additives cannot be toler-
It is not a good idea to allow the addition of ated. A FSC of <25FS is usually acceptable for
Sb20 3 in the corrosion resistant barrier. Alone, most applications. However, some specifiers
Sb20 3 will not improve a non-halogenated will claim that 15-20FS is better than 20-25FS,
resin; its use becomes an unnecessary expen- when under actual burning conditions there
sive filler with no flame retardant properties. is a negligible difference in the combustibility
The most commonly used test method for of FRP laminates. When selecting a flame
evaluating flammability is ASTM Method E-84 retardant resin, it is important to qualify a
(the tunnel test) also known as NFPA 255. This system to meet the properties required. Flame
test method measures the comparative burn- retardant resins are available in a wide vari-
ing characteristics of a material by evaluating ety of formulations, including lower cost
the flame front propagation over the surface of general purpose to premium grade types
the test material, which is exposed to con- with better high temperature properties.
trolled temperatures in a forced air chamber or Class I flame spread thermosetting resin
tunnel. A flame spread (FS) classification (FSC) systems can also be achieved without the need
is obtained which measures the ignition time for a synergist like antimony oxide. These
 Design considerations 45

systems have a higher halogen content which product would essentially 'unzipper' and fail.
obviously increases resin cost. However, these The polyester and vinyl ester resins described
materials allow for the production of translu- in this chapter offer a wide selection of materi-
cent FRP products that are desirable in many als which will accomplish the need to protect
construction applications. Such a resin system the fiberglass and, at the same time, provide
has been specified by Disney engineers for optimum performance properties dictated by
their architectural applications at Disney the end use application.
entertainment centers. The first agenda in proper resin selection
involves an analysis of the key performance
requirements of the end use application. This
2.5 DESIGN CONSIDERATIONS should be very thorough as follows:
Since most thermosetting resin systems are • strength requirements;
used with fiberglass reinforcements, it is • thermal requirements;
important to consider material selection and • chemical exposures;
the fabrication process in establishing the • electrical requirements;
design of the FRP composite. • color requirements;
• surface requirements;
• environmental exposures;
2.5.1 MATERIAL SELECTION • fire resistance needed;
• smoke requirements;
The type of fiberglass reinforcement, place-
• potential upset conditions;
ment in the composite and fiberglass content
• number of parts required;
determines the strength of an FRP composite
• life expectancy.
and provides the mechanical properties dic-
tated by the end use requirements. In any part Resin selection is obviously very important in
made of FRP, the strength of the part will any FRP application as the above list of design
increase directly in relation to the percentage criteria illustrates, but is absolutely vital in
of fiberglass in the total weight of the compos- corrosion applications. Corrosive attack on a
ite. In addition, the arrangement and type of FRP laminate along with fire is the most criti-
fiberglass will have important effects on the cal situation the composite will face. A
resultant physical properties since the strength fiberglass building panel properly made can
obtainable in the finished part will be in the perform for an indefinite number of years, but
direction of the fibers. even a properly fabricated FRP tank exposed
The selection of the thermosetting resin sys- to concentrated acids at elevated temperatures
tem will determine the chemical, electrical and may only be good for 10-15 years. The best
thermal performance of the FRP product. example of this is in the pulp and paper indus-
However, the most significant contribution by try where vigorous chemical attack on FRP
the resin relates to 'life' of the composite, equipment can dictate replacement on a rou-
since the resin must protect the fiberglass. tine basis, say every five years. In spite of this,
Accordingly, if the resin fractures or blisters in FRP equipment may still offer the most cost
any manner that permits an attack on the glass effective material of construction.
fibers, the composite will lose strength rapidly Although all FRP composites will be
or delaminate. An interesting way to visualize attacked in the same manner in a particular
this is to consider a FRP pipe made by the fila- environment, certain types (chlorendic,
ment winding process. If the continuous fiber bisphenol A, vinyl esters and isopolyesters)
strands providing the hoop strength to the are significantly more resistant. These then
product are severed by chemical attack, the make up the list of corrosion and heat resistant
46 Polyester and viny ester resins

resins to choose from in addressing a specific process. Simply, 'you can select the best resin
application. A good rule to follow here is that and fiberglass in the world and if you don't
no single FRP resin can handle every kind of envi- put them together correctly - failure will prob-
ronmental problem, so resin selection is of the ably result'. Material selection added to design
utmost important. It should also be under- and production requirements leads to a deter-
stood that knowledge of the molecular mination of the fabrication process. Many
structure of these higher performance resins methods of fabrication are used to manufac-
does not eliminate the need for actual testing ture products for the numerous FRP markets.
to determine resin suitability in a given appli- These methods vary from hand lay-up I spray-
cation. For example, certain vinyl esters are up, filament winding and resin transfer
reasonably resistant to alkaline exposure while molding which utilize low temperature curing
other types are poor. The corrosion fabricators' to various high temperature molding com-
guide for the suitability of a thermosetting pound (SMC), pultrusion, and continuous
resin in a corrosive environment is AS1M C- panel. The designer must analyze the end use
581. The procedure involves complete property requirements such as color, surface
exposure of a FRP test laminate for one year, characteristics, strength and chemical resistance
with intermittent strength testing, to establish requirements and then add-in cost factors,
a curve which depicts loss of flexural strength part volume, part size and finishing to finalize
versus time. It is absolutely essential that the the selection of process. For example, trans-
resin selected for that environment form a portation body panels would be a high
plateau during the one year test period. volume application requiring outstanding sur-
Obviously, it is also important that this plateau face finish and excellent strength properties.
be achieved at a satisfactory retained flexural All of these can be satisfied with a isopolyester
strength. Table 2.3 summarizes a comparison sheet molding compound that is compression
of FRP properties of various thermosetting molded under heat and pressure. This process
resin types versus carbon and stainless steel. can be automated and delivers the highest vol-
ume and highest part uniformity of any
thermoset molding method. Lower part finish-
2.5.2 EFFECT OF PROCESS AND END USE
ing cost is achievable because subsequent
REQUIREMENT
trimming machining is minimized.
There is an old saying in the FRP business that Corrosion resistant equipment would be
heightens awareness of the fabrication fabricated, on the other hand, by either filament

Table 2.3 Comparison of properties of various types of FRP

Isophthalic Orthophthalic Chlorendic Bisphenol A Vinyl Carbon Stainless

fumarate ester steel steel
Corrosion resistance
Acids B C A A B C B
Alkalis B C C A A B B
Peroxides C C A B B C C
Hypochlorites C C A B A C C
Solvents B C B B B A A
Flame retardance C C A C A A A
Structural strength A B A A A A A
Thermal insulation A A A A A C C

A = High, B = Moderate, C = Low

 References 47

winding or hand lay-up / spray-up processing. parallel to the length of stock. Pultrusion is an
The former gives the highest strength to automated, low labor system which can use
weight ratio of any FRP manufacturing any type of thermosetting resin. However,
process. However, the most important consid- resilient resins such as isopolyesters and vinyl
eration is that these two process methods esters are much preferred because of the very
allow the easy creation of an effective resin- high glass content in the finished part. Low
rich corrosion barrier which is mandatory to cost reinforcement is adaptable to putrusion
satisfactory FRP life expectancy in corrosion because the glass weight percentage is high.
applications. The purpose of this barrier is to Pultrusion is used for FRP structural and elec-
isolate the fiberglass reinforcements from trical applications primarily, but the weight
attack that would result in wicking, blistering and density of the finished product does pro-
and delamination. The satisfactory corrosion vide moderate corrosion resistance properties.
barrier should be about 125 mils (3175 ~m) The value received from good design based
thick, fabricated with glass or polyester veil on on proper selection of materials and process
the surface, backed up by 2-3 plies of type E can be very rewarding for FRP composites.
chopped strand mat. The resin-rich corrosion The systematic analysis of end use require-
barrier should be constructed with the very ments, economic requirements and
best resin available in terms of chemical resis- competitive materials will enable the compos-
tance to the expected environment. For ite designer / specifier to optimize the
example, the corrosion barrier for 26% cost/performance of FRP as a material of con-
hydrochloric acid should employ a vinyl ester struction. The matrix materials available today
and two layers of polyester veil. It is always give the fabricator sufficient opportunities to
wise in dual laminate construction (different meet his final objective of providing the right
liner and wall resins) to utilize the material product at the lowest cost.
with the higher resiliency (higher tensile elon-
gation) in the liner portion of the laminate
REFERENCES
structure.
A final example of process effect on lami- 1. Ellis, c., US Patent 2255313; appl. August 6,
nate properties should address the rapidly 1937.
growing world of FRP pultrusion. This fabri- 2. Amoco Chemical Company, Bulletin IP-70a,
Chicago, nlinois.
cation method provides very high strength
due to high fiber concentration and orientation
EPOXY RESINS 3
L.S. Penn and H. Wang

3.1 INTRODUCTION crosslinked network and how they lead to dif-

ferent processing parameters and final
Epoxy resin systems have achieved acceptance
properties in the formulated system. The
as adhesives, potting compounds, molding
chapter also describes the role of cure moni-
compounds and as matrices for continuous fil-
toring and property evaluation in epoxy resin
ament composites used in structural
system technology.
applications. In this chapter, we discuss epoxy
resins in their role as matrices in fiber compos-
ites. In this role, they possess several 3.2 GENERAL DESCRIPTION OF
advantages over other types of polymers. The TIIERMOSETTING SYSTEMS
main advantages are:
3.2.1 DEFINITIONS
• inherently polar nature that confers excel-
lent adhesion to a wide variety of fibers; According to common chemical practice, mol-
• relatively low cure shrinkage that makes ecules can be classified by the functional
dimensional accuracy of fabricated struc- groups they contain. Thus, a molecule contain-
tures easier to obtain; ing the epoxide group (shown below) as part
• no volatile by-products of the curing reac- of its structure is called an epoxide, regardless
tion to cause undesired bubble or void of the remaining details of the molecule.
formation;
• crosslinked structure that confers excellent
resistance to hostile environments, both
aqueous and nonaqueous. /0"
-c-c-
In addition to these advantages, epoxy resins I I
have tremendous versatility because they can
be formulated to meet a broad range of specific
processing and performance requirements. To In practice, other types of molecules are added
know how to take advantage of the formula- to the epoxide to formulate a thermosetting
tion options, the engineer needs to have an system, i.e. one that will undergo a curing
elementary understanding of epoxy resin reaction to harden into a rigid form. The con-
chemistry and structure-property relation- fusing practice has been followed of referring
ships. This chapter attempts to provide that by to both the epoxide alone and the formulated
presenting information about the resin system system as epoxy resin'. One must determine
I

constituents, how they react together to form a which is truly intended by the context in
which it is used. For the purposes of clarity in
Handbook of Composites. Edited by S.T. Peters. Published this chapter, we will use epoxide when refer-
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 ring to the epoxide constituent alone and will
 General description of thermosetting systems 49

use the term epoxy resin when referring to the When the amine nitrogen contains two hydro-
uncured or cured formulated system. gens, each reacts with a different epoxide ring.
This scheme is shown in Fig. 3.1, where the
developing network is evident. This scheme
3.2.2 THE THERMOSETIING (CURING)
applies to both aliphatic and aromatic amine
REACTION
curing agents. The reaction between epoxide
The thermosetting reaction is the joining of and amine produces a C-N bond, whose envi-
many small molecules by chemical reaction to ronmental resistance is good, but whose
produce an extended network structure. stability to elevated temperature is highly
Although this process is a polymerization, it is dependent on the adjacent molecular structure.
distinct from the type of polymerization that
forms many individual long chains; the ther-
mosetting reaction unifies all the constituent
monomers into a single large molecule extend-
ing to the boundaries of the material.
~rn-CH2
°
/"- /0,,-
CH2-CB~
Epoxide molecules in the pure state at room
temperature normally do not react with each /0,,- H2N-R-NH2 /O,,-
other and can sit for years in a dry container ~rn-CH2 CH2- r n ...............
without mutual reaction. The types of chemi-
cals added to the epoxide to effect network
formation fall into two categories: curing
agents and catalysts. Curing agents, some-
times called hardeners, are added in
significant amounts to the epoxide and react
with it to become a part of the crosslinked net-
work. These curing agents can be aliphatic
amines, aromatic amines, or anhydrides.
Catalysts, on the other hand, are added in Fig. 3.1 Reaction scheme for the reaction between
extremely small amounts to cause the epoxide epoxide and amine curing agent. Each hydrogen
molecules to react directly with each other, i.e. reacts with an individual expoxide group. Thus the
to homopolymerize. Sometimes the chemicals primary amine group acts bifunctionally and the
used as catalysts for homopolymerization can secondary amine group acts monofunctionally. The
be used for another purpose; when added in epoxide group acts monofunctionally.
small amounts to epoxide-curing agent mix-
tures, they will accelerate the curing reaction.
In this role they are called accelerators rather From the reaction scheme, it is obvious that
than catalysts. the correct relative amounts of epoxide and
In the sections below, we discuss the chem- amine curing agent must be used. If there is .an
ical reactions involved in network formation, imbalance, unreacted functional groups will
both when the different curing agents are used be present and the full properties of a com-
and when homopolymerization occurs. plete network will not be developed. The
correct amounts by weight to combine are
determined by computing the weight of cur-
Amine curing
ing agent that contains one chemical
In amine curing agents, each hydrogen on an equivalent of amine hydrogens and matching
amine nitrogen is reactive and can open one that with the weight of epoxide that contains
epoxide ring to form a covalent bond 1-5 one chemical equivalent of epoxide groups.
50 Epoxy resins

Anhydride curing functions as a catalyst to speed up the epox-

ide-curing agent reaction. (Accelerators will be
In anhydride curing agents, the anhydride
discussed later.) Figure 3.2 shows an idealized
groups themselves must be cleaved asymmet-
scheme, where the cleaved anhydride reacts
rically to start the reaction1-4,6. Initial cleavage
with an epoxide ring carbon, opening the ring
is accomplished with the help of a small
in the process. The negatively charged oxygen
amount of accelerator, which is a chemical that
formed by the opening of the epoxide ring can
proceed to react with a different anhydride
group, perpetuating the reaction. In this ideal-
o o- ized developing network, each anhydride

CC\ - - " C
II j group is bifunctional, i.e. it links to two differ-
NR3 (Accelerator)
c=o ent epoxide molecules. In practice, the high
temperatures required for anhydride cure, plus
C/
II C=O the presence of accelerator, provides conditions
+
o for some extent of epoxide homopoly-
merization (described later) to take place,
making the actual curing reaction much more
o- complex than depicted in Fig. 3.2. The reaction
j
C=O between epoxide and anhydride produces pri-
marily ester linkages, which have good
CC=O stability to elevated temperatures and to most
+ hostile environments except bases.
o- o- Not surprisingly, the correct amount of
j
j anhydride curing agent relative to the epoxide
C=O C=O must be used to obtain a well developed net-
work and the associated good properties. The
CC=O CC=O
+

0-
1 o-
+ correct amounts by weight to combine are esti-
mated by examining the reaction scheme and
computing the weight of curing agent needed
j I to react completely with a given weight of
C=O C=O epoxide. The simplified reaction scheme of
Fig. 3.2 proposes that one anhydride group
CC=o CC=o reacts with two epoxide groups and one epox-
j I

O-CH-~C-CH-O
9 ? ide group reacts with two anhydride groups,
making the number of anhydride groups con-
j 2 2 j
sumed equal to the number of epoxide groups
C=O C=O consumed in the reaction.
CC=O
+
CC=~+ Catalytic curing (homopolymerization)
The remaining route to formation of a
crosslinked network from epoxide molecules
Fig. 3.2 Simplified reaction scheme for the reaction requires homopolymerization1-4,7. This can be
between epoxide and anhydride curing agent. After
cleavage, each anhydride group reacts with two brought about if small amounts of certain
epoxide groups and each epoxide group reacts with Lewis acids or Lewis bases are added. These
two anhydride groups. Thus both the epoxide operate as true catalysts by initiating a self-
group and the anhydride group act bifunctionally. perpetuating cationic (Lewis acid) or anionic
 Constituents used in formulated systems 51

+
R3NCH2-QI- CHi'-'.....' .....·~
I
oI
CH2-QI~
I
0-
Fig. 3.3 Reaction scheme for the homopolymerization of epoxide. After ring opening, each epoxide group
reacts with two other epoxide groups. Thus the epoxide group acts bifunctionally.

(Lewis base) polymerization. A simplified 3.3 CONSTITUENTS USED IN FORMULATED

reaction scheme initiated by a Lewis base is SYSTEMS
shown in Fig. 3.3. Homopolymerization
results in the formation of a network of ether 3.3.1 EPOXIDES
linkages, which has outstanding elevated-tem- Although many different types of epoxides are
perature stability and resistance to hostile available, only a few are favored for use in
environments. matrices in fiber composites8--10. These are
shown in Table 3.1.

Table 3.1 Structures and characteristics of commonly used epoxides

Epoxy equivalent Viscosity at 25°C, Comments

weight, g/eq Pa s (cP)

Diglycidyl ether of bisphenol A (DGEBA)

171-177 3.5-5.5 May crystallize on storage.

(3500-5500) Example: DER 332 (Dow).
180-188 6.5-9.6 Contains small amount of higher polymer
(6500-9500) to prevent crystallization. Examples: DER 330
(Dow) Epon 826 (Shell).
185-200 10.0-19.0 Contains small amount of higher polymer to
(10 000-19000) prevent crystalliza1ion.
Example: Epi-Rez 510 (Hi-Tek Polymers),
DER 331 (Dow), and GY 6010 (Ciba).
450-550 Melting point n '" 2; used in prepregs. Examples: Epon 1001
65-76°C (149-167°F) (Shell) and DER 661 (Dow).
52 Epoxy resins

Table 3.1 Continued

Epoxy equivalent Viscosity at 25°C, Comments

weight, g/eq Pa s (cP)

Diglycidyl ether of bisphenol F (DGEBF)

158-165 5.0-8.0 Isomeric mixture that will not crystallize on

(5000-8000) storage. Example: GY 281 (Ciba).

Polyglycidyl ether of phenol-formaldehyde Novolac

H------CH2
n

172-179 1.1-1.7 (1100-1700) Example: DEN 431 (Dow).

at 52°C (126°F)
17Cr181 20.0-50.0 (20 000-50 000) Example: DEN 438 (Dow).
at 52°C (126°F)
191-210 4.0-10.0 (4000-10 000) Example: DEN 439 (Dow).
at 52°C (126°F)

Polyglycidyl ether of o-cresol-formaldehyde Novolac

/0" /0"
O-CH2 -CH- CH2 O-CH2 -CH- CH2
'" __ "--' CH3 /.-......",rCH3
V-
};_C_H_3_ __
CH2
n

200 Melting point Used for high-temperature service: R represents

35°C (95°F) chlorohydrins, glycols, and/ or polymeric ethers.
Molecular weight = 540. Example: ECN 1235
(Ciba).
225 Melting point Same as above, but molecular weight = 1080.
73°C (163°F) Example: ECN 1273 (Ciba).
230 Melting point Same as above, but molecular weight = 1170.
BOoC (176°F) Example: ECN 1280 (Ciba).
235 Melting point Same as above, but molecular weight = 1270.
99°C (210°F) Example: ECN 1299 (Ciba).
 Constituents used in formulated systems 53

Table 3.1 Continued

Epoxy equivalent Viscosity at 25°C, Comments

weight, g/eq Pa s (cP)

N, N, N', N', - Tetraglycidyl methylenedianiline

117-133 10.0-15.0 (10 DOG-IS 000) Used for prepregs. Example: MY 720 (Ciba).

95-107 0.55-0.85 (550-850) Used extensively for prepregs and adhesives.

Example: MY 0510 (Ciba).

Conspicuously, these epoxides all contain aro- achieve cure with amine curing agents fall in a
matic rings in their structures. Aromatic rings wide range, from 25°C to nearly 200°C,
confer mechanical rigidity and thermal stabil- depending on the chemical structure of the
ity to the crosslinked network. It is also worth amine. The first five entries in the table are
noting that some of the epoxides in the table aliphatic amines, which can cure epoxides at
have two epoxide functional groups, while room temperature or only slightly above.
others have three or four or more. Whether a Aliphatic amine-cured systems also tend to
network is developed by mixing the epoxide have low glass transition temperatures, (the r
with a curing agent or is developed from epox- temperature at which the mechanical beIJvior
ide alone by catalytic homopolymerization, a changes from rigid to rubbery) and cannot be
large number of the molecules in a given for- used in composites that will experience high
mulation must be able to react with more than temperature use.
two other molecules in order to form a Most other entries in the table are aromatic
crosslinked network instead of merely form- amines, whose ring structures confer solidity
ing linear chains. and mechanical rigidity. These amine curing
agents require elevated temperature cure, but
the networks they produce have high glass
3.3.2 AMINE CURING AGENI'S
transition temperatures and are suitable for
Table 3.2 presents commonly used amine cur- use in composites that will be exposed to ele-
ing agents. The temperatures required to vated temperatures in service.
54 Epoxy resins

Table 3.2 Structures and characteristics of commonly used amine curing agents

Amine hydrogen Viscosity at 25°C (lJOn Comments

equiv. weight, g/eq Pa s (cP)

Oiethylenetriamine (OETA)

H2N-CH2-CH2-NH - CH2-CH2 - NH2

20.6 0.0055-{).0085 Available from Dow as DEH 20.

5.5-8.5)

Triethylenetriamine (TETA)

24.4 0.020-0.023 Available from Dow as DEH 24.

(20-23)

Diethylaminepropylamine (DEAPA)

65 50 Available from Union Carbide.

(5000)

Tetraethylenepentamine (TEPA)

2~27 0.055 Available from Dow as DEH 26.

(55)

Aliphatic polyether triamine (APTA)

H 2C+0 -CH2 - CH -(CH3)

I
+ x
NH2

CH3 - CH2 - C-
I
CH2 + 0 - CH2 - CH -(CH3 )+Y NH2

H2C + 0 - CH2 - CH -(CH3) + z

NH2

77~2 0.072-0.080 x + y + z ... 5.3. Available from Texaco as

(72~0) Jeffamine T 403.
 Constituents used in formulated systems 55

Table 3.2 Continued

Amine hydrogen Viscosity at 25°C (77°F), Comments

equiv. zoeight,g/eq Pa s (cP)

4,4'-Methylenedianiline (MDA)

50 Melting point Available from Ciba as HI 972 and from

89°C (192°) Pacific Anchor as Ancamine DDM.

m-Phenylenediamine (MPDA)

5 NH2

27 Melting point Available from E.!. duPont de Nemours &

60°C (140°F) Company.

4,4' -Diaminodiphenylsulfone (DDS)

o
H2N -0- ~ -0- o
NH2

62 Melting point 170-180°C Used mainly in prepregs; yields good shelf life and
(338-356°F) high-temperature properties. Available from
Ciba as HT 976.

3,3' -Diaminodiphenylsulfone

62 Melting point 174-178°C Used mainly in prepregs; reacts more slowly than
(345-352°F) 4,4' analog. Available from Ciba as HI 9720.
56 Epoxy resins

Table 3.2 Continued

Amine hydrogen Viscosity at 25°C (77°F), Comments

equiv. tueigh~gleq Pa s (cP)

40% MPDA-60% MDA

0- NH, H,N -Q-CH,-Q-NH,

H2N

38 1.50 Eutectic mixture. Available from UniRoyal

(1500) as Tonox 60-40

Dicyandiamide (DICY)

NH2
I
H 2N-C=N-C=N

28 Melting point 207-209°C Slow reacting. Used for prepregs. Available from
(405-408°P) Cytec Ind.

3.3.3 ANHYDRIDE CURING AGENTS Elevated temperature, typically in the range

Table 3.3 presents commonly used anhydride 100-200 C, is required to achieve cure with
D

curing agents. Their structures vary widely anhydride curing agents. The glass transition
and some are liquid at room temperature temperatures of anhydride-cured systems are
whereas others must be heated to liquefy. correspondingly high.

Table 3.3 Structures and characteristics of commonly used anhydride curing agents

Anhydride equiv. Melting point, DC (OF) Comments

tueight, g/eq

Phthalic anhydride (PA)

148 130 (266) Available from Amoco.

 Constituents used in formulated systems 57

Table 3.3 Continued

Anhydride equiv. Melting point, °C (OF) Comments

weight, g/eq

Hexahydrophthalic anhydride (HHPA)

154 40 (104) Available from Pacific Anchor as Ancadride

MHHPA and from Ciba as HT 907.

Nadic methyl anhydride (NMA) maleic anhydride adduct of methyl cyclopentadiene

180 Liquid at 25°C (77°F) Widely used for prepegs. Available from Pacific
(0.200 Pa s) 200 cP Anchor as Anhydride METHPA and from
Ciba as HY 906.

Dodecenyl succinic anhydride (DDSA)

CH3
I C
I?
CH3 -CH2 -CH2-CH-CH2-C=CH- C ---T"' \
I I I'T~ /0
CH3 CH3 CH3 C
~

°
270 Liquid at 25°C (77°F) Available from Dixie Chemical and from
(0.200 Pa s) 200 cP Humphrey Chemical.
58 Epoxy resins

Table 3.3 Continued

Anhydride equiv. Melting point, °C (oF) Comments

weight, g/eq

Chlorendic anhydride (CA)

371 231-235 Needs no cure accelerator, but is high melting and

(448-445) hard to handle; good fire retardant. Available
from Velsicol.

Trimellitic anhydride (TMA)

193 161-164 Available from Buffalo Color.

(322-327)

Maleic anhydride (MA)

100 53 (127) Available from Amoco.

Succinic anhydride (SA)

IP
C
C\
C
/0
~

100 120 (248)

°Available from Buffalo Color.
 Constituents used in formulated systems 59

Table 3.3 Continued

Anhydride equiv. Melting point, °C (oF) Comments

weight, g/eq

Methyltetrahydrophthalic anhydride

166 Liquid at 25°C (77°F) Available from Lindau.

(0.06 Pa s) 60 cP

3,3',4,4' - Benzophenone-tetracarboxylic dianhydride (BillA)

<:©r "rUe>
O~ ~ p

II ~
o 0
161 221 (430) Used mainly in powder coatings; when used as
minor component in fiber composite matrix, it
improves high-temperature properties. Available
from Allco.

3.3.4 CATALYSTS FOR useful in prepregs that must be stored for

HOMOPOLYMERIZATION some time without cure advancement prior to
Table 3.4 presents some Lewis acids and Lewis being used in fabrication of a structural com-
bases that have been found effective as cata- ponent.
lysts for homopolymerization of epoxides.
Certain Lewis acids, such as boron trifluoride 3.3.5 ACCELERATORS FOR CURING AGENT
(BF3) produce rapid and very exothermic SYSTEMS
homopolymerization and need to be used in
Some Lewis acids and Lewis bases can also be
blocked form. BF3' when blocked with
used as accelerators in epoxide-curing agent
monoethyl amine to form the complex
mixtures to speed up a sluggish reaction. They
BF3MEA, is latent at room temperature and
are added in small amounts, only a few weight
becomes active only above 90°C, the tempera-
per cent, empirically determined to give the
ture at which the complex separates. The
best results. Most often they are used to speed
epoxide homopolymerization that occurs
up the curing reaction in epoxide--anhydride
above 90°C is rapid, but has controllable
systems. Table 3.4 indicates which catalysts are
release of heat. Epoxides containing blocked
also used as accelerators and lists additional
Lewis acids as catalysts have been found to be
chemicals used as accelerators only.
60 Epoxy resins

Table 3.4 Structures and characteristics of commonly used catalysts and accelerators

Amine hydrogen Melting point, °C (OF) Comments

equiv. weight, g/eq

Benzyldimethylamine (BDMA)

Liquid at 25°C (77°F) Lewis base used as an accelerator mainly for

(0.1 Pa s, 100 cP) anhydride mixtures. Avalible from Ciba
as DY062.

2, 4, 6-Tris(dimethylaminomethyl)phenol

H3C~N_CH2~OOH CH2 _< CH

3

H3C ¥ CH3

CH2
I
N
H3C
/ "-CH3
Liquid at 25°C (77°F) Lewis base used as an accelerator for epoxide
(0.3 Pa s, 300 cP) anhydride mixtures to provide room-temperature
cure. Available from Rohm & Haas as DMP-30 and
from Ciba as DY 064.

2-Ethyl-4-methylimidazole (EMI)

Liquid at 25°C (77°F) Lewis base used as an accelerator for epoxide-

(4-8 Pa s, 4000-8000 cP) anhydride mixtures to provide long pot life and
good elevated-temperature properties. Available
from Air Products as EMI-24.
 Constituents used in formulated systems 61

Table 3.4 Continued

Amine hydrogen Melting point, °C (OF) Comments

equiv. weight, g/eq

Boron trifluoride-monoethylene amine (BF3MEA)

F
I
F - B:NH2 - CH2 - CH3
I
F
85--90 (185--194) Blocked Lewis acid; used as an accelerator for
epoxide DDS systems in high temperature
service; provides latency. Available from ATOTech.

3.3.6 DILUENTS epoxide groups. Some of these diluents are

For some types of processing the viscosity of monofunctional, i.e. have only one epoxide
the uncured resin system needs to be lowered. group, so they cannot form crosslinks in the
When it is not advisable, as in the case of wet network like bi-, tri- and tetrafunctional mole-
filament winding or hand lay-up, to use cules do. However, by chemically attaching to
volatile solvents as diluents, reactive diluents the network, the reactive diluents become a
must be used. Table 3.5 lists some acceptable permanent and stable part of it. Other diluents
diluents, all low viscosity liquids containing are bifunctional and will form crosslinks.

Table 3.5 Structures and characteristics of commonly used, commercially available epoxide reactive dilutents

Epoxy equivalent Viscosity at 25°C, Comments

weight, g/eq Pa s (cP)

Butyl glycidyl ether (BGE)

130--149 0.002-0.003 Example: RD-1 (Ciba).

(2-3)

Oetyl, deeyl glycidyl ether blend

215--235 0.005--0.015 Example: Epotuf 37-057 (Reichhold) and Dy 027

(5-15) (Ciba).
62 Epoxy resins

Table 3.5 Continued

Epoxy equivalent Viscosity at 25°C, Comments

weight, g/eq Pa s (cP)

p-t- Butyl phenyl glycidyl ether

220-245 0.015-0.30 Example: Epi-Rez 5014 (Hi-Tek Polymers).

(15-30)

Phenyl glycidyl ether (PGE)

ISO 0.006 Example: Heloxy WC-63 (Wilmington Chemical).

(6)

Cresyl glycidyl ether (CGE)

~ /0"
C H 3 V 0 - CH2 - CH- CH2

170-190 0.005-0.050 Less volatile than BGE. Example: Epotuf 37-053

(5-50) (Reichhold), and by DY 023 (Ciba).

Oiglycidyl ether of 1,4 - butanediol (BOE)

120-140 0.010-0.025 Example: RD-2 (Ciba)i not a pure compound.

(10-25)
 General principles offormulation 63

Table 3.5 Continued

Epoxy equivalent Viscosity at 25°C, Comments

weight, g/eq Pa s (cP)

Diglycidyl ether of neopentyl glycol

130-145 0.OO~.015 Example: AZ epoxy N (AZS Corp.); not a pure

(5-15) compound.

Diglycidyl ether of polypropylene glycol

175-205 0.30-0.060 (30-60) n = 4. Example DER 736 (Dow).

305-335 0.5~.100 (55-100) n = 9. Example DER 732 (Dow).

Vinyl cyclohexene dioxide

76 0.20 Example ERL 4206 (Union Carbide).

(20)

3.4 GENERAL PRINCIPLES OF

FORMULATION
the cured network results in lower rigidity and
lower T as compared to aromatic rings, or
Epoxy resin systems must be formulated on a even sa~ated rings. Thus, aromatic amine-
rational basis and the chemical structure of the cured systems or homopolymerized systems
constituents forms this basis. Many detailed both have high stiffness and high T g' whereas
structure-property relationships can be an aliphatic amine-cured system Will have a
reduced to rules of thumb and some of those lower stiffness and lower Tg •
we present in this section. The reader should In terms of processing and cure, the flexibil-
use these with caution, because often a more ity and mobility of aliphatic segments imparts
detailed examination of the chemical structure low viscosity (if a liquid) or low melting point
reveals conflicting trends and experimentation (if a solid). Constituents whose structures are
is needed. mainly aliphatic react rapidly at room temper-
First, the presence of significant amounts of ature. Thus, an aliphatic amine-cured system is
aliphatic segments in the chemical structure of recommended over an aromatic amine-cured
64 Epoxy resins

system if low viscosity for processing and a in all but a few cases; the anhydride-cured sys-
room temperature cure are needed ll,12. tem is degraded in strong base due to basic
Constituents with aromatic ring structures hydrolysis of its ester linkages and both the
react sluggishly or not at all at room tempera- anhydride-cured and the homopolymerized sys-
ture13,14. Thus aromatic-amine cured systems tems are vulnerable to swelling by the strong
require elevated temperature cures. solvent, trichloroethylene.
As cited earlier in the discussion on cure
reactions, the relative amount of curing agent
3.5 PROCESSING CONSIDERATIONS IN
to epoxide is important in achieving a well-
EPOXY RESIN COMPOSITES
developed network. The formulations
presented in Table 3.6 are approximately what The goals of the processing procedures used to
would be used for some specific epoxide-curing make a good quality fiber composite are to
agent formulations. The reader should verify ensure that the resin forms a void-free contin-
the correctness of the formulations by comput- uous phase, surrounds each filament, is evenly
ing the appropriate weight ratios from the distributed, is present in the desired amount
molecular structures given in Tables 3.1, 3.2 relative to the fiber and is fully cured. The ease
and 3.3. with which these goals can be achieved is
highly dependent on the rheological proper-
Table 3.6 Formulations for selected epoxy resin ties of the resin as it progresses through its
systems cure and on the engineer's ability to evaluate
rheology and degree of cure throughout the
Constituents Epoxide Curing agent processing cycle.
pbw" pbw" In the early years of epoxy resin technol-
Diglycidyl ether of ogy, processing procedures were developed
bisphenol A triethylene by a trial and error approach. This is still a
tetramine 100 14 viable approach, especially for engineers
Diglycidyl ether of skilled in the art of thermoset composite pro-
bisphenol A meta- cessing. However, in recent years,
phenylene diamine 100 16 rheologicaP6-18 and chemorheologicaF9,20
models that relate viscosity, rigidity and
Diglycidyl ether of
bisphenolA degree of cure to time and temperature have
hexahydrophthalic been developed. These models predict rheo-
anhydride 100 90 logical changes during cure and can serve as
a Parts by weight an aid to processing and cure cycle develop-
ment.

Finally, through the chemistry of the constituents

3.5.1 PROCESSING VARIABLES EARLY IN CURE
and the network they form, the formulation
influences the environmental resistance of the One of the major concerns in the early stages
resin. Thus the formulation must be selected of composite processing is resin viscosity.
with the future environmental exposures in Sometimes, the freshly mixed, uncured resin
mind. Figure 3.4 compares the environmental formulation is inherently fluid enough to pen-
resistance of four important types of epoxy resin etrate the fiber bundle and surround each
systems: aliphatic amine-cured, aromatic amine- filament. (Tables 3.1, 3.2 and 3.3 give viscosity
cured, homopolymerized epoxide (BF3MEA values for several epoxides and curing
catalyst) and anhydride-cured1s. The excellent agents.) Often, however, the viscosity of the
resistance of the crosslinked network is evident uncured resin system is too high and must be
 Processing considerations in epoxy resin composites 65

1()()

*-
I 25
&
c:
'=
ta
U
>
x Sodium hydroxide, 50% Sulfuric acid, 25% Hydrochloric acid, 25%
0

-
Co 8~C (180" F) 8:ZOC (18d'F) 8~C (180"F)
OJ
0
c:
o
.....
.;
c:
OJ
GJ

.!:::l
-g 1()()
E
...
i;j

~ 75
GJ
u:::
50

25

Distilled water Trichloroethylene Sodium hypochlorite, 6%

54D C (13(1' F) 54 D C (130°F) 54 D C (130°F)
Exposure conditions

Fig.3.4 Environmental resistance of common cured epoxy systems as indicated by flexural modulus reten-
tion after environmental exposure 15.
Aliphatic amine-cured (TETA), ~ Homopolymerized (BF3MEA), ~
Aromatic amine-cured (MPDA), ~ Anhydride-cured (PA), IlIill
reduced to achieve the desired flow require- the simplest approach, can result in bubble or
ments. The two major approaches to reducing void formation within the composite if the
viscosity are thinning the mixture with low solvent cannot escape completely. This could
viscosity organic solvents and adding low vis- be a problem for component fabrication by
cosity reactive diluents, such as those wet filament winding, where layers contain-
presented in Table 3.5. ing the freshly mixed epoxy resin system are
Thinning with organic solvents, although placed sequentially on top of one another.
66 Epoxy resins

Filament winding processors do not use sol- By contrast, reactive diluents, being them-
vents to reduce viscosity, rather selecting selves epoxides, chemically react to become a
lower viscosity resins, reactive diluents, or permanent part of the crosslinked network.
diluting with heat. On the other hand, the use Figure 3.5 shows the relation between viscos-
of small amounts of solvent to reduce resin ity and the amount of diluent added to a
viscosity during fabrication of prepreg viscous epoxide. Ideally, the engineer wants to
(pre-impregnated fiber) presents no prob- use just enough diluent to lower viscosity as
lems, since prepreg is made in the form of needed without dramatically altering the
single, thin-layer sheet, tape or tow from properties of the final cured network.
which solvent can vaporize easily and the The length of time that an epoxy resin for-
prepreg is heat treated to eliminate solvent mulation remains fluid is important.
and to advance cure. Liquid-like flow becomes impossible once the
gelation stage, marked by an abrupt increase
in viscosity, is reached. TIme to gelation is
called gel time, or sometimes pot life.
Aliphatic amine curing agents produce pot
-....- 17"....
~
•
lives of the order of minutes or a few hours,
while aromatic amine curing agents produce
!:::. pot lives of 24 h or more13,22,23. Anhydride cur-

.-
.111
0
•." ing agents typically produce very long pot
N lives (e.g. two months for NMA) when mixed
ii with epoxides. This is because, as already

'.
1-
:i- mentioned, the anhydride group is not very
0
1100
reactive with epoxides unless it is cleaved with
U
II)
:.; 1100 the aid of an accelerator molecule. Once the
accelerator is activated, the pot life of the mix-
IGOO
ture will be shortened to a few hours. Pot life
,o0 can be controlled over a wide range by careful
100
use of accelerators.
Each resin formulation has a unique chem-
700
istry that imparts a set of processing variables
.00 with unique values. Standard laboratory test
SOO
methods for processing variables that are
important early in cure are described by the
400 American Society for Testing and Materials,
100 Philadelphia, Pennsylvania. ASTM numbers of
the procedures for determining resin viscosity,
lOO
gel time and melt flow are listed in Table 3.7.
100

o JO 10 10 40 \0 .0 70 10 90 100 Table 3.7 Laboratory test methods for measuring

DILUENT, % processing variables

Processing variable Reference

Fig. 3.5 Viscosity versus per cent of diluent in epox-
ide--diluent mixtures21 • The epoxide is DGEBA and Viscosity ASTMD2393
the diluent is BDE (diglycidyl ether of 1, 4-butane Gel time (pot life) ASTMD2471
diol). Melt flow ASTM D3795, ASTM 04473
 Processing considerations in epoxy resin composites 67

3.5.2 MONITORING OF CURE ods where samples were taken and tested at
Once the cure is underway in an epoxy resin intervals in the cure cycle. The time needed to
system, all of the properties of the system develop one data point depended on the par-
change rapidly until the final crosslinked net- ticular laboratory method. Off-line testing has
been made more convenient with the develop-
work. ~ ~ac.hed. N.0t only is flow decreasing
~ent of continuous monitoring techniques
and ngIdlty mcreasmg, but all other properties
(I.e. data points in real time) used on small
~electrical, chemical, optical, etc.) are chang-
dedicated specimens of resin or prepreg.
mg. When the curing epoxy resin system is
Recently, real time, in situ methods where cure
subjected to temperature and pressure
changes, the resin response will be characteris- can be monitored in the composite structural
component itself have been developed.
tic of its degree of cure at that moment. Cure
cy~le events, such as temperature and pressure
changes, need to be carefully timed with this Traditional off-line methods
in mind.
Intermittent off-line methods include chemical
The degree of cure can be defined in terms
of anyone of a large number of chemical or titration of the unreacted epoxide groups pre-
sent24,25, specific gravity to measure
physical (including mechanical) properties that
dens~fication26 and differential scanning
change continuously during the curing reac-
calonmetry to measure the residual cure
tion and reach a constant value at end of the
e~?therm.20,27,28. Figure 3.6 shows a plot of spe-
cure. Evaluation of the degree of cure is usually
Cific gravity data obtained on specimens cured
based on one of these properties and full cure
for increasing lengths of time at a single tem-
is then defined as the point at which this
selected property reaches a constant value. perature. Figure 3.7 shows differential
Originally, the only available methods for scanning calorimeter scans for two epoxy resin
evaluating cure were off-line laboratory meth- specimens of the same formulation, but with
different degrees of cure.

1.24 1 - - - - - - - - - - - - - - - - - - .
1.22 f-

+
1.20 -

E 1.18 r
"+t -tft'+ + \ + + + +

~
C)

~
1.16 -:t=
'ii)
c:
C
Q)
1.14 ~
1.12
1.10 I I I I I I I I I I I I

0 50 100 150 200 250 300 350 400 450 500 550 600 650
Time (min)
Fig. 3.6 Specific gravity compared with cure time for an aromatic amine-cured epoxy system cured at
120°C 26. Voltl~e reduction (densification) during early network formation is rapid and levels off as cure
nears completion.
68 Epoxy resins

Exotherm

Energy
calls

Endotherm
100 300 400
Temperature - °c
Fig. 3.7 Differential scanning calorimeter scans for an epoxy resin system28• Scans for two different speci-
mens, each having a different original degree of cure, are shown, with scan A displaced upward from
scan B for graphical visibility. The height of the residual cure exotherm is inversely related to the original
degree of cure. The higher exotherm peak in B indicates an original degree of cure lower than in A.

Continuous off-line methods include infrared The off-line methods, real time or not, are
spectroscopyS,29,30, parallel plate-type bulk useful for developing a cure schedule for a
dielectrometry31,32 and dynamic mechanical new epoxy resin formulation, for optimizing
spectrometry33-38. Figure 3.8 shows the results processing variables and for quality control of
of infrared monitoring for five different neat incoming resins or prepregs. They have also
resin specimens, each cured at a different tem- been used successfully for the development of
perature and Fig. 3.9 shows data obtained by mathematical models of cure kinetics.
bulk dielectrometry. However, because they are off-line, they can-
not be used for process control.

105°e
0~
I 95°e

-
~
:::l
0
77°e
0
Q)
~
C)
Q)
0

20
Time-h
Fig. 3.8 Degree of cure compared with time for an aromatic amine-cured epoxy system28 • Data for five
specimens, each cured at a different temperature, are presented.
 Property data for cured epoxy resin systems 69

200~--~--~--'----r---r--~--~--~
~

II \ rDecreasing molecular dipole mobility

Vdue to curing
150 , , Temperature profile
\
o
(J
\
I \ Increasing molecular dipole mobility
~ \ due to temperature elevation
~ 100 \
iE , c
o
\ 'i
-----7-
1-........... _
~ \
\ I 'iii
Ci
50
\ Nofunher
\ cure at
\.J temperature
OL---~--~--~--~--~--~--~--~
o 2 4 6 8
Tirne- h
Fig. 3.9 Bulk dielectrometry data compared with time for an aliphatic amine-cured epoxy system28,
Dissipation factor (dashed line) is inversely related to rigidity in the developing network. The applied tem-
perature is shown by the solid line.

Modern in situ methods Other examples of in situ monitoring are

In situ methods for cure monitoring are real based on fluorescence spectroscopy of tag
time methods that require sensors small molecules in the resin43,44 and on infrared45,46
enough to implant and leave in the composite and Raman47 spectroscopies of the resin mole-
itself. When the information from in situ mon- cules themselves. In these methods, optical
itoring is used as continuous input to an fibers are the sensors that transmit the appro-
appropriate process model, it can be used in priate wavelength of light into the curing resin
process control loops that adjust the process- and also transmit the spectral information
ing conditions automatically. back ours. The spectral changes relate directly
One example of an in situ method is low fre- to the chemical changes that occur as the cur-
quency dielectrometry using a very small ing reaction progresses to completion,
assembly of interdigitated electrodes called a
fringe field sensor. This technique measures
3.6 PROPERTY DATA FOR CURED EPOXY
changes in ability of permanent dipoles within
RESIN SYSTEMS
the resin chemical structure to align themselves
with the applied oscillating electric field and Property data for cured, unreinforced epoxy
also measures changes in the mobility of ions resin systems are needed for two purposes.
present as impurities in the resin39-41. Both of First, they are useful when selecting the best
these quantities correlate with resin viscosity in fiber and matrix combination for a particular
the early stages of cure and with mechanical application. Resin system choices can be
rigidity in the later stages of cure39,4Q,42. rapidly narrowed down to a few alternatives
70 Epoxy resins

when comparisons of key properties are made and many other properties of the composite can
from existing data tables or manufacturers' be computed in advance, if one has the corre-
data sheets. sponding values for the fiber and the matrix.
Second, epoxy resin data are required in Tables 3.8, 3.9 and 3.10 present property
micromechanics computations of composite data for three major resin systems: aliphatic
properties. Elastic constants, thermal expansion amine-cured, aromatic amine-cured and anhy-
coefficients, moisture absorption coefficients, dride-cured. Property data for commonly used
Table 3.8 An aliphatic amine-cured epoxy resin system, room-temperature curable12,28

Resin system constituents Parts by weight

Epoxy: DGEBA, e.g., DER 332 (Dow) 100
Curing agent: APIA, e.g. Jeffamine T-403 (Texaco) 45
Cure cycle: 16 h at 6Q°C (140°F) for improved properties over room temperature cure
Viscosity at 25°C (77°F) Heat distortion temperature at
0.8 Pa s (800 cP) 1820 kPa (264 psi) 62°C (144°F) a

Density of cured resin at 25°C (77°F) Coefficient of linear thermal expansion

1.16 gcm-3 • from 298 to 374 K 66 X 10-5 0C-I
Volumetric shrinkage Average specific heat from
After gelation 4.4% • 286 to 367 K 1.75 X 103 J kg-1 K-l a
After cure 4.4% • Coefficient of thermal conductivity
Water absorption, wt. gain after 2 h in At 298 K 0.133 W m-1 K-l •
boiling water 0.75% • At 318 K 0.174 W m-1 K-l •
At 336 K 0.210 W m-1 K-l a
Impact strength
(lzod notched bar test) 11.0 J m-1 of notch Compressive properties
Modulus of elasticity
Shear properties 3.48 GPa (504 ksi)
Failure stress 61 MFa (8.85 ksi)
Modulus of elasticity 1.27 GPa (184 ksi)
Tensile properties
Modulus of elasticity 3.24 GPa (470 ksi)
80

120
16 ~
100 ::!: 60
U>
U>

80 I!!
~ ~ ti5
:E U> 40
U>
U> 60 8
U>
e
I!! ti5
ti5 40
28d @23°C
4
20 8d @ 23"C

0 0
0 2 3 4 5

Tensile strain % Compressive strain %

 Property data for cured epoxy resin systems 71

Table 3.9 An aromatic amine-cured epoxy resin system13

Resin system constituents Parts by weight

Epoxy: DGEBA, e.g. Epon 826 (Shell) 100
Diluent: BDE, e.g. RD-2 (Ciba-Geigy) 25
Curing agent: MDA-MPDA eutectic, e.g. Tonox 60-40 (UniRoyal) 29
Cure cycle: 3 h at 60°C (140°F) + 2 h at 120°C (248°F)
Water absorption, wt. gain
after 6 h in boiling water 0.93%
1.2 Pa s (1200 cP)
Glass transition temperature
T1Ille to reach 2.0 Pa s 6h
Gel time for a 30-g mass at
25°C (77°F) 23h Heat distortion temperature at
1820 kPa (264 psi) 121°C (250°F)
Density at 25°C (77°F)
Uncured 1.15 Mgm-3
Cured 1.21 Mgm-3 Coefficient of linear thermal expansion
from 298 to 3755 K 6.81 x 10-5 °C-l
Volumetric shrinkage
After gelation 3.7% Specific heat
After cure 5.4% At363K 1.54 x 103 J kg-1 K-l
At424K 1.71 X 103 J kg-1 K-l
Tensile properties
Modulus of elasticity 2.68 GPa (389 ksi) Coefficient of thermal conductivity
At 325 K 0.243 W m-1 K-l
At 356 K 0.244 W m-1 K-l
100r_---r--~~--~----r_--~
14 At 389 K 0.256 W m-1 K-l
Shear properties
12 Failure stress 52 MPa (7.54 ksi)
80
10 Compressive properties
Maximum stress 111 MPa (16.1 ksi)
8!. 60
Strain at maximum stress 8.0%
:::i!: 8
/I)
/I) Modulus of elasticity 2.9 GPa (420 kSi)
~ 6
Ci.i 40
4
20
2

0
2 4 6 8 10
Strain %

epoxy resin formulations are often available procedures exist, e.g. as from the American
from resin suppliers. Data for new or unusual Society for Testing and Materials, they should
formulations must be generated by the user. be followed. Where they do not exist, litera-
Whether the data are generated by the resin hire references are helpful. Table 3.11 lists
supplier or the user, it is important that stan- some commonly tested properties and the
dard test procedures be followed. This will standard methods (American Society for
ensure that the resin systems can be compared Testing and Materials) describing the tests.
on an equal basis. Where standardized test
72 Epoxy resins

Table 3.10 An anhydride-cured epoxy resin system l5

Resin system constituents Parts by weight

Epoxy: DGEBA, e.g. Epon 828 (Shell) 100
Curing agent: NMA 90
Accelerator: BDMA 1
Cure cycle: 3 h at 120°C (248°F) + 24 h at 150°C (302°F)

Viscosity at 27°C (81°F) 1.78 Pa s (1780 cP)

Tune to reach 100 Pa s (1000 cP) 5-6 days
Pot life of a 500 g mass at 23°C (73°F) 4-6 days
Heat distortion temperature at 1820 kPa (264 psi) 121°C (250°F)
Solvent absorption, wt. gain
After 24 h in boiling water 0.67%
After 3 h in boiling acetone 1.9%
Tensile properties At 23°C (73°F) At 100°C (212°P)
Maximum stress 72.4 MPa (10.5 ksi) 46.2 MPa (6.70 ksi)
Strain at maximum stress 2.7% 7.2%
Modulus of elasticity 3.45 GPa (500 ksi) 1.38 GPa (200 ksi)

Table 3.11 Standard test methods for cured epoxy resin systems

Property ASTM
Standard Method
Physical and chemical properties:
Specific gravity D792
Chemical resistance D543
Water absorption D570
Light and water exposure D1499
Electrical properties:
Volume resistivity D257
Surface resistivity D257
Dielectric strength D149
Dielectric breakdown voltage D149
Permittivi~ dielectric constant D150
Dielectric loss D150
Thermal properties:
Heat deflection temperature D648
Glass transition temperature D4065
Coefficient of linear thermal expansion D696
Coefficient of thermal conductivity C177
Mechanical properties:
Tensile modulus and strength D638
Compressive modulus and strength D695
Flexural modulus and strength D790
Impact resistance D256
Dynamic mechanical properties:
Storage modulus D4065
Loss modulus D4065
Transition temperature D4065
 References 73

REFERENCES 17. Lee, W.I., Loos, AC and Springer, G.S. 'Heat of

reaction, degree of cure and viscosity of
1. Lee, H. and Neville, K. Handbook of Epoxy Resins,
Hercules 3501-6 resin,' J. Compos. Mater.,
Classical Re-issue, McGraw-Hill, New York,
1982,16, 510-520.
1982.
18. Springer, G.S., 'Compaction and consolidation
2. Goodman, S.H. 'Epoxy Resins', ch. 6 in
of thermoset and thermoplastic composites,'
Handbook of Thennoset Plastics, (S.H. Goodman, Compos. Polym., 1990,3, 155-175.
Ed.), Noyes Publications, Park Ridge, New
19. Mijovic, J. and Ott, J. 'Modeling of chemorheol-
Jersey, 1986, pp. 132-182.
ogy of an amine-epoxy system of the type used
3. Tanaka, Y. and Bauer, RS. 'Curing Reactions,'
in advanced composites,' J. Compos. Mater.,
ch.3 in Epoxy Resins: Chemistry and Technology, 1989,23,163-185.
(CA May, Ed.), Marcel Dekker, Inc., New York,
20. Mijovic, J. and Wang, H.T. 'Cure kinetics of neat
1988, pp. 285-463. . and graphite-fiber-reinforced epoxy formula-
4. Fischer, M., Lohse, F. and Schmld, R
tions,' J. Appl. Polym. Sci., 1989, 37, 2661-2673.
'Struktureller Aufbau and Physikalisches
21. Product data sheet for Araldite RD-2, Ciba-
Verhalten venetzter Epoxiharze,' Makromol. Geigy, Ardsley, New York, 1979.
Chern., 1980,181, 1251-1287.
22. Penn, 1.5. and Chiao, T.T. 'A long pot-life epoxy
5. Dusek, R and Bleha, M. 'Curing of epoxide
system for filament winding,' in Proceedings of
resins: model reactions of curing with amines,' J.
the 7th National SAMPE Technical Conference,
Polym. Sci., Polym. Chern. Ed., 1977, 15, 2393-2400. Albuquerque, New Mexico, October 14-16,
6. Fisch, W. and Hofmann, W. 'Uber den 1975, Vol. 7 (1975), pp. 177-187.
Hartungsmeehanismu. der Athexylinharze,' J.
23. Chiso, T.T., Jessop, E.S. and Penn, 1.S.,
Polym. Sci. 1954,12,497-502.
'Screening of epoxy systems for high perfor-
7. Morgan, RJ. and Mones, E.T. 'Cure reactions,
mance filament winding applications,' in
network structure, and mechanical response of
Proceedings of the 7th National SAMPE Technical
diamino diphenyl sulfonecured tetraglycidyl
Conference, Albuquerque, New Mexico, October
4,4'-diamino diphenyl methane epoxy matri-
14-16,1975, Vol. 7 (1975), pp. 167-175.
ces,' J. Appl. Polym. Sci., 1987,33,999-1020.
24. Jay, RR 'Direct titration of epoxy compounds
8. Resins and Hardeners Product Guide, Ciba-Geigy
and Aziridines,' Anal. Chern, 1964, 36, 667-668.
Corporation, Hawthorne, New York, 1990.
25. Jahn, H. and Goetzky, P. 'Analysis of epoxides
9. Epoxy Resin Manual: Industry Edition, Dow
and epoxy resins,. ch. 13 in Epoxy Resins:
Chemical Company, Midland, Michigan, 1988.
Chemistry and Technology, (CA. May, Ed.),
10. Specification Guide for Epon and Eponol Epoxy
Marcel Dekker, Inc. New York, 1988, pp.
Resins and Epon Curing Agents, Shell Chemical 1049-1087.
Company, Houston, Texas, 1990. 26. Chou, C.T. and Penn, 1.S. 'Mechanism by which
11. Chiao, T.T., Jessop, B.S. and Newey, H.A. 'A
an orthocarbonate reduces residual stress in a
moderate- temperature-curable epoxy for composite,' J. Compos. Mater., 1992, 26, 171-184.
advanced composites,' SAMPE Quart., 1975, 6,
27. Carpenter, J.F. and Bartels, TT,
38-42.
'Characterization and control of composlte
12. Chiao, T.T. and Moore, R1. 'A room-tempera-
prepregs and adhesives,' in Proceedings of the 7th
ture-curable epoxy for advanced fiber
National SAMPE Technical Conference,
composites,' 29th Ann' Cont, Reinforced Albuquerque, New Mexico, October 14-16,
Plastics/Composites Institute, SPI, Washington, 1975, Vol. 7 (1975), pp. 43-52.
D.C, February 5-8, 1974, section 16-B.
28. Penn, 1.5. and Chiao, T.T. 'Epoxy resins,' ch. 15
13. Chiao, T.T., Jessop, E.S. and Newey, H.A. 'An
in Handbook of Composites, (George Lubin, Ed.),
epoxy system for filament winding,' SAMPE
Van Nostrand Reinhold, New York, 1982, pp.
Quart., 1974, 6,28-32. 57-88.
14. Product data sheets, Ciba-Geigy, Hawthorne,
29. Dannenberg, H. 'Determination of functional
New York, 1993.
groups in epoxy resins by near infrared spec-
15. Epoxy Resins for Casting, Shell Chemical
troscopy,' SPE Trans., 1963, 3, 78-88.
Company, New York, 1967.
30. Morgan, R 'Structure-property relations of
16. Springer, G.S., 'Resin flow during the cure of epoxies used as composite matrices,' Adv.
fiber reinforced composites,' J. Compos. Mater., Polym. Sci., 1985, 72, 1-43.
1982,16,400-410.
74 Epoxy resins

31. Yalof, S. and Wrasidlo, W. 'Crosschecking 40. Day, D.R., Lewis, TJ., Lee, H.L. and Senturia,
between dielectric measurements, DTA, and S.D. 'The role of boundary layer capacitance at
other methods of thermal analysis in research blocking electrodes in the interpretation of
and production,' J. Appl. Polym. Sci., 1972, 16, dielectric cure data in adhesives,' J. Adhesion,
2159-2173. 1985, 18, 73-90.
32. Delmonte, J. 'Electric properties of epoxy resins 41. Micromet Instruments, Inc., Cambridge,
during polymerization,' I. Appl, Polym. Sci. 1959, Massachusetts, Technical literature, 1991.
2,108-113. 42. Ciriscioli, P.R and Springer, G.S. 'Dielectric cure
33. Lewis, A. 'Dynamic mechanical behavior dur- monitoring - a critical review,' SAMPE I., 1989,
ing the thermosetting curing process,' SPE 25,35-42.
Trans., 1963, 3, 201-212. 43. Sung, C.s.P., Pyun, E. and Sun, H.L.
34. I<reahling, RP. and Kline, D.E. 'Thermal con- 'Characterization of epoxy cure by UV-visible
ductivity, specific heat, and dynamic and fluorescence spectroscopy: azochromic
mechanical behavior of diglycidyl ether of labeling approach,' Macromolecules, 1986, 19,
bisphenol A cured with m-phenylene diamine,' 2922-2932.
I. Appl. Polym. Sci., 1969,13,2411-2425. 44. Sung, e.S.P. and Mathisen, R 'Cure characteri-
35. Arridge, R and Speake, J. 'Mechanical relax- zation of an epoxy network by fluorescence
ation studies of the cure of epoxy resins: 1. behavior of trans-diaminostilbene,' Polymer,
Measurement of cure,' Polymer, 1972, 13, 1987,28,941-945.
443-449. 45. Compton, D.A., Hill, S.L., Wright, N.A. et al. 'In
36. Kline, D.E. 'Dynamic mechanical properties of situ FTIR analysis of a composite curing reac-
epoxy resins during polymerization,' I. Appl. tion using a Mid-Infrared Transmitting Optical
Polym. Sci., 1960, 4, 123. Fiber,' Appl. Spectros. 1988,6,972-979.
37. Babayevsky, P. and Gillham, J. 'Epoxy ther- 46. Young, P.R, Druy, M.A., Stevenson, W.A. and
mosetting systems: dynamic mechanical Compton, D.A. 'In situ composite monitoring
analysis of the reactions of aromatic diamines using infrared-transmitting optical fibers,'
with the diglycidyl ether of bispenol A,' I. Appl. SAMPE J., 1989, 25, 11-16.
Polym. Sci., 1973, 17, 2067-2088. 47. Myrick, M.L., Angel, S.M., Lyon, R.E. and Vess,
38. Enns, J.B. and Gillham, J.K. 'The time-tempera- T.M. 'Epoxy cure mOnitoring using fiber-optic
ture transformation (TTf) cure diagram: Raman spectroscopy,' SAMPE I., 1992, 28,
modeling the cure behavior of thermosets,' J. 37-42.
Appl. Polym. Sci., 1983,28,2567-2591. 48. Mijovic, J., Kenny, J.M., Nicolais, L. and
39. Senturia, S.D. and Sheppard, N.F. 'Dielectric Pejanovic, S. 'Present and future trends in in-
analysis of thermoset cure,' Adv. Polym. Sci., situ monitoring of processing of advanced
1986,80,1-47. composites,' SAMPE J., 1992, 28, 39-46.
HIGH TEMPERATURE RESINS 4
Hugh H. Gibbs

4.1 INTRODUCTION too much from their high temperature mechan-

The high temperature resins discussed in this ical properties and thermal-oxidative stability.
chapter are defined as a family of aromatic As a result of a extensive work on the part of
polyimides having glass transition tempera- polymer scientists, mostly within the USA, a
tures (T) greater than 316°C (600 0 P). Other
wide variety of products has been developed
• g
resm systems such as the bis-maleimides and possessing various trade-offs between their
various aromatic thermoplastics (including processibility and properties. In writing this
lower Tg thermoplastic polyimides) are dis- chapter the author has attempted to explain the
evolution of high temperature polyimides as
cussed in Chapters 5 and 6 respectively.
Over the years it has been found that the key matrix resins for advanced composites together
to achieving outstanding high temperature with a description of their processing character-
mechanical properties and thermal-oxidative istics (where available), physical properties and
stability is to have a polymer made with aro- long term high temperature performance.
matic heterocyclic repeat units where there is a
minimum aliphatic content (e.g. aliphatic C-H 4.2 CONDENSATION POLYIMIDE
and C=C groups). Such groups can contribute CHEMISTRY
to thermal-oxidative instability. Although
many types of aromatic heterocyclic polymers 4.2.1 GENERAL COMMENTS ON
are possible, one type, polyimides, has turned CONDENSATIONPOLYMrnIDZATION
out to be the most commercially successful.
The highly aromatic character achievable in Studies carried out in DuPont in the 1950s and
such polymers is the reason behind their ther- 1960s established that polyimides can be pre-
mal-oxidative stability. Also, provided that pared by the reaction of an aromatic
flexibilizing linkages in the monomers are kept dianhydride and an aromatic diamine in a
to a minimum, the inherent rigidity of the polar solvent such as dimethyl acetamide or N-
repeat units results in the high T which is methyl-2-pyrrolidone (NMP). This is
essential if an adequate level of higt tempera- illustrated in Fig. 4.1 for the polyimide based
ture mechanical property retention is to be on pyromellitic dianhydride (PMDA) and
achieved. 4,4'-oxydianiline (aDA). The intermediate
Over the past 25 years various strategies polyamide acid solution is the basis for
have been developed to introduce processibil- DuPont's Pyre ML® wire enamel. During the
ity into aromatic polyimides without detracting second step, the imidization of the polyamide
acid, 2 moles of water are eliminated per repeat
unit. Two other DuPont products, Kapton®
Handbook of Composites. Edited by S.T. Peters. Published polyimide film and Vespel® SP polyimide pre-
in 1998 by Chapman & Hail, London. ISBN 0 412 54020 7 cision parts are also based on this chemistry.
76 High temperature resins

°
o~=O=~::o
I
°
H2NO°o-NHZ (::~~)F~-('rLHNOOo-NHt-
!
+
n HO-C~\:-OH
°
oPyromellitic Dianhydride 4 4' Ox dianilin"
,- Y e ° Heat
II
0
II PMDAlODA Polyamide Acid n
(PMDA) (ODA)

t
(-2H20)

<!-('r!>o(;OIV:t)00-+'
o
lI~rr
0
- - n

PMDAlODA Polyimide

Fig. 4.1 Typical reaction sequence for a polyimide from a dianhydride and a diamine.

It turns out that the particular polyimide temperatures in excess of IOOoe (212°F) are
shown in Fig_ 4_1. is unsatisfactory for the pur- reached. Thus high solids monomeric solu-
pose of making prepregs for high temperature tions are possible which are ideal for
composite parts. For one thing the prepregging. If, on the other hand, the dianhy-
PMDA/ODA polyimide is too intractable, dride is the commercially available starting
having no detectable Tg or melting point material then it must be converted in situ to
below its decomposition temperature which is the open ring diester diacid form by prereact-
well in excess of soooe (932°F). In addition, the ing with an alcohol such as ethanol or
polyamide acid solutions which are immedi- methanol.
ately generated by dissolving the diamine in a
suitable solvent and then adding the reactive
4.2.2 CHEMISTRY OF SKYBON1)®
dianhydride, are unsuitable for prepregging.
Ideally monomeric solutions are preferred that Skybond from Monsanto is a product which
possess modest viscosities even when the has been available commercially since the
solids contents are in the S0--60 wt% range. mid-1960s_ The chemistry of Skybond is illus-
This desired combination of properties cannot trated in Fig. 4.2_ The relatively low cost
be achieved in polymeric solutions such as dianhydride, 3,3',4,4'-benzophenonetetracar-
polyamide acids. Instead, the solutions boxylic dianhydride (BTDA), is first
rapidly become unacceptably viscous when prereacted with ethanol using NMP as the sol-
solids contents in excess of 15--20% range are vent Then, if m-phenylenediamine (MPD) is
reached thus making them too difficult for added to the solution, Skybond 700 results. If
prepregging using commercially available 4,4'-methylenedianiline (MDA) is employed,
equipment then Skybond 703 is produced. During cure
In order to prepare polyimide binder solu- the application of heat causes the elimination
tions (or polyimide precursor solutions as they of the solvent along with 2 moles of water and
are sometimes called) it is necessary to have 2 moles of ethanol per repeat unit to produce
the aromatic dianhydride in either one of the the polyimide. The molecular weight initially
two possible open ring forms (tetraacid or achieved will depend on the monomer imbal-
diester diacid). If the tetraacid form is com- ance employed. It has long been speculated
mercially available then the binder solution that during the cure process branching can
can be made directly because there is essen- occur by the reaction of amine end-groups
tially no reaction with the diamine until with the bridging carbonyl group of the BTDA
 Condensation polyimide chemistry 77

3,3' ,4,4'-benzophenonetetracarboxylic Diethyl ester of 3,3',4,4'-benzophenone

dianhydride tetracarboxylic acid
BTDA BTDE

m-phenylene diamine 4,4'-methylene dianiline

Skybond 700 Skybond 703

Fig. 4.2 Chemistry of Skybond® binder solutions.

moiety leading to branching and intractability reactive and present in stoichiometric propor-
of the matrix resin. This is probably one of the tions, extremely high molecular weight
reasons why it is difficult to fabricate low void polyimide is ultimately produced. Also, since
composites using this type of chemistry. the bridging hexafluoroisopropylidene group
is inert under normal cure conditions, the poly-
imide produced is essentially linear and in that
4.2.3 CHEMISTRY OF NR-lS0
sense thermoplastic.
The formation of DuPont's monomeric binder
solution NR-150 (used to make Avimid® N
4.2.4 CHEMISTRY OF 3F /36F POLYIMIDES
prepreg) is illustrated in Fig. 4.3. The
2,2-bis(3',4'-dicarboxyphenyl) hexafluoro- One of the keys to success in producing an
propane (6F tetraacid) is dissolved in a suitable essentially linear condensation polyimide is to
solvent (e.g. ethanol, diglyme or NMP) along have a chemically inert flexibilizing linkage in
with a 95/5 mixture of p-phenylenedi- the dianhydride moiety. As indicated in
amine/m-phenylenediamine to form a low Section 4.2.3 one of the ways of doing this is to
viscosity, stable monomeric solution1 suitable employ a hexafluoroisopropylidene bridging
for prepregging. During the cure both solvent group. Since the phenyl group is also very
and 2 moles of by-product water per repeat inert another approach is to replace one of the
unit are eliminated to initially form the tran- CF3 groups with a phenyl group3. This results
sient intermediate polyamide acid. According in the so-called 3F dianhydride. When the 3F
to the studies of Sonnett et aP, although a low monomer is used along with PPD the poly-
concentration of amide acid persists for a rela- imide produced is called 3F-PPD polyimide.
tively long time period during the early stages If, on the other hand, a mixture of the 3F and
of cure, most is rapidly converted to the imide 6F monomers are employed along with the
form by the thermally induced elimination of same diamine the copolyimide is designated
water (2 moles per repeat unit). Since the 36F-PPD.
tetraacid and diamines are very pure, highly
78 High temperature resins

I
6
o CF 3 0
b
/I
NH2

HO-C~I~C-OH
HO-C-V
/I
CF 3 V-C-OH
/I
+
+ ~
6 .# NH z
+ Solvent

° 0 NH2
95/5 mixture PPDIMPD

Monomeric binder solution

j Heat
(-2H 20)
(- Solvent)

o CF3 0
II I II-G~
HN-C~T~C-NH '1_~
HO-CV

°
/I
CF3 VC-OH
I
0 n
Intennediate transient polyamide acid

Heat
J (-2H20)
(- Solvent)

n
NR-150 Polyimide

Fig. 4.3 NR-1S0 polymerization chemistry.

4.3 ADDmON POLYIMIDE CHEMISTRY difficult for these volatiles to escape. Therefore,
ideally one would want to eliminate all of the
4.3.1 OVERALL CHEMISTRY OF ADDmON volatiles prior to pressurization. In order to
POLYIMIDE PRECURSOR SOLUTIONS accomplish this in a conventional condensation
One of the important concerns in the cure of polyimide it is necessary to have a high enough
monomer imbalance so that the polymer mole-
conventional condensation polyimides is the
proper management of the evolution of volatile cular weight and melt viscosity will be low
enough to allow for complete consolidation.
by-products. During the early stages prior to
The problem is that this molecular weight is
pressurization the composite is generally some-
generally so low that the matrix resin properties
what porous and volatiles can readily diffuse
out. However, once pressure has been applied are adversely affected. Properties such as
and a low void state is achieved the diffusivity strength, toughness and T will be much lower
than desired. g
dramatically decreases making it much more
 Addition polyimide chemistry 79

A common way of making a low void com- 4.3.2 CHEMISTRY OF PMR-15

posite is to produce in situ an imide oligomer
The first approach to a commercially successful
having a low enough molecular weight so that
addition polyimide was PMR-15. The basic
it has good melt flow and readily consolidates
PMR (Polymerization of Monomeric Reactants)
when pressure is applied and then, with further
chemistry was originally invented and devel-
heating, can go on to produce much higher
oped at NASA Lewis Research ~ent~. ~
molecular weight polymer without the further
illustrated in Pig. 4.4 a monomenc solution IS
evolution of volatile by-products. This can be
first prepared consisting of the dimethyl ester
accomplished through the use of reactive end-
of BIDA (BIDE), the diamine, MDA and the
groups which are capable ~f reacting with ~ne monomethyl ester of nadic anhydride (NE), the
another in various ways WlthOut the evolution
reactive end-capping agent. The monomer
of volatiles. However, it must be pointed out
ratios are n:n+1:2 respectively. Because the high
that such improvements in processibility are
solids solution is monomeric it is ideally suited
only achieved with some sacrifices in other
to prepregging. As a result of the application of
properties. Because of the presence of cross-
heat, solvent is eliminated along with the water
links the toughness can be adversely affected.
and alcohol of imidization to produce the inter-
Also, the kinds of linkages produced during
mediate imide oligomer having a formulated
these addition polymerization reactions are
molecular weight of about 1500. This normally
generally aliphatic in nature thus making them
occurs between 121°C (250 0 P) and 232°C
much more prone to thermal-oxidative attack. (450 0 P). At this point essentially all of the
Many approaches, which are now discussed,
volatile by-products have been eliminated. The
have been taken in order to achieve various
stage is now set for pressurization. At some-
trade-offs in processibility against properties.
o 0
~ II II
rj1'r~OCH3 ~~ H3~~~~~=~3
l!j)-.-OH 2-V
+ H N
V- NH2 +
II
o
II
o 0
2 Moles NE 2MolesMDA 1 MoleBTDE
PMR-15 Binder Solution

j Heat
- 2 Moles H2O
- 2 Moles Methanol + Methanol Solvent

- 0 0
II II 1:)0 II

rj1'rlN-O'D-<~~~~ J(~
2 0 2

~. II U ~
o

J-
0 PMR-150UgOlllrr 0

bl
0
~ W II W

~.- V-N(~~~:::N
II II
o o 0 2
CrossliDked PMR-15
Fig. 4.4 PMR-15 polymerization chemistry.
80 High temperature resins

what higher temperatures the imide oligomer 4.3.3 CHEMISTRY OF PMR-II

undergoes melting and, if pressure is applied, During the early 1970s it became clear that
consolidation to form a low void structure although composites based on PMR-15 pos-
occurs. At temperatures in the 275-316°C sessed good enough properties to make them
(527-600°F) range ring opening of the nadic a commercial success there was still room for
end-groups takes place in a reverse Diels-Alder
improvement. For instance, the relatively hi~h
type of reaction first discovered by Lubowitz9,10
aliphatic content which came from the nadlc
in 1970 and a complex series of reactions takes end-groups and the methylene group of the
place involving these end-groups leading to MDA contributed to thermal-oxidative insta-
high molecular weight polymer without the bility. Also, the relatively high cross-link
further evolution of volatile by-products. density in a polyimide with a formulated mol-
Hence the low void state originally achieved ecular weight of 1500 resulted in a rather
during the initial consolidation is maintained brittle matrix resin. In order to overcome these
and a high quality laminate usually results. deficiencies PMR-II was inventedll~14 (Fig 4.5).
For PMR-II polyimides the 6F dianhydride is
used in place of the BTDA and is converted in
situ to the diethyl ester diacid derivative

PMR-II imide prepolymer

V-CAP imide prepolymer

A-~2
Y I
CH2 -CH2
A'

CYCAP imide prepolymer

AFR700B imide prepolymer

Fig. 4.5 Structures of 6F based addition polyimide prepolyrners.

 Addition polyimide chemistry 81

(6FDE) by pre-reacting with ethanol. The sin- 4.3.6 CHEMISTRY OF AFR700B

gle ring diamine, PPD, is employe~ in place?f With any of the reactive end-groups discussed
the MDA. The end-capping agent IS still nadlC
thus far in this chapter the concentration of
anhydride. The formulated molecular weight end-capping agent has always been such that
was increased from 1500, which is what it is in
there are end-caps on both chain ends.
PMR-15, to the 3000-5000 range (yields
However, by adding just enough end-capping
PMR-II-30 and PMR-II-50 respectively). While
agent to cap one chain end only and assuming
the polymerization is basically the same in
that high molecular weight polymer can still
both PMR-II and PMR-15, since the aliphatic
somehow be obtained under reasonable pro-
content has been greatly reduced, significant cessing conditions, a further significant
improvements in thermal-oxidative stability improvement in thermal-oxidative stability
and toughness can be realized.
should be possible simply because the concen-
tration of unstable groups will have been cut
4.3.4 CHEMISTRY OF V-CAP in half. This is exactly what has been done
with AFR700B based on the work of Serafini et
Assuming that the 6F /PPD backbone has near al. at TRW21-23.
optimal therrna1-oxidative stability and formu-
In AFR700B the ratio of 6FDE:PPD:NE is
lated molecular weights in the 3000-5000 range
8:9:1 and the formulated molecular weight is
are about right from the stand-point of melt about 4400 (see Fig. 4.5 for its structure)24. This
flow, if further improvements in stability are to
means that the nadic end-group component is
be realized then one approach would be to only 3.7% of the overall weight compared with
make changes in the end-group chemistry. One 22% for PMR-15 and 6.5% for PMR-II-50).
effective method is the replacement of nadic
During cure of AFR700B composites it is nec-
anhydride as the end-capping agent with p-
essary not only to achieve the nor~al
aminostyrene (PAS)15-18. The imide prepolymer
imidization and nadic end-group couplmg
that is produced is illustrated in Fig. 4.5. In this
reactions but also other unspecified reactions
case there is only one weak bond per end-cap
involving the amine end-groups undoubtedly
compared with eight for the nadic end-cap.
occur leading to the formation of the desired
During the addition polymerization phase of
strong, stiff, tough, high Tg polyirnide.
the cure free radical polymerization of the vinyl
(-CH=CHz) end-groups occurs leading to high
molecular weight polymer. 4.3.7 CHEMISTRY OF TRW-R-8XX
The most recent addition to the growing fam-
4.3.5 CHEMISTRY OF CYCAP ily of commercially available polyimides
comes from TRW and has been designated
Another approach to a more stable polymer TRW-R-8XX25. Although the chemical struc-
has been to replace the nadic end-capping
ture of this polyimide has not yet been
agent with 2-amino-p-cyclophane (APC)19,20.
disclosed it is reported to be a condensa-
The structure of the imide prepolymer con-
tion/ addition polyirnide based on relatively
taining the CYCAP end-groups that is initially
low cost monomers, making its cost compara-
produced is illustrated in Fig. 4.5. This type of
ble to that of PMR-15. It is reported to be free
end-capping agent has only two weak bonds
of the carcinogen, MDA.
per end-cap. Thermolytic cleavage of the
-CH -CH - groups during the final stages of
cure2produces bi-radicals (-CH2· ·CH2-) which
undergo coupling with other bi-radicals to
build up the molecular weight.
82 High temperature resins

4.4 COMMERCIAL AVAILABILITY OF BINDER Table 4.1 Typical properties of neat cured NR-150 38
SOLUTIONS
Property Units Value
Most of the binder solutions described in this
Tg DC 350-370
chapter are not commercially available but OF 662-700
rather are prepared by the prepregger on an g crn-3 1.43-1.45
Density
as-needed basis just prior to prepregging. One Coefficient of thermal °C-l 5.6 X 10-5
notable exception is Monsanto's Skybond. The expansion °F-l 3.1 X 10-5
monomeric solutions (Skybond 700 and 703) Char yield % 60
have solids contents in the 45-52% range and Tensile strength MPa (ksi) 110 (16)
solution viscosities of 3000-7000 poise26 • Elongation, RT % 6
316°C (600 DF) 65
Fracture toughness J m-2 (in lb in-2) 2000 (11.4)
Rochwell hardness
4.5 COMMERCIAL AVAILABILITY OF
(E scale) 70
PREPREGS
At the time of writing of this chapter all of the
different polymide prepreg systems were com- Table 4.2 Typical properties of neat cured PMR-15
mercially available from one prepregger or
Property Units Value Reference
another. However, the reader should appreci-
ate the fact that as time goes by some prep reg T DC (OF) 335 (635) 28
Dgenslty
• gcm-3 1.30-1.32 28
types will disappear from the market place
Coefficient of °C-l 16 X 10-6 35
and others with an improved balance of pro- DF-l 28 X 10-6
thermal 35
cessing characteristics, properties and expansion
economics will come along to take their place. Tensile strength MPa (ksi) 55 (8.0) 35
Also, some companies will go out of business Tensile modulus MPa (ksi) 3200 (470) 35
or will be bought out by other companies as Elongation % 1.5 36
consolidation in the industry occurs. Compressive MPa (ksi) 110 (16) 35
Therefore, if a given type of prepreg is yield strength
Compressive MPa (ksi) 186 (27) 35
required the reader should contact their
strength
favorite prepregger. If that particular company Equilibrium % 4.2 28
does not offer the product the reader requires moisture
then advice should be obtained as to where absorption
such prepreg could be commercially pur- Fracture Jm-2 500 (2.86) 37
chased, if at all. Another fruitful source of toughness (in lb in-2)
information would be to search the Internet.
In neat resin thermal-oxidative stability stud-
ies carried out by Scola27 it was found that
4.6 NEAT RESIN PROPERTIES
after 24 h at 316°C (600°F) the NR-150 resin
The neat resin mechanical properties for cured had lost 9% of its weight compared with about
NR-150 and PMR-15 are summarized in Tables 76% for the PMR-15. Neat AFR700B, which has
4.1 and 4.2 respectively. One of the important been post-cured under nitrogen, has been
differences between these two resin systems is reported to have a room temperature tensile
their toughness. Cured NR-150 has been strength of 93.8 MPa (13.6 ksi) with 18% reten-
found to be dramatically tougher (2000 J m-2 tion of this value at 371°C (700DP). No
fracture toughness) compared with a value of properties were available for Skybond, PMR-II,
500 Jm-2 for the cured PMR-15. Another signif- V-CAP and CYCAP. Very little information
icant difference is thermal-oxidative stability. was available for the 3F-PPD and 36F-PPD
 Processing characteristics 83

polyimides except for neat resin densities (1.35 because such data was either considered pro-
and 1.42 g cm-3 for the 3F and 36F resins prietary or classified as secret by the
respectively) and T of 365-370°C (689-698°F) Government Laboratories. In other cases,
for both systems ~ the as-molded state and where some processing information was avail-
405-41O°C (761-770°F) for the post-cured able in the open literature, very little was
state3• TRW-R-8XX has been reported25 to yield usually said about the quality of the part pro-
polyimides having T in the 400-426°C duced by a given cycle so it is difficult to
(75G-800°F) range and ~omposite weight loss compare the processibility of one resin system
characteristics at 371°C (7000 P) up to 10 times with another since the quality of the laminates
better than PMR-15. All of the resin systems produced is generally unknown. To make mat-
described in this chapter appear to possess ters worse there are also the issues of the
good strength and stiffness. Thus, provided processing characteristics of thick compared
that complete cure is achieved during process- with thin sections and how processing can be
ing and low void composites are produced handled, if at all, when both thin and thick sec-
possessing good fiber/matrix adhesion, high tions are present simultaneously in a given
levels of composite mechanical properties part. It is, therefore, of paramount importance
should be obtained with good retention (at for workers in this field to have as clear an
least 50%) of these properties to just below understanding as possible of the chemistry
their T g • involved at every stage of the cure so that they
can quickly and efficiently develop the opti-
mum cure cycle for a given part. The cure
4.7 PROCESSING CHARACTERISTICS
cycles presented below will give the reader an
approximate idea of the kinds of conditions
4.7.1 GENERAL COMMENTS
that have been employed to produce a part.
One of the features that can clearly differenti-
ate one resin system from another is the ease
4.7.2 SKYBOND PROCESSING CONDITIONS
with which a fully cured low void composite
can be produced having a specified fiber vol- The following autoclave cure cycle has been
ume. Also, although all polyimides can have recommended by Monsant0 26 for 12 ply
their Tg increased from post-c~re, some sys- (3.2 mm, 0.125 in thick) 181 style E-glass (soft
tems respond much more readily than others. A-t100 finish) fabric/Skybond 700 laminates:
In all cases there is a problem of properly man-
aging the release of a significant amount of • apply full vacuum;
• heat to 177°C (350 0 P) at 1.7-2.8°C/min
volatiles (normally 10-15% of the weight of
(3-5°F/min);
the prepreg). The ways in which this can be
accomplished can vary significantly from one • hold 5 min at 177°C (350 0 P);
• apply 0.69 MPa (100 psi);
resin to another depending on the chemistry
involved. Factors such as the techniques • hold 30 min;
• cool under pressure and vacuum.
employed during lay-up of the vacuum bag
assembly, pressure, heat-up rate, maximum In order to maximize high temperature prop-
cure temperature, vacuum application and erties it is recommended that a post-cure
intermediate holds all must be carefully con- should be carried out up to and including the
trolled and optimized for each system. expected use-temperature. For laminates of
Unfortunately the story on the processing of the type described above a post-cure is sug-
the various systems covered in this chapter is gested in which the part is heated to 200°C
very incomplete. In a majority of cases infor- (392°P), 225°C (437°F), 250°C (482°F), 300°C
mation on processing was simply not available (572°P), 325°C (617°P), 350°C (662°F) and
84 High temperature resins

371 °C (700°F) and held for 2 h at each temper- noted that pressurization does not occur
ature. It is recommended that if thicker until a temperature of about 238°C (460°F)
laminates are involved the post-cure cycle is reached. At that temperature essentially
should be extended. all of the solvent and imidization volatiles
For the same type of laminate based on have been eliminated. Normally PMR-15
Skybond 703 a similar autoclave cure cycle can laminates are subjected to an oven post-cure:
be employed. The only significant differences • heat from room temperature to 204°C
are a slower heat-up rate (1.1-1.7°C/min, (400°F) at 5.6°C/min (lO°F /min);
2-3°F /min) with the pressure being applied at • heat from 204°C to 288°C (400°F to 550°F) at
121°C (250°F) on the way to the final cure tem- 1.1 °C/min (2°F/min);
perature of 177°C (350°F). Also a similar post- • dwell at 288°C (550°F) for 1 h;
cure cycle is suggested in order to achieve • heat from 288°C to 316°C (550°F to 600°F) at
maximum heat resistance. 1.1°C/min (2°F/min);
While there is very little definitive informa- • dwell at 316°C (600°F) for 10-16 h;
tion on the void content of Skybond based • cool to room temperature at 2.8°C/min
laminates it is believed that they are generally (5°F/min) maximum.
in the 5-20% range. A most important feature
of this particular polyimide system is that the
4.7.4 PMR-II AND V-CAP PROCESSING
maximum autoclave processing temperature
CONDITIONS
is only 177°C (350°F). No other resin system
described in this chapter can make that claim. The autoclave cure cycles for PMR-II and
V-CAP based composites are similar to that of
PMR-15. The main difference is that the maxi-
4.7.3 PMR-15 PROCESSING CONDITIONS mum processing temperature has been
increased from 316°C to 371°C (600°F to
The following represents a typical autoclave
700°F). The following typical autoclave cure
cure cycle28 cited for PMR-15:
and post-cure cycles for graphite reinforced
• apply 7-21 kPa (1-3 psi) vacuum; PMR-II composites has been reportedl 4.15:
• raise autoclave temperature to 227°C
• apply full vacuum at room temperature;
(440°F) at 0.83-1.1°C/min (1.5-2.0°F/min);
• heat to 149°C (300°F) at 3.9°C/min
• at 163-177°C (325-350°F) apply full vacuum;
(7.0°F /min);
• dwell at 227°C (440°F) for 1-4 h depending
• hold for 30 min at 149°C, then apply
on part thickness (up to 2.8 mm, 0.11 in =
172 kPa (25 psi);
1 h, 2.8-6.4 mm, 0.11-0.25 in = 2 h,
• heat to 288°C (550°F) at 3.9°C/min
6.4-12.8 mm (0.25-0.50 in) = 3 h);
(7.0°F/min) with the pressure being
• raise temperature to 238°C (460°F) at
increased to 344 kPa (50 psi) at 177°C
1.1°C/min (2.0°F/min);
(350 0 P) and then to 1.38 MPa (200 psi) at
• hold at 238°C (460°F) while 1.38 MPa 232°C (450°F);
(200 psi) autoclave pressure is applied.
• heat from 232°C to 371 °C (450°F to 700°F) at
Do not hold longer than 15 minutes while 2.8°C/min (5°F/min);
pressure is being applied • cool under full pressure and vacuum to
232°C (450°F) slowly;
• raise temperature to 316°C (600°F) at
• cool from 232°C to room temperature
2.2-3.3°C/min (4-6°F /min);
rapidly.
• dwell at 316°C (600°F) for 3 h;
• cool to room temperature. Vent autoclave As with PMR-15, a post-cure is normally car-
pressure below 204°C (400°F) It should be ried out in a circulating air oven:
 Mechanical properties before and after air aging 85

• heat from room temperature to 260°C their high Tg (377-418°C, 710-785°F) the reten-
(500°F) at 20°C/min (36°F/min) tion of properties was excellent out to
• heat from 260°C to 385°C (700°F to 725°F) at temperatures as high as 360°C (680°F).
1°C/min (1.8°F /min) with 2 h holds at 316°C At the writing of this chapter the long term
(600°F), 343°C (650°F) and 20 h hold at 385°C. air aging characteristics of laminates of this
particular type had not been completed.
However, because of its all aromatic character,
4.7.5 PROCESSING CONDITIONS FOR OTHER
ultra-high molecular weight and the complete
RESIN SYSTEMS
absence of any aliphatic character from reac-
At the time of the writing of this chapter no tive end-capping agents it should air age well
unclassified processing information was avail- and the matrix in Avimid N-150 has been pre-
able for CYCAP, AFR700B, TRW-R-8XX or the viously shown to possess outstanding
3F /36F polyimides. The compression molding thermal-oxidative stability'l°.
conditions used by DuPont to make the lami-
nates whose properties are described in
4.8.3 3F /36F POLYIMIDES
Table 4.4 were not disclosed. However, details
concerning the autoclave processing of Although this family of all aromatic poly-
Avimid N have been previously discussed29 • imides is relatively new, preliminary data
indicates that high quality laminates possess-
ing good mechanical properties and excellent
4.8 MECHANICAL PROPERTIES BEFORE AND long term thermal-oxidative stability can be
AFTER AIR AGING produced. According to the work of Scola3
both the 3F-PPD and the 36F-PPD systems
4.8.1 SKYBOND resulted in G40-600 laminates having the
The most common type of reinforcement that expected good room temperature mechanical
has been employed with Skybond binders is properties (flex strength and short beam shear
E-glass fabric. The mechanical properties of strength) with excellent retention of these' dry'
as-molded and air aged laminates based on properties out to at least 371°C (700°F). Also,
Skybond 700 are summarized in Table 4.3. This after 100 h in air at 371 °C (700°F) these lami-
particular resin system has been tailored for nates retained at least 100% of these properties
extended exposures at temperatures up to with weight loss values ranging from 1.4% for
371 °C (700°F). For 343°C (650°F) applications the 36F-PPD resin and 2.4% for the 3F-PPD
Skybond 703 is recommended by the manu- polyimide. Surprisingly neat resin studies car-
facturer. It is interesting to note that in spite of ried out by Scola have similarly shown that
the relatively high porosity levels (5-20%) the 36F-PPD polyimide appeared to be some-
Skybond binders are still in demand for cer- what more stable than the 3F-PPD polymer.
tain specialty applications. For instance, after 100 h exposure at 371°C
(700°F), the 36F-PPD copolyimide had lost
about 2% of its weight compared with 3.2% for
4.8.2 AVIMID N the 3F-PPD polymer. Much more work needs
to be carried out on polyimides based on the
Some mechanical properties for compression
3F monomer before any final decision can be
molded/post-cured Celion G30-500 uniweave/
made as to its long term viability in the market
Avimid N laminates are tabulated in Table 4.4.
place and how it will ultimately compete with
All laminates had a very low void content
the 6F based polyimides such as Avimid N,
(<1 %) and, therefore, high levels of mechanical
PMR-II, V-CAP, CYCAP and AFR700B.
properties at room temperature. Because of
86 High temperature resins

Table 4.3 Mechanical properties of Skybond® 700/181 style E-glass laminates (Soft A-I 100 Finish) 26

Property High temperature, Vacuum bag

high pressure
Flex strength, MPa (ksi)
24°C (75 oF) 517-586 (75-85) 524-576 (76-84)
371 °C (700°F) after 0.5 h at 371 °C 310-414 (45-60) 152-221 (22-32)
371 °C (700°F) after 100 h at 371°C 138--241 (20-35) 138-166 (20-24)
316°C (600°F) after 500 h at 316°C 200 (29)
316°C (600°F) after 860 h at 316°C 138 (20)
316°C (600°F) after 1850 h at 316°C 76 (11)
288°C (550°F) after 2300 h at 288°C 283 (41)
288°C (550°F) after 4500 h at 288°C 221 (32)
288°C (550°F) after 9000 h at 288°C 103 (15)
Flex modulus, GPa (msi)
24°C (75°F) 22 (3.1) 19 (2.8)
299°C (570°F) after 335 h at 299°C 22 (3.1)
371 °C (700°F) after 100 h at 371°C 12 (1.8)
316°C (600°F) after 500 h at 316°C 18 (2.6)
316°C (600°F) after 860 h at 316°C 18 (2.6)
316°C (600°F) after1850 h at 316°C 14 (2.1)
288°C (550°F) after 2300 h at 288°C 18 (2.6)
288°C (550°F) after 4500 h at 288°C 20 (3.0)
288°C (550°F) after 9000 h at 288°C 14 (2.0)
Ultimate tensile strength, MPa (ksi)
24°C (75°F) 393 (57) 347 (50.1)
299°C (570°F) after 335 h at 299°C 290 (42)
24°C (75°F) after 100 h at 250°C (482°F) 336 (49)
24°C (75°F) after 100 h at 300°C (572°F) 332 (48.2)
Elongation, %
24°C (75°F) 1.9 2.0
299°C (570°F) after 335 h at 299°C 1.4
24°C (75°F) after 100 h at 250°C (482°F) 1.7
24°C (75°F) after 100 h at 300°C (572°F) 2.0
Weight loss, %
After 100 h at 371 °C (700 0 P) 3.0 <5
After 500 h at 316°C (600 0 P) 2.2
After 860 h at 316°C (600 0 P) 3.4
After 1850 h at 316°C (600°F) 7.9
After 2300 h at 288°C (550°F) 3.6
After 4500 h at 288°C (550 0 P) 5.0
After 9000 h at 288°C (550°F) 12.0
 Mechanical properties before and after air aging 87

Table 4.4 Mechanical properties of compression molded Celion® G30-SOO Uniweave/ Avimid® N laminates 39

No. ofplies Orientation Tg' °C (OF) Temperature, Property

0C (OF)
Short beam shear strength, MPa (ksi)
16 0° 418 (785) 24(75) 98.6 (14.3)
218 (425) 61.4 (8.9)
316 (600) 46.2 (6.7)
360 (680) 38.6 (5.6)
Flex strength, MPa (ksi)
16 0° 418 (785) 24 (75) 1344 (195)
316 (600) 731 (106)
360 (680) 565 (82)
Flex modulus, GPa (msi)
16 0° 418 (785) 24(75) 126 (18.0)
316 (600) 123 (17.9)
360 (680) 117 (17.0)
In-plane shear strength, MPa (ksi)
16 (±45°) 418 (785) 24 (75) 174 (25.2)
371 (700) 104 (15.1)
Tensile strength, MPa (ksi)
10 0° 377 (710) 24 (75) 1261 (183)
416 (780) 1027 (149)
Tensile modulus, GPa (msi)
10 0° 377 (710) 24 (75) 124 (18.0)
416 (780) 104 (15.1)
Open hole tensile strength, MPa (ksi)
10 0° 377 (710) 24 (75) 1027 (153)
416 (780) 854 (124)
Open hole tensile modulus, GPa (msi)
10 0° 377(710) 24 (75) 132 (19.1)
416 (780) 114 (16.5)
Tensile strength, MPa (ksi)
16 0,90,±45° 377 (710) 24 (75) 460 (66.7)
360 (680) 314 (45.5)
416 (780) 236 (34.3)
Tensile modulus, GPa (msi)
16 0,90,±45° 377 (710) 24 (75) 56.5 (8.2)
360 (680) 46.8 (6.8)
416 (780) 30.3 (4.4)
Open hole tensile strength, MPa (ksi)
16 0,90,±45° 377 (710) 24(75) 389 (56.5)
360 (680) 250 (36.3)
416 (780) 177 (25.5)
Open hole tensile modulus, GPa (msi)
16 0,90,±45° 377 (710) 24 (75) 54.4 (7.9)
360 (680) 46.8 (6.8)
416 (780) 38.3 (5.7)
Compressive strength, MPa (ksi)
16 0,90'±45° 377 (710) 24 (75) 389 (56.5)
Compressive strength, MPa (ksi)
16 0,90,±45° 377 (710) 24 (75) 458 (66.4)
360 (680) 312 (45.4)
416 (780) 270 (39.2)
Compressive strength, MPa (ksi)
16 0° 377 (710) 24 (75) 868 (126)
316 (600) 448 (65.1)
360 (680) 409 (59.4)
Interlaminar fracture toughness
J m-' (in lb in-')
16 0° 418 (785) 24 (75) 630 (3.6)
316 (600) 682 (3.9)
343 (650) 718 (4.1)
88 High temperature resins

Table 4.5 Mechanical properties of fiber reinforced PMR-15 laminates 35

Reinforcement
High strength High strength
Property 7781 Style E-glass fabric [>3447 MPa (500 ksi)] [>3447 MPa (500 ksi)]
(50-55 Vol. % fibers) Standard modulus Standard modulus
[228 CPa (33 msi)] class [228 CPa (33 msi)] class
Carbon fiber unidirectional tape Carbon fiber 8-harness
(57-63 Vol. % fibers) satin fabric
(55-60 Vol. % fibers)
Compressive
strength
MPa (ksi)
23°C (73°P) 517-586 (75-85) 827-965 (120-140) 552-689 (80--100)
288°C (550°F) 758-896(110--130) 414--552 (60-80)

Compressive
modulus
GPa (msi)
23°C (73°P) 28-34 (4--5) 97-117 (14-17) 62-76 (9-11)
288°C (550 0 P) 83-110 (12-16) 48-62 (7-9)

Flex strength
MPa (ksi)
23°C (73°P) 483-621 (70--90) 965-1103 (140--160)
288°C (550°F) 689-896 (100-130)
316°C (600 0 P) 414-552(60-80)

Flex modulus
GPa (msi)
23°C (73°F) 21-34 (3--5) 55-69 (8-10)
288°C (5500 P) 55-69 (8-10)
316°C (600°F) 21-34 (3--5)

Tensile strength
MPa (ksi)
23°C (73°F) 1241-1448 (180-210) 689-896(100--130)
288°C (5500 P) 1241-1448(180--210) 758-965(110-140)

Tensile modulus
GPa (msi)
23°C (73°P) 117-138 (17-20) 62-76 (9-11)
288°C (550°F) 103-124 (15-18) 62-76 (9-11)

Interlaminar
shear strength
MPa (ksi)
23°C (73°P) 62-76 (9-11) 55-69 (8-10)
288°C (5500 P) 34--48 (5-7)
316°C (6000 P) 34--48 (5-7)
 Mechanical properties before and after air aging 89

4.8.4 PMR-15 perature and either 232°C (450 0 P) or 288°C

(550 0 P) at heat-up and cool-down rates of
Prom the data cited in Table 4.5 it can be seen
278°C (532°P)/min (achieved by employing a
that PMR-15 based composites reinforced with
heated fluidized sand bed) significant microc-
either E-glass or graphite fibers are character-
racking was detected at 5000 thermal cycles
ized by good room temperature strength
and beyond. The data plotted in Fig. 4.6
properties with excellent retention of these
clearly indicate that cycling to the higher tem-
properties at temperatures in the 288-316°C
perature resulted in a much higher
(550-600 0 P) range.
concentration of microcracks. According to the
Over the past 20 years a great deal of effort
results plotted in Pig. 4.7 the development of
has gone into the study of the weight loss
~crocracks was found to have a very delete-
characteristics of PMR-15 composites at vari-
nous effect on the matrix-dominated
ous temperatures and air pressures and the
compressive strength of the graphite fabric
effects of these exposure conditions on the
laminates. However, GE also showed that the
laminate mechanical properties. One of the
fiber-dominated tensile strength was not
significant new property features to be uncov-
affected by the presence of the microcracks.
ered has been the effects of temperature
!here is every reason to believe that compos-
cycling on the development of microcracks
ites based on other resin systems will also
and the effects of these microcracks on the
undergo micro cracking from temperature
strength and stiffness properties. In acceler-
cycling and can be expected to exhibit
ated thermal cycling studies31 carried out at
improved microcracking performance to the
General Electric - Aircraft Engines in which
extent to which they possess improved tough-
the cycling was carried out between room tem-
ness compared with PMR-15.

35

-E
i
30
()
!
.
A
.E. 25
~
Q..
20 A
A
A
. •
==as A
A
J:
~

!. 15
.~
10
C
CD 10
C o
..III: • RT - 288 ~C (550 oF) Data
()
5 o
!! o RT - 232°C (450 oF) Data
0
o
o
0
0 5000 10000 15000 20000 25000
Number of Cycles

Fig. 4.6 Crack density compared with accelerated thermal cycles for graphite fabric/PMR-15 laminates. 31
90 High temperature resins

--
100

fI.
I 80
[J
.9 iii
-.6 iii
.,...
at

-t
c 60 [J iii

en IB
'.
ID
!Q.
E
40

B
t
...... * .. RT - 232 °C (450 oF) Data
...
0 20
0
[J RT - 288 °0 (550 oF) Data

0
0 5000 10000 15000 20000 25000
Number of Thermal Cycles

~~~.~1 Effect of accelerated thermal cycling on the compressive strength of PMR-15 / graphite fabric lam-

4.8.5 PMR-II !he 343°C (650°F) interlaminar shear strength

mcreased for the first 20Q-4(){) h of aging. Even
The greatly improved weight loss performance
after 600 hours aging the interlaminar shear
of PMR-II based composites compared with
strengths were about the same as they were at
PMR-15 as a result of aging in air at elevated
the beginning, However, the reader must be
temperatures (343°C, 650°F) is illustrated in
aware that there has been a considerable
Fig. 4.814•15 • Note that very little difference in
weight loss (about 8%) with the resulting
weight loss was found between PMR-II-30 and
shrinkage and induced stresses. Also, the
PMR-II-50. Although a stability improvement
matrix resin has presumably become embrit-
would have been expected with the higher for-
tied. Therefore, in assessing the useful lifetime
mulated molecular weight it usually is not
of a composite at elevated temperatures it is
detected because of the difficulty in making
im~ortant to not only consider the changes in
low void laminates with the higher viscosity
vanous strength and stiffness properties but
resin. Porosity itself, of course, can contribute
also the effects that weight loss itself can have
to thermal-oxidative instability.
on dimensional stability and matrix resin
As a general rule, for any of the polyimides
toughness.
discussed in this chapter, the high temperature
Carbon fiber selection can also be an impor-
mechanical properties tend to increase as a
tant factor in determining the overall high
result of the air aging process. This is presum-
temperature performance of a composite. As
ably because the oxidative crosslinking that
illustrated in Fig. 4.1()32 PMR-II-50 laminates
occurs during the aging process tends to
reinforced with Celion® G40-600 or Thomel®
increase the Tg • In Fig. 4.914,15 it can be seen that
650-35 retained a higher level of flex strength
 Mechanical properties before and after air aging 91

14

12 . PMR·15 •
0 PMR·/I·30
10
<> PMR·/I·50

~ 8
<>
0

l-
I 6
~
4 <>
0
<>

••
0
2
C
C D
0

0 100 200 300 400 500 600 700

Time. Hotn
Fig. 4.8 Weight loss of Celion® 6K/PMR laminates at 343°C (650°F).14,lS

60 8.70

«I
I:L 50 0 7.25 j

-
0
::I 0 0.

•
i
.&:
Q Q

-...
<> c

-:a
40 5.80
Qo
t!
! en
en 0

.
4.35
1CIJ
30
PMR-15 •
«I
.&:
en
...as
•...c
0 PMR·tI·30 2.90
20
PMR-II .. 50
c
0
'E
I as
;::
!
10 1.45
-•
;::

.E
0 0

0 100 '200 300 400 500 800 700

Time. Hours

Fig. 4.9 Interlaminar shear strength of Celion® 6K/PMR laminates after aging in air at 343°C (650°F).14,15
92 High temperature resins

200 180
&1 T40R

160 •E;l
G40-600
G40-700
160

.;;
as 0 GotO-800 .If!
D..
:i 6J T 650·42 120 .s:.
120
"6
i
-
rA c
CD
Lo
C
! CIJ
e 80 80 )(
CD
I u:
u:
40 40

o o
Inllial After 500 Hours
at 371 OC (700 OF)

Fig. 4.10 Effect of fiber selection on the flex strength of PMR-II-50 laminates air aged at 371 °C (700°F).32

o
T40R G40-600 G40-700G40-800 T850-42 T650-3S
Fiber

Fig. 4.11 Thermal-oxidative stability of various graphite fibers exposed to 371 °C (700 OF) air. 32
 Mechanical properties before and after air aging 93

after 500 h at 371 °C (700 0 P) than laminates rein- performance in comparison with PMR-II.
forced with other fibers such as Thomel 40R, These same researchers have also shown that
Thomel 650-42, Cellon G40-700 and Celion V-CAP / graphite composites possess improved
G40-800. It is of interest to note that according thermal-oxidative stability compared with
to the data in Pig. 4.1P2 there was not a one-to- PMR-II-50. However, with the advent of
one correlation with the basic carbon fiber AFR700B where significant improvements in
stability. While Cellon G40-600 had one of the thermal-oxidative stability have been realized
lowest weight losses in 371°C (700 0 P) air aging by simply capping one end only with the low
the Thome1650-35 had one of the highest. cost nadic anhydride, it is not clear whether
either the V-CAP or CYCAP polyimides with
their more expensive end-capping agents will
4.8.6 V-CAP AND CYCAP
be successful in the market place.
As a result of the reduced aliphatic character
of the reactive end-groups in CYCAP, Meador
et al. 19 have reported improved weight loss

Table 4.6 Mechanical properties of unidirectional AFR700B/S2 glass tape laminates 24

Material Fiber Vol. % Test temperature, °C (OF)

condition orientation fibers 23 (73) 316 (600) 371 (700)
Flex strength, MPa (ksi)
As-molded 0° 55 1123 (163) 634 (92) 496 (70)
Air aged 100 h 0° 55 958 (139) 517 (75)
at 371 °C (700°F) (1 atm)
90 0 Flex strength, MPa (ksi)
As-molded 0° 55 48 (6.9) 29 (4.2) 27 (3.9)
Air aged 100 h 0° 55 23 (3.3) 13 (1.9) 23 (1.7)
at 371 °C (700°F) (1 atm)
Flex strength, MPa (ksi)
Cycled 100 x from 0° 55 972 (141) 421 (61)
RT to 371 °C (700°F) in air
50 h total at 371 °C
In-plane shear strength, MPa (ksi)
As-molded ±45 59 81 (11.8) 53 (7.7) 41 (5.9)
Air aged 100 h ±45 59 29 (4.2) 21 (3.0) 26 (3.7)
at 371°C (700°F) (1 atm)
Tensile strength, MPa (ksi)
As-molded 0° 54 1103 (160) 917 (133) 834 (121)
Air aged 100 h 0° 55 814 (118) 703 (102) 676 (98)
at 371 °C (700°F) (1 atm)
0° Compression strength*, MPa (ksi)
As-molded 0° 58 662 (96) 386 (56) 303 (44)
Air aged 100 h 0° 58 621 (90) 352 (51) 331 (48)
at 371°C (700°F) (1 atm)
.. Non-standard test; open hole compression fixture used with no-hole specimen. For comparison the room temperature
llTRI compression strength (D3410) was 1152 MPa (167 ksi)
94 High temperature resins

........ AFR700BIS2
- 0 - AFR700B/T650-42
1.6
• "¢o' .. AFR700BJAstroquartz® III

j
1.2
.,."
.
~

..... ..... If •••

0.8
, . II
" ....
••
" • .01' "

,-OrrtI'.. til
..d'
0.4
'--
l" ·
o
o 1 000 2000 3000 4000 5000 8000 7000 8000
Time, Hours
Fig. 4.12 Effect of air aging at 260°C (500°F) on the weight loss of AFR700B composites. 24

4.8.7 AFR700B one can also expect significant improvements

in both matrix resin toughness and thermal-
One of the most promising new resin systems
oxidative stability compared with PMR-15.
that has come along in the last several years is
AFR700B. One indication of this is the excel-
lent properties tabulated in Table 4.624 which 4.8.8 TRW-R-8XX
had been obtained on unidirectional 52
Although very little unclassified property data
(933 finish) glass tape laminates, both as-
molded as well as after 100 h in air at 371 °e was available at the time of the writing of this
chapter, according to TRW25, the room temper-
(700 0 P).
In long term isothermal air aging studies24
ature mechanical properties of composites
based on TRW-R-8XX have been found to be
carried out at 2600 e (SOOOP) the weight loss
equivalent to those based on PMR-15 but with
performance of AFR700B laminates reinforced
with Astroquartz® III fibers was superior to superior toughness, higher T (371 °e to >
449°e, 700°F to >840°F), sup~rior retention
that exhibited by either 52-glass or Thomel
(>70% compared with <20% for PMR-15) of
650-42 laminates (Fig. 4.12). The retention of
properties at 371 °C (700 OF) and up to 10 times
the 260 0 e (500°F) flex strength of 52-glass lam-
better weight loss characteristics after 100 h at
inates is illustrated in Fig. 4.13. 5imilar results
371 °C (700 OF) compared with PMR-15. Based
were obtained with Astroquartz III and
Thomel 650-42. Overall, composites based on on these results and assuming reasonable pro-
APR700B offer the promise of reasonable cessibility, TRW-R-8XX could tum out to be a
processability, cost and excellent mechanical serious contender in the world of high tem-
properties in the dry-as-molded state out to perature polyimides, especially in view of the
about 371 °e (700 0 P). Based on its chemistry projected low cost.
 Electrical properties 95

90 90

80 80
•
~ :=
i 70 70
i
i ~
..
0..

US 60
US
60
Iu..
)II

u..
50 50

40 40
o 10002000 3000 4000 5000 8000 7000
Time, Hours

Fig. 4.13 Effect of air aging at 260°C (500°F) on the flex strength of AFR700B/S2-glass laminates.24
4.9 ELECTRICAL PROPERTIES reported to be 3.2 and 3.7 respectively. Also, in
the same types of laminates the dissipation
One of the useful properties of aromatic poly-
factors have been measured to be 0.001 for
imides is their good all around electrical
Avimid N and 0.015 for the Skybond. The par-
properties. Low dielectric constants and dissi-
tial fluorocarbon character of the 6F monomer
pation factors have been measured in quartz
is undoubtedly the reason behind the
fabric reinforced Avimid Nand Skybond
improved properties of Avimid N compared
based composites33• For instance, the room
with Skybond. Some other miscellaneous elec-
temperature dielectric constants for Avimid N
trical properties26 of E-glass reinforced
and Skybond Astroquartz laminates have been
Skybond based composites are tabulated in

Table 4.7 Electrical properties of Skybond® 700/181 style E-glass fabric laminates 26

Property Units Value

Dielectric strength
short time parallel to laminate volt 55000
step by step parallel to laminate volt 38000
short time volts/J.Im 179
stepwise volts/J.Im 140
Dielectric constant (1 MC) 4.1
Dissipation factor (1 MC) 0.0045
Insulation resistance megohm 1.9 X 107
Volume resistivity ohm-cm 2.47 X 1015
Surface resistivity ohm 3.35 X 1014
96 High temperature resins

Table 4.7. Skybond 700/E-glass fabric lami- there were also definite indications that voids
nates have been shown to have a dielectric in the composite could serve to facilitate the
constant and dissipation factor measured at drying out process, thus resulting in a higher
X-band (8.5 KMC) frequency at room temper- apparent 'wet' T . This means that although
ature of 3.74 and 0.016 respectively. When voids are well known to adversely affect
measured at 300°C (572°F) there was essen- mechanical properties their presence could
tially no change. conceivably result in significant improve-
ments in a composite's hot/wet properties to
the point where an overall better balance of
4.10 HYGROTHERMAL PROPERTIES
properties might be possible.
All of the mechanical properties discussed In the design of high temperature poly-
thus far in this chapter, either before or after imide parts it is strongly suggested that
air aging, have been determined on 'dry' spec- moisture effects be fully taken into account. It
imens. Although there are many reports in the seems apparent that the full potential of poly-
literature on the absorption of water by a wide imide composites will not be realized until
variety of polymers and the effects that this effective ways are found to reduce the adverse
water can ~ave on the Tg, strangely there have effects of moisture at elevated temperatures
been relatively few reports of the effects of without seriously affecting the other impor-
moisture on polyimide composites. Hot/wet tant properties such as strength, toughness
properties are normally reported for epoxy and thermal-oxidative stability.
and bismaleimide composites, but usually not
those based on polyimides. Unfortunately for
4.11 END-USE APPLICATIONS
polyimide composites water absorption is to
be expected since the equilibrium water con- In spite of their recognized limitations (e.g.
tent of the neat resins being normally in the microcracking and hygrothermal problems)
2-4% range. polyimide composites have been successfully
One of first references to moisture effects employed in a wide variety of applications.
occurred in 197634 in which studies at DuPont For instance, autoclave molded PMR-15
on Avimid N composites indicated that low graphite fabric composite has been employed
void «1 %) E-glass fabric laminates under- in the manufacture of ducts for the F-404
went an almost lOO°C (212°F) reduction in the engine used in the United States Navy's F-18
T when that property was measured on a fighter (Fig. 4.14). Other successful applications
Jater saturated laminate by thermal mechani- for PMR-15 include a fire wall for the GE-90
cal analysis at a heat-up rate of 50°C
(90°F)/min. The original 'dry' Tg of 342°C
(648°F) might have indicated good mechanical
property retention out to at least 316°C
(600°F). However, the 'wet' Tg of 245°C (473°F)
indicated a dramatically lower projected end-
use temperature «<245°C). It was also
reported in the same reference that at the
slower heat-up rate of SoC (9°F) per minute a
much higher apparent •
'wet' Tg was possible
(305°C, 581°F). Thus, If a part can be heated up
slowly enough, it can dry out as it is heated
and the deleterious affects of absorbed mois- Fig. 4.14 PMR-15 /graphite duct for the F-404 jet
ture can be greatly reduced. In this same paper engine.
 End-use applications 97

variable stator vanes in a variety of military

and commercial jet engines. Parts of this type
have also been made using PMR-15 and
PMR-II. Another interesting non-aerospace
application for Avimid N from DuPont has
been the product, NO-CHX® takeout jaws (Fig.
4.16), which have been used in the glass bottle
industry to grab the hot blow-molded glass
bottles to transport them during fabrication
Fig. 4.15 Avimid® N / graphite jet engine variable without cracking or checking.
stator vane bushings. Other applications for polyimide compos-
ites have been radomes, missile fins, jet engine
engine as well as various splitters and fairings nozzle flaps, fairings, cowls and inlet guide
for the F-110 engine. vanes, gear cases for helicopters and heat
Compression molded Avimid N/graphite shields.
variable stator vane bushings (Fig. 4.15) and
washers, now available from DuPont/Tribon
Composites, have been extensively employed
since the early 1980s in conjunction with the

Fig. 4.16 NO-CHX® Take out jaws based on Avimid N / graphite.

98 High temperature resins

REFERENCES 19. Meador, M.A., Cavano, P.J. and Malarik, D.C

Proc. Sixth Ann. ASM/ESD Adv. Compo Conf.,
1. Gibbs, H.H. 20th Nat!. SAMPE Symp. Exhib., Oct. (1990).
April 1975.
20. Sutter, J.K. et al., NASA CP-I0039, 12-1 (1989).
2. Sonnett, J.M., McCullough, RL., Beeler, A.J. and 21. Serafini, T.T. et al., US Patent 5091505 (Feb.
Gannett, T.P. 24th Intern SAMPE Tech. Conf., 1992).
October 1992, p. T735.
22. Serafini, T.T. et al., US Patent 5 149 760 (Sept.
3. Scola, D.A. United Technologies, private com- 1992).
munication.
23. Serafini, T.T. et ai., US Patent 5149772 (Sept.
4. Serafini, T.I., Delvigs, P. and Lightsey, G.R J.
1992).
App. Polym. Sci., 1972, 16, 905. 24. Johnson, K.M. Air Force Materials Laboratory,
5. Serafini, T.T., Delvigs, P. and Lightsey, G.R US
private communication.
Patent 3 745149 (July 1973).
25. Serafini, T.T. TRW-Redondo Beach, private com-
6. Serafini, T.T. and Delvigs, P. Appl. Polym. Symp.,
munication.
1973, 89, (22).
26. Monsanto trade literature.
7. Serafini, T.T. Proc. 1975 Intern. Conf. Composite
27. Scola, D.A. 34th Intern. SAMPE Symp., p. 246
Materials, AIME, New York, 1976, 1,202.
(1989).
8. Serafini, T.T., Delvigs, P. and Lightsey, G.R.
28. Hexcel trade literature.
NASA TN D-6877 (1972).
29. Gibbs, H.H. 10th National SAMPE Tech. Conf,
9. Burns, E.A., Lubowitz, H.R and Jones, J.P. 1978, p. 21.
NASA CR-72460 (1968).
30. Gibbs, H.H. J. Appl. Poly. Sci., Appl. Poly. Symp.,
10. Lubowitz, HR US Patent 3 528 950 (1970). 1979, 3S, 207.
11. Serafini, T.T., Vannucci, RD. and Alston, W.B.
31. Ward, D. General Electric - Aircraft Engines, pri-
NASA TMX-71984 (1976).
vate communication.
12. Vannucci, RD. 32nd Intern. SAMPE Symp.
32. Meador, M.A. NASA Lewis Research Center,
Exhib., April 1987.
private communication. .
13. Vannucci, RD. and Cifani, D. NASA TM-100923
33. Tanikella, M.S. Mosco, J.A. and Rafalski, T.J.,
(1988).
24th Intern. SAMPE Tech. Conf., 1992, p. 687.
14. Vannucci, RD. and Cifani, D. 20th Intern.
34. Gibbs, HH. 21st Nat!. SAMPE Symp. Exhib.,
SAMPE Tech. Conf., Sept. 1988.
1976, p. 607.
15. Serafini, T.T., Delvigs, P. and Vannucci, RD. 35. Fiberite trade literature.
36th Ann. Tech. Conf. SP! Reinforced 36. Engineered Materials Handbook - Composites,
Plastics/Composites Inst., Feb. 1981. 1987,1,79.
16. Vannucci, R.D., Malarik, D.C, Papadapoulos 37. Wilson, D. High Performance Polymers, 1991,3(2).
D.S. and Waters J.P. NASA TM 103233. 38. DuPont trade literature.
17. Vannucci, R.D., Malarik, D.C, Papadapoulos,
39. Iuliano-Picho, D. DuPont Composites, private
D.S. and Waters, J.P. 22nd Intern. SAMPE Tech. communication.
Conf., Nov. 1990.
18. Malarik, D.C and Vannucci, R.D. NASA CP-
10039,15-1 (1989).
SPECIALITY MATRIX RESINS 5
David A. Shimp

5.1 INTRODUCTION resins by resin suppliers accomplishing from

15-50% of the cyclotrimerization curing reac-
Bismaleimide (BMI) and cyanate ester (CE)
resins were first commercialized in the 1970s as ti?n using ~osely controlled reactor processing.
FIgure 5.1 illustrates formation of the s-triazine
250--300°C Tg class laminating resins for circuit
ring (cyanurate trimer) by the cycloaddition of
board substrates. In the early 1980s structural
three cyanate groups. Formulators catalyze
prepregs were introduced to an aircraft indus-
amorphous monomers or prepolymer resins
try searching for primary structure composites
with higher service temperatures and with latent catalysts which promote full con-
version to the thermoset polycyanurate with
improved damage tolerance relative to multi-
subsequent heating to 177-250°C. Cyanates
functional epoxy based composites. Both BMI
and CE resins have since evolved as easy-to- also serve to cure epoxy resins, forming cost-
e~ective .hybrids retaining an anomalously
process thermosetting resins qualified for
177°C (350°F) hot-wet service. Toughening high fraction of CE homopolymer properties
(Shimp, 1992).
technologies provide compression-after-impact
Bism~eimide monomers are prepared by
ratings approaching or matching the damage
the reaction of aromatic diamines with maleic
tolerance of thermoplastic resin composites.
anhydride in the presence of dehydrating
Bismaleimides, with higher modulus values
agents (Stenzenberger, 1990). Homopolymers
and established higher thermal ratings, earned
of BMI monomers are excessively brittle and
a strong position in military aircraft primary
~ pr~ctice a~ co-reacted with chain-extending
structures with recent selection for the F-22
fighter. Cyanate esters, with superior dielectric dlamm:s, dlallyl bisphenols or dipropenyl
loss properties and lower moisture absorption, phenoxldes to develop toughness via reduced
are strong contenders for radomes, skins cover- cross link density. Figure 5.2 illustrates chain
extension with aromatic diamine (Bargain,
ing phase-array antennae, advanced Stealth
1971) to form longer linear segments which
composites and space structures.
ultimately crosslink by homopolymerization
of maleic double bonds. Kerimid® resins and
5.2 RESIN CHEMISTRY Fiberite PI molding compounds are examples
of commercial BMIs using aromatic amine
Cyanate ester monomers are prepared by react-
modification.
ing bisphenols or polyphenols with cyanogen
Figure 5.3 depicts a series of reactions
chloride in the presence of an organic base
whereby o-allyl-phenols add across the maleic
(Rottloff, 1977). Crystalline monomers are ther-
double bond via the ene' reaction and a second
I
mally advanced to amorphous prepolymer
?-,aleimi~e ent~rs into a Diels-Alder ring-form-
mg reaction WIth the now conjugated propenyl
Handbook of Composites. Edited by S.T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 residual double bond. The proposed reaction
100 Speciality matrix resins

o 0

2 <N-R-N> +H,N-R-NH,

o 0
I
MICHAEL ADDITION
~

I
HOMOPOLYMERIZATION
~

Curing via
cyclotrimerization
~N-R-NR,
~
~ ~ ~N-R-N~
rr"N-R-N/~ ~
o 0 0 0

R=-o-X-o-
Y Y

x = alkylidene Y = alkyl or H

Fig. 5.2 Sequence of chemical reactions for advanc-

ing BMI monomers with aromatic diamines to resin
adducts by chain extension and ultimately to
toughened thermoset plastics.

mechanism includes isomerization to form the

aromatic ring. Crosslinking-can. occur by a con-
Thermoset plastic
(polycyanurate) tinuation of these reactions (difunctional
components) or via residual maleic double
Fig. 5.1 Dicyanates cure by forming triazine rings
bond homopolymerization. The chemistry by
on heating, advancing to prepolymers (up to 50% which o,o'-diallyl bisphenol Acoreacts with and
conversion) and to thermoset plastics at -O-C=N toughens 4,4'-bismaleimido-diphenylmethane
conversions >60--64%. (BMI-DAB) is described by Zahir (1978) and
 Matrix formulation 101

King (1984). Allyl functional phenoxy com- 5.3 COMMERCIAL RESINS

pounds follow the same reaction path while The chemical structure of seven commercial
propenyl functional phenoxides eliminate the
di(poly)cyanate ester monomers is shown in
'ene' reaction step (Stenzenberger, 1990).
Table 5.1 along with supplier information,
phYSical state and key homopolymer proper-
ties. The three crystalline monomers are
R usually supplied only as amorphous prepoly-
mers in semisolid, hard resin, or ketone

O~
+
laminating solution form. CE homopolymer

1
properties are not affected by prepolymer
OH advancement, which is only an interruption of
"ENE" REACTION the ring-forming curing reaction to alter phys-
ical state and rheological properties. Monomer
asymmetry, e.g. AroCy® L-lO, can yield low RT
viscosity. CEs have a low toxicity profile and
storage stability comparable to epoxies.
Table 5.2 describes several BMI monomers,
advanced resins, reactive tougheners with
allyl and propenyl functionality and RTM
resins. Most BMI monomers have a crystalline
DIELS-ALDER physical state. Eutectic blends of monomers
REACTION are available as resolidified melts of lower
melt point. Resins prepared by prereacting a
molar excess of BMI with tougheners have a
powder or hard resin physical state. Allyl and
propenyl functional reactive tougheners are
HOMOPOLY- usually viscous liquids which serve to dissolve
MERlZATION crystalline BMIs at temperatures below nODc
to offer convenient melt processibility.

AROMATIZATION
5.4 MATRIX FORMULATION

5.4.1 MEETING RHEOLOGICAL

HOMOPOLY- REQUIREMENTS
MERlZATION
Both resin classes offer a wide selection of
monomers and prepolymers enabling the
fomulator to satisfy the rheological properties
of fluid RTM compounds, tack and cohesive
integrity of compliant prepreg and the short
flow of compression molding compounds.
Fig. 5.3 Sequence of reactions between allyl phe- Figure 5.4 illustrates the limiting direct rela-
nols and BMI monomers involves grafting via the
tionship in families of thermosetting resins
'ene' reaction and fused ring formation via
Diels-Alder. Crosslinking with di(poly) functional between fluid monomer (ISO mPa s viscosity)
components involves completion of these reactions temperature and Tg developed on curing. High
and/or maleic double bond homopolymerization. temperature polyimides locate off-scale in the
102 Speciality matrix resins

Polycyanate monomer structure! Trade name! Homo'[!olymer '[!ropert!L

precursor supplier! Tg wt.% Dk G]C
physical state °C Hp MHz J m-2
AroCyB

~
rec-o-QCHo-0 0
O-C:N
CHs
Ciba Specialty Chern.*

BT-2000 289 2.5 2.91 140

Mitsubishi GC

BisphenolA Crystal

~H3 , 3
CH
AroCyM
N:C-o{} <Q)-o-c:N
,
~ \
Ciba Specialty Chern. 252 1.4 2.75 175

Crystal
CH3 CH3
Tetramethylbisphenol F

N=C-~<::»~~)~~~~C:N
AroCyF
Ciba Specialty Chem. 270 1.8 2.66 140
CF 3
Crystal
Hexafluorobisphenol A

0 0
N:C-o { } C~H o -o-CEN
AroCyL-10
Ciba Specialty Chem. 258 2.4 2.98 190

Liquid
Bisphenol E

tEe
-o-O,CH3
0 ~C~C~
0 :N
XU-366
Ciba Specialty Chern. 192 0.7 2.64 210
CH CH a
Semisolid
BisphenolM

~~¥i
PrimasetPT 270

O~
Lonza, Inc.
CH to 3.8 3.08 60
XU-371
n
Ciba Specialty Chern. >350
Novolac resin Semisolid

NiC- XU-71787
Dow Chemical 244 1.4 2.80 125

Dicyclopentadienyl bisphenol Semisolid

* The complete name of the Ciba company supplying AroCy cyanate resins is Ciba Specialty Chemicals Corp.,
Performance Polymers Group.
 Matrix formulation 103

Table 5.2 Commercial BMI monomers, resins (adducts), reactive tougheners and compounds

Supplier Trade name/structure Description

~¢
Basic BMI monomer
Crystalline powder
Cilia
m.p. 150-160°C
o 0

Matrimid® 5292A

o,o'-Diallyl Bisphenol A
Ciba Reactivetoughener
12 000-20 000 mPa s at 25°C

Matrimid 5292B

Inspec Compimide® MOAB Basic BMI monomer

Inspec Compimide 353 Resolidified BMI melt

Eutectic monomer blend 400-1400 mPa s at 110°C

Inspec Compimide 796 Resolidified BMI resin melt

Proprietary BMI adduct 1000-4500 mPa s at 110°C

Bisallyl polyphenoxide
Inspec Reactive toughener
Compimide TM 121 120-250 mPa s at 71°C

~~~~
Bispropenyl phenoxy
benzophenone
Reactive toughener
Inspec 1000-1600 mPa s at 71 °C
b
o
Compimide TM 123

Inspec Compimide 15 MRK Powder for molding compounds

Inspec Compimide 65 FRW Resolidified melt for filament

winding and RTM
Ciba designates Ciba Specialty Chemicals Corp., Performance Polymers. Inspec designates Inspec Fine Chemicals Co.

Table 5.1 (on facing page) Commercial cyanate ester monomers, suppliers, physical states and homopoly-
mer properties. Water absorption is wt.% at saturation. Dk = dielectric constant
104 Speciality matrix resins

400~--------------------------~
106r-~----rT----------~
Tack

B
&
Drape
•CI.
o
o
300

o
AROCY
S SYM. DAi
ANAT S '--J
8M.'-
U

• 105,+--='.-"",*-\.
•
....ci TETRA- CI.
L-10 RTX L...-.....I E

o
388 EPOXIDE
~200
UJ
U
o

..•
w r----1
10
II: DIEPOXIDE
C
w 104+------'-~~-r;-------~
W
II:
>-
5 100 l-
i
o
o
; 1 03J..WlndItna~......
o~----~~----~----~----~----~~
o ~ 100 1~
Resin
MONOMER TEMP. (OC> at 150 MPA'$ Transfer
Molding
Fig. 5.4 Relationship between fluid monomer tem-
102+-~~~~~-r~~--~
perature and cured T in families of thermosetting o 20 40 60 80 100
resins. Higher service lemperature is normally asso- AROCY L-10, WEIGHT or.
ciated with increasing processing difficulties.
Fig. 5.5 The asymmetric structure of AroCy L-IO dis-
upper right quadrant. AroCy L-10, derived rupts crystallinity, permitting optional use of this
from an asymmetric bisphenol, breaks the pat- ring-forming resin as a reactive diluent. AroCy num-
tern and can be used as a 250°C T resin or as bers are the % cyanate conversion of prepolymers.
a reactive diluent of 120 mPa'; viscosity
(Fig. 5.5). but predissolving in 2-6 phr (parts per hun-
dred resin) alkyl phenol, e.g. nonyl or dinonyl
phenol, forms stable liquid packages which
5.4.2 CURE CATALYSTS
are readily miscible (Shimp, 1988). The alkyl
Catalysts are not required to cure BMI resins at phenol provides the active hydrogen co-cata-
temperatures above 200°C, but several types lyst and can serve as a monofunctional
provide effective cure acceleration. Tertiary reactant to increase conversion and resistance
amines, imidazoles and free radical generators to boiling water at marginal cure temperatures
are noted by Zahir (1978). Boyd (1987) (Fig. 5.6). Extension of this principle to AroCy
describes the preferred latency of tri- XU-366 enables this monomer to convert satis-
phenylphosphine and its phosphonium halide factorily at 121°C for use with high modulus
derivatives as prepreg catalysts. polyethylene fibers (Shimp, 1994a) and with
Cyanate esters require catalysis to cure at composite tools.
practical rates. Copper (most active at low
temperatures) and cobalt (latent) acetylaceto-
5.4.3 TOUGHENING TECHNIQUES
nates provide ~5% conversion within 2-6 h at
post cure temperatures in the range of Concentrated effort over the last decade has
200-250°C. Metal coordination catalysts in produced composite toughening techniques
general are difficult to solubilize in neat resins, which satisfy damage tolerance requirements
 Matrix and composite properties 105

the CAl performance level of -::::.45 ksi was

demonstrated (Boyd, 1993a) by combining in
situ epoxy extension of o,o'-diallyl bisphenol A
~ 1600 with thermoplastic polyimide particles.
"
z
ILl
IE:
CE monomers and prepolymers dissolve
powdered amorphous thermoplastics (Tps) of
t-
en the polysulfone, polyethersulfone, polyether-
~1200 imide, polyphenylene oxide and copolyester
II)
families, then subsequently phase separate
ot- these thermoplastics during cure. Co-continuous
Z morphologies are developed at Tp concentra-
oen 800 tions -::::.15% which increase G,c values in a non-
IE: linear response to concentration (Shimp,
ILl
:& 1994a). Lee (1991) describes the development
! of a CE matrix formulation toughened with
: 400 RESISTANCE OF AROCY B polyoxazolidinones, polyethersulfone and co-
:;:) polyester Tp resins. Table 5.3 lists a number of
o: HOMOPOLYMER TO Tp resins used to toughen both CE and BMI
HYDROLYSIS AT 100·C resins as well as reactive rubbers used to elim-
inate microcracking in orbital service.
O~~---r--~~r-~--~
o 2 4 6
NONYLPHENOL CONC. (phr)
, i i i 5.4.4 EPOXY RESIN MODIFIERS
10 20 30 40
Meq OH/OCN Epoxy resins derived from epichlorohydrin
co-react with CE resins at equivalent ratios of
Fig. 5.6 Increasing concentrations of alkyl phenols up to 1.2 epoxides per monomer cyanate.
in cyanate ester homopolymers increase conversion Hybrids with typical epoxy weight fractions of
(numbers at right) for a given cure temperature and 50-70% develop Tg values in the ISO-200°C
increase resistance of 3 mm thick castings to hydrol- range and retain dielectric constants ~.1 with
ysis in boiling water. Cure temperature: .: 250°C;
....,: 210°C; .: 177°C. (for AroCy B) loss tangents generally below 0.010 (Shimp,
1992). The use of diepoxides to toughen BMI
resins via chain extension of alkenyl (bis)phe-
of primary aircraft structures and microcrack nols was described in the previous section.
resistance in earth orbit and cryogenic service. Epoxides can also react with the secondary
BMI resins earlier required the development of amine formed by the Michael addition of aro-
allyl, propenyl and amine functional reactants matic diamine to the BM! maleic double bond.
to achieve >2% tensile elongation-at-break, a
minimum requirement for efficient secondary
5.5 MATRIX AND COMPOSITE PROPERTIES
toughening with thermoplastic polyimides.
Boyd (1990) describes the development of The following acronyms will be used to
improved chain extending reactants by cou- denote BMI resin system properties in figures
pling 2- or 4-propenyl phenol with diepoxides. used thoughout this section:
Incorporation of thermoplastic polyimide fine
BMI-MDA = the reaction product of 4,4'-bis-
particles of 5-15 /lm diameter, described by
maleimidodiphenyl methane (molar excess)
Boyd (1991a), increased the compression-after-
with methylene dianiline.
impact (CAl) performance of BM! composites
to -::::.276 MPa (-::::.40 ksi). Further toughening to BMI-DAB = the equimolar reaction product of
106 Speciality matrix resins

Table 5.3 Thermoplastic and reactive rubber tougheners

Classification Product Tg' °C Used Supplier

with

Soluble T/
Polyethersulfone Victrex 5003P 203 CE ICI/Mitsui
Polysulfone Udel P-1700 175 CE Amoco
Polyetherimide Ultem 1000(P) 215 CE General Electric
Polyphenyleneoxide PPO 202 CE General Electric
Polyimide Matrimid 5218 300 CE Ciba
Elastomeric Tp
Copolyester Vitel PE-307 14 CE Bostik
Reactive rubbers
Soluble" (OH) ATX-013 <25 CE Echo Resius
Soluble" (Epoxy) HycarETBN <25 CE B.E Goodrich
Preformed Core/Shellb CRS (exp.) <25 CE Dow Chemical
Polysiloxanes
Epoxy functional Experimental <25 CE Proprietary
Maleimide functional PAP Series <25 CE National Starch
Particulate T b
Polyimide P P-84 290 BMI LenzingAG
Polyimide Matrimid 5218 300 BMI Ciba
Polyamide 1002 DNAT 85 CE Atochem Corp.
a Initially soluble but phase separate during cure.
b Small particles swell but do not completely dissolve with cure.
Note: Most of this toughening technology is described in patents.

4,4'-bismaleimido diphenylmethane with Although CE resins demonstrate higher tem-

o,o'-diallylbisphenol A (Matrimid 5292). perature onsets of rapid thermal degradation
in TGA tests, available long term isothermal
aging tests in air indicate superior BMI perfor-
5.5.1 MECHANICAL PROPERTIES
mance. Boyd (1993b) classifies I-year CAl
Table 5.4 compares properties of representa- retention life of BMI composites as >177 but
tive CE and BMI castings. Significant <205°C. Stenzenberger (1991) rates 2000 h life
differences are the higher room temperature of BMI castings as >200°C but <250°C based
modulus values of BMI matrices (superior on retention of shear and flexure strengths.
stress transfer to fiber) and higher CE elonga- CE/E-glass laminates are rated at 162-180 DC
tion-at-break values. for 25 000 h retention of flexure strength at
50% of the unaged values (Shimp, 1989).

5.5.2 THERMAL PROPERTIES 5.5.3 DIELECTRlC PROPERTIES

T values average about 20 DC higher for BM! Figure 5.7 ranks the dielectric constant (Dk )
g .
matrIces (Table 5.5). CTE values below Tare
g
and dissipation factor (Df ) or loss tangent of
comparable while CE resins retain higher char thermoset matrix resin castings compared
yields as a result of increased aromaticity. with reference thermoplastics. Effects of
 Matrix and composite properties 107

RESIN DIELECTRIC PROPERTIES 7.-----------------------,

25°C OK
01< Of 6
TGMDAlOOS 10- 1
4
FR-4 EPOXY
5

BMI-OAB ~FR-4 EPOXY

10- 2 4
BMI-OAB
~roCy L AroCy B, L
3
AroCy B AroCy M ,F 3
AroCy M
AroCy F 10- 3
XU-366 2
POLYETHYLENE "'XU-366
2 PTFE
10- 4 POLYETHYLENE 20 40 60 80 100
PTFE RESIN CONTENT, Volume %
Fig. 5.8 Effect of reinforcement and concentration
on dielectric constant of AroCy M composites. Test
1 AIR data at 25°C and 1 MHz.

terns. Moisture locates primarily in the CE free

Fig. 5.7 Thermosetting and thermoplastic resins are
ranked for dielectric constant (Dk ) and dissipation volume fraction, resulting in less swelling
factor (Df) at 1 MHz frequency. (Fig. 5. 10) than is caused by association with
strong dipoles. Plasticization of moisture-con-
moisture absorption, test temperature and fre- ditioned matrix castings, compared as a
quency are summarized by Shimp (1994b). function of flexural modulus retention at ele-
Low Dk values of CE homopolymers and vated test temperatures, is minimized by the
CE/epoxy hybrids are attributed to the sym- low absorption of AroCy M o-methylated CE
metrical arrangement of electronegative resin (Fig. 5.11). Hydrolysis of AroCy M cya-
oxygen and nitrogen atoms around a central nurate linkages in 121°C steam requires >600 h
electropositive carbon atom in these struc- exposure (Fig. 5.12).
tures, resulting in weak dipoles. Dielectric
constants of CE composites compared with
5.5.5 PROPERTIES OF UNIDIRECTIONAL
fiber type and loading are plotted in Fig. 5.B.
COMPOSITES
Dk and loss tangent values of quartz reinforced
BMI, CE and epoxy composites are compared Properties of intermediate modulus carbon
at four radar bandwidths in Fig. 5.9 (Speak, fiber reinforced BMI composites (Table 5.6)
1991). and CE composites (Table 5.7) indicate good
translation of fiber strength for both classes.
Damage tolerance ratings based on CAl
5.5.4 MOISTURE ABSORPTION AND EFFECTS
results at 6.7 KJ m-I impact energy fall in the
CE homopolymers absorb less moisture than 200-345 MPa (30-50 ksi) class, approaching or
BMI and TGMDA/DDS epoxy matrix sys- equaling the damage resistance of thermoplastic
108 Speciality matrix resins

Table 5.4 Mechanical properties of CE and BMI resins

AroCy AroCy AroCy Matrimid

B M L 5292

Composition (PBW)
AroCyB-30 100
AroCyM-20 100
AroCyL-lO 100
Matrimid 5292A 100
Matrimid 5292B 85
Nonylphenol 2 2 2
Cobalt acetylacetonate 0.13 0.13 0.13
Property of casting"
Tensile strength, MPa 88 76 87 82
ksi 12.7 11.0 12.6 11.9
Tensile elongation, % 3.2 2.7 3.8 2.3
Flexure strength, MPa 174 159 187 167
ksi 25.2 23.0 27.1 24.2
Young's modulus Flexure Flexure Flexure Tensile
25°C, GPa 3.17 2.97 3.24 4.28
msi 0.46 0.43 0.47 0.62
149°C, GPa 2.42
msi 0.35
163°C, GPa 2.55 2.35 2.28
msi 0.37 0.34 0.33
204°C, GPa 2.00
msi 0.29
G 1C' Jm-2 140 175 190 170
inlb in-2 0.80 1.00 1.08 0.97
aStep-cure with post cure of 2 h at 250°C for CE; 6 h at 250°C for BM!. Data courtesy of Ciba Specialty Chemicals Corp.,
Performance Polymers Group.

composites. BMI composites have demon- Korex™ aramid/phenolic honeycomb core

strated hot-wet performance in 177°C rated (DuPont) are recommended for use with CE
aircraft. CE composites may attain that goal resins catalyzed with copper or cobalt acety-
with AroCy M resin, but insufficient 177°C lacetonates to eliminate blistering associated
hot-wet compression data has been published with post cures >190°C. (Shimp, 1993; 1994a).
for commercial materials of this class.
5.6.2 GALVANIC CORROSION
5.6 DESIGN CONSIDERATIONS
Both BMI and CE carbon fiber composites
have been reported to undergo resin degrada-
5.6.1 SELECTION OF ARAMID FillER AND
tion in accelerated galvanic cell tests
CORE
producing strongly alkaline conditions at
Aramid fiber and core reinforcements for CE cathodic sites (Boyd, 1991b; Olesen, 1991).
composites should be selected from second Figure 5.13 compares the onsets and rates of
generation materials which absorb <2% mois- alkaline hydrolysis (etching) for CE and BMI
ture in the workplace. Kevlar® aramid fiber and matrix castings. (See also Boyd, 1991b).
 Suppliers of prep reg and other formulated products 109

Table 5.5 Thermal properties of CE and BMI resins

AroCy AroCy AroCy Matrimid

B M L 5292

Composition (PBW)
AroCyB-30 100
AroCyM-20 100
AroCyL-10 100
Matrimid 5292A 100
Matrimid 5292B 85
Nonylphenol 2 2 2
Cobalt acetylacetonate 0.13 0.13 0.13
Property of casting'
HDT, DC
Dry 254 252 249 273
Wet 197 226 183 217
T, DC
g byDMA 289 267 270 295
byTMA 257 255 259 273
CTE by TMA, ppm;oC
40 to 200 DC 64 66 64 63
TGA at 10°C / min
Onset in air, °C 411 406 408 371
Char in N 2, % 41 46 43 29
Specific gravity at 25 DC 1.201 1.151 1.228 1.232
aStep-cure with post cure of 2 h at 250°C for CE; 6 h at 250°C for BM!. Data courtesy of Ciba Specialty Chemicals Corp.,
Performance Polymers Group.

Effective design practices for susceptible com- strength and generates destructive heat, limit-
posites are use of titanium rather than ing power and range. CE composites curing at
aluminum rivets, placement of a fiberglass 121°C (250°F), e.g. Bryte Technologies'
reinforced insulating ply and/or modification EX-1S1S, are thermally compatible with high
of CE resin with 55-70% epoxy resin. modulus polyethylene reinforcement. Such
composites are characterized by Dk values as
low as 2.6 and Df values as low as 0.004 when
5.6.3 MICROWAVE TRANSPARENT
measured at 10 GHz.
COMPOSITES
Composite design for radomes, antennas and
5.7 SUPPLIERS OF PREPREG AND OTHER
advanced stealth structures should utilize low
FORMULATED PRODUCTS
dielectric loss materials (Speak, 1991; Shimp,
1994b; Stonier, 1991a,b). Figure 5.14 summa- Table 5.8 lists suppliers of BMI and/or CE
rizes microwave interactions with a radome prepreg, adhesive, syntactic foam, RTM/ fila-
wall. Reflection weakens returning signals and ment winding systems and chopped fiber
overheats emitter sources; refraction distorts reinforced molding compounds formable by
signal quality; absorption decreases signal compression, injection or transfer processes.
110 Speciality matrix resins

4.0
DIELECTRIC

~
CONSTANT

3.5 ~ EPOXY
BMI

CE

3.0
X BAND Ka BAND U BAND W BAND
8-12 26-40 40-60 75-100
GHz GHz GHz GHz
0.030
LOSS
TANGENT
0.020 to-
EPOXY
8MI
0.010 --.;;;:::

CE
0.000
X BAND Ka BAND U BAND WBAND

Fig. 5.9 Comparison of typical quartz reinforced radome composites for dielectric loss properties measured
at four radar bandwidths. Redrawn from Speak, S.c., Sitt, H and Fuse, R.H .. 1991. Novel cyanate ester
based products for high performance radome applications. Int. SAMPE Symp., 36 pp. 336-347.

, 5.4
+3 6.2
A "
of.
or
0»
c +2
/~ 4.8

"
.c
0
II)
E
71"
,
,
VOLUME '"
WATER

:s
+1 , I--TGMDA
DDS ABSORBED

~ , pO
... ' 0 a6 AROCYB

"
a:I o
0-0-060-0
RTX366
2.6
0.7
0
10 1 10 2 103 10 4
Hours at 25°C & >95% RH
Fig. 5.10 Changes in 3 mm thick bar volumes during water immersion for a period of one year indicate
swelling rates and limits of thermoset resins. The ratio of volume increase to total volume of water
absorbed (numbers on right) indicates the fraction of water associated with dipoles.
/::,.: BMI-MDA; X: BMI-DAB; 0: TGMDA-DDS; 0: AroCy B; 0 :XU-366
 Suppliers of prep reg and other formulated products 111

Table 5.6 Properties of BMI/IM-7 unidirectional composites

Meclumical strength Rigidite Rigidite

5250-4" 526()b

0° Tensile, MPa (ksi)

25°C 2618 (380) 2691 (390)
0° Compression, MPa (ksi)
25°C Dry 1820 (235) 1746 (253)
105°C Wet 1346 (195)
149°C Wet 1276 (185)
177°C Dry 1310 (190)
177°C Wet 966 (140)
0° Compressive modulus
25°C, GPa (msi) 158 (23) 152 (22)
Open hole compression, MPa (ksi)
25°C Dry 420 (61) 352 (51)
177°C Dry 351 (51) 269 (39)
191°C Wet 303 (44) 221 (32)
Compression after impact
At 4.5 kJ m-I, MPa (ksi) 248 (36) 380 (55)
At 6.7 kJ m-I, MPa (ksi) 214 (31) 345 (50)
Edge delamination, MPa (ksi)
25°C 241 (35) 358 (52)
• Data courtesy of Cytec. Post cure 6 h at 227°C; 60% fiber vol.
b Data courtesy of Cytec. Post cure 6 h at 215°C; 60% fiber vol.

Table 5.7 Properties of CE unidirectional composites

Cytee Fiberite Hexeel

5245C 954-2 HX-1562

Reinforcement/cure
Carbon fiber IM-6 IM-7 IM-7
Max. cure temp., °C 210 232 177
Mechanical strength
0° Tensile, MPa (ksi) 2439 (356) 2814 (408) 2610 (378)
0° Compression, MPa (ksi)
25°C, Dry 1690 (245) 1573 (228) 1700 (246)
121°C, Wet 1350 (196) 1331 (193)
132°C, Wet 1310 (190) 1140 (165)
149°C, Wet 987 (143) 1290 (187)
CAl, MPa (ksi)
At 6.7 kJ m-1 214 (31) 262 (38) 317 (46)
Edge delamination, MPa (ksi) 262 (38) 269 (39)
112 Speciality matrix resins

Fig. 5.11 (left) Moisture plasticization of cast

1oo,-------------------~100
matrix systems is inversely related to the percent-
age of dry room temperature flexural modulus
retained at elevated test temperatures.
80
..........
U) • : at 149°C wet;
:::)

:::)w
• : at 177°C wet.
a~
0 .... 60
::Ea
... w
C Z
a::-
:::)c 40
>et.
...
wa::
II. 20 Fig. 5.12 (below) Hydrolysis of unsubstituted CE
~ (bisphenol A dicyanate) homopolymer begins to
reduce mechanical properties after 200 h exposure
to 121°C steam autoclave at 15 psig. Ortho-methy-
AroCy AroCy BMII lation is an effective technique for increasing
M-20 B-30 DAB hydrolytic stability of cured CE resins in aggressive
environments. 0: AroCy B; .: AroCy M.

% WEIGHT GAIN
12~--------------------------------------'

9
AROCV B
6 N=C-O-o~~o-o-C:N AROCV M

3 /~SET
';'Pi~~ ~ ~~~ CHa CHa

______~________~T=M=B~F=D~C~:y__~
O~~~
o 200 400 600
TIME. HOURS

Table 5.8 Sources of formulated/ compounded CE and BMI products

Supplier Prepreg Adhesive Syntactic RTM Compression molding

foam compound compound

Bryte Technologies CE CE CE CE
Cytec BMI,CE BMI,CE BMI,CE BMI
Hexcel BMI,CE BMI
Fiberite, Inc. CE BMI,CE
YLA CE CE CE CE
 Applications 113

+1.0r------------...,
AROCY B I EPOXY

....----
l: t ABSORPTION
~--
{ / ~~ :.
Ji:. ",
I
I

10 20 30 40 50 60
\
. ~

DAYS IMMERSION IN 20% NaOH AT sooe TRANSMISSION

Fig. 5.14 Interactions of microwaves with a radome
Fig.5.13 Cured CE and BM! resins hydrolyze (etch) wall.
in strongly alkaline solutions, as indicated by the
onset of weight loss. Ortho-methylated CE resin
and blends with epoxy resin (50/50 blend shown) were used to construct EFA (Eurofighter) pro-
increase resistance to alkaline environments gener- totypes and are used in construction of the
ated in galvanic cells. Dassault Rafale. Both materials are candidates
for HSCT (High speed civil transport) use.
Principal applications for CE composites
5.8 APPLICATIONS
(McConnell, 1992) include radomes for mili-
Toughened BMI/ carbon fiber composites have tary aircraft, fighter aircraft retrofitted with
been specified as the principal composite mate- improved tracking systems, skins over phase
rial for F-22 fighter primary and secondary array radar, weather tracking aircraft radar
structures (Fig. 5.15). BMI service temperatures and missile nose cones. CE prepreg reinforced
are sufficiently high for cowlings, nacelles and with high modulus pitch-based carbon fibers
thrust reversers of jet engines. CE composites are preferred materials for earth orbit service,

Fig. 5.15 F-22 fighter constructed with BMI composites. Photograph courtesy of Lockheed.
114 Speciality matrix resins

demonstrating low outgassing, microcrack Rottloff, G. et al. 1977. US Patent 4 028 393.
resistance and resistance to 109 rads of ionizing Shimp, D.A. 1988. US Patent 4785 075.
radiation (Willis, 1991). Applications in space Shimp, D.A., S.J. Ising and J.R. Christenson. 1989.
Cyanate esters: a new family of high tempera-
include communication satellites, solar arrays,
ture thermosetting resins. SPE/Case Western
parabolic antennas, optical benches and preci- Conf on High Temperature Polymers and Their
sion segmented reflectors. Uses, 1, 127-140.
BM! film adhesives are employed in jet Shimp, D.A. and I.E. Wentworth. 1992. Cyanate
engine or high speed aircraft sandwich panels ester-cured epoxy resin structural composites.
where hot-wet service up to 190°C is required. Int. SAMPE Symp., 37, 293--305.
CE film and paste adhesives are used together Shimp, D.A and M. Southcott. 1993. Controlling
moisture effects during the curing of high T
with syntactic foams in the construction of
cyanate ester/aramid composites. Int. SAMPt,
radomes. BM! molding compounds reinforced Symp., 38, 370-379.
with up to 65 wt. % of chopped reinforcements Shimp, D.A. 1994a. Technologically driven applica-
are used to mold ducts, drive sprockets for tions. In Chemistry and Technology of Cyanate
heated rolls in copy machines, helicopter gear Ester Resins (I. Hamerton Ed.) Chap 10. Blackie,
boxes and missile strongback mounting sup- Glasgow, pp. 282-327.
ports. Shimp, D. and B. Chin. 1994b. Electrical properties
and their significance for applications. In
Chemistry and Technology of Cyanate Ester Resins
REFERENCES (I. Hamerton Ed.) Chap 8. Blackie, Glasgow, pp.
230-257.
Bargain, M. et al. 1971. US Patent 3 562 223. Speak, S.c., H. Sitt and R.H. Fuse. 1991. Novel
Boyd, J.D. and D.A. Shimp. 1987. US Patent cyanate ester based products for high perfor-
4644039. mance radome applications. Int. SAMPE Symp.,
Boyd, J.D. and Hon-Son R. 1990. US Patent 36,33~347.
4923928. Stenzenberger, H.D. 1990. Chemistry and properties
Boyd, J.D. 1991a. US Patent 5 037 689. of addition polyimides. In Polyimides (D.
Boyd, J. et al. 1991b. Galvanic corrosion effects on Wilson, P.M. Hergenrother and H.D.
carbon fiber composites. Int. SAMPE Symp., Stenzenberger, Eds) Chap 4. Blackie, Glasgow.
36,1217-1231. Stenzenberger, H.D. et al. 1991. BMI/bis(allylphe-
Boyd, J.D. and L.N. Repecka. 1993a. US Patent noxy phthalimide)-copolymers: improved
5189116. thermal oxidative stability. Int. SAMPE Symp.,
Boyd, J.D. and G.E.C. Chang. 1993b. Bismaleimide 36 pp. 1232-1243.
composites for advanced high temperature Stonier, R.A. 1991a. Stealth aircraft and technology
applications. Int. SAMPE Symp., 38, 357-369. from World War II to the Gulf, Part I. SAMPE
King, J.}., Chaudhari M. and Zahir. S. 1984. Nat. Journal, 27(4), 9-16.
SAMPE Conf, 29 392. Stonier, R.A. 1991b. Stealth aircraft and technology
Lee, EW. and K.S. Baron. 1991. US Patent 5 045 609. from World War IT to the Gulf, Part II. SAMPE
McConnell, v.P. 1992. Tough promises from cyanate Journal, 27(5), 9-18.
esters. Adv. Comp., May/June pp. 28-37. Willis, P.B. and D.R. Coulter. 1991. Applications of
Olesen, K. 1991. Degradation of graphite/polymer cyanate resins to spacecraft composites. Paper
composites in the presence of a corroding metal. read at 8th Int. Conf Composite Materials,
Read at the High Temple Workshop, 11, Reno, ECCM/VIII, Honolulu, 15-19 July 1991.
Nevada, 4 Feb 1991. Zahir, Sheik A-C. and A. Renner. 1978. US Patent
4100140.
THERMOPLASTIC RESINS 6
Lars A. Berglund

6.1 INTRODUCTION po~it~s o~er advantages. They have very low

tOXICIty smce they do not contain reactive
Thermoplastic composites form a fairly new
chemicals (therefore storage life is infinite).
group of materials. Commercial prepreg tape
Because it is possible to remelt and dissolve
such as CF/PEEK (carbon fiber/polyether
such thermoplastics, their composites are also
etherketohe) and later CF /PPS (polyphenyle-
easily recycled or combined with other recy-
nesulfide) was introduced in the early 1980s.
cled materials in the market for molding
However, as early as 1966, Menges reported on
compounds.
improved static strength and fatigue resistance
In the aerospace market, composites based
when epoxy was replaced by polyamide 6 as
~n toughened epoxies dominate. The poten-
a composite matrix (Menges, 1966). In the mid
tially cheaper manufacturing of thermoplastic
1970s there was interest in CF /PSU (polysul-
composites has not yet been realized to the
fone) due to expectations of better processing
extent necessary to motivate large-scale invest-
methods and improved toughness characteris-
ment in new manufacturing equipment.
tics. However, solvent resistance was found to
However, for the next generation of aircraft
be a problem. Composites later introduced
based on semi-crystalline thermoplastics, such in~erest in. thermoplastic composites is high:
as PEEK and PPS, which have been introduced HIgher flymg speeds require higher tempera-
tures in the materials than the maximum
more recently, have excellent chemical resis-
temperature available from epoxy-based com-
tance and are superior to epoxy-based
pos~tes in use t.oday. Since the release of gases
composites in this respect.
Enthusiasm for thermoplastic composites is durmg processmg and inherent brittleness are
serious disadvantages of thermoset polyimides,
generated for, basically, three different rea-
thermoplastic composites are of great interest.
sons. First, processing can be faster than for
In tt:e automotive market, thermoplastic
~ermo~et composites since no curing reaction
compOSItes are used extensively. Matched-die
IS re~=JUlred. 1?ermoplastic composites only
requue heatmg, shaping and cooling. compression molding of glass mat thermo-
~lastics (GMT), primarily based on glass
~econdly, ~e properties are attractive, in par-
hcular, hIgh delamination resistance and fIber/polypropylene (GF/PP), is common,
because it permits fast processing cycles for
damage tolerance, low moisture absorption
fairly large components. In the established
and the excellent chemical resistance of semi-
field of injection molded components, materi-
crystalline polymers. Thirdly, in light of
als are used with long fibers (5-10 mm) in
environmental concerns, thermoplastic com-
molding pellets. This leads to improved
mechanical properties compared with materi-
als based on shorter fibers (Truckenmueller
Handbook of Composites. Edited by S.T. Peters. Published
in 1998 by Chapman & Hail, London. ISBN 0 412 54020 7
and Fritz, 1991).
116 Thermoplastic resins

6.2 MANUFACTURING METHODS processing, for conversion into a high-viscos-

ity melt. Shaping then takes place and the
The principles for thermoplastic composites
processing are very different from those for material solidifies on cooling. In Table 6.1,
different manufacturing methods used for
thermoset composites. During the processing
of thermosets, the polymer is initially a liquid thermoplastic composites are outlined.
Further discussion of thermoplastic compos-
which then solidifies due to the formation of
ites processing is available in Chapter 24 of
a three-dimensional molecular network from
this book and in previous reviews (Carlsson
chemical reactions. Thermoplastics are in the
1991; Cogswell, 1992; Kausch, 1993).
solid state before processing because of their
As with thermoset composites, materials
high molecular weight. They are heated
with low fiber volume fractions show ease of
above their softening temperature during

Table 6.1 Manufacturing routes for composites based on thermoplastic resins

Manufacturing route Outline offabrication and processing methods

Open mold processes
1. Autoclave Unidirectional or woven fibers pre-impregnated by the resin (prepreg) are
used. Other forms of prepreg have reinforcing fibers in combination with
the resin as fibers or as powder. The prepreg layers are stacked on the
mold surface and covered with a flexible bag. Consolidation is obtained by
external pressure applied in an autoclave at elevated temperature.
2. Filament winding Prepreg tape or tape with the resin as fibers or powder are wound onto a
mandrel at pre-determined angles. Heat and pressure are applied to the
tape in order to continuously weld it onto the underlying material.
3. Folding Preconsolidated sheets are heated. Simple fixtures are then used to shape
the sheets into the desired geometry.
Closed mold processes
4. Injection molding A mixture of molten thermoplastic and short fibers is injected into a colder
(short fibers, 0.1-10 mm) metal mold at very high pressure. The component is allowed to solidify
and is automatically ejected.
5. Compression molding Semi-finished sheets of glass mat thermoplastics are heated and placed in
(short fibers, 5-50 mm) the lower part of the mold in a fast press. The press is quickly closed and
pressure is applied so that the material can flow to fill the mold.
Technology is also available where the hot molding compound reaches the
mold from an extruder.
6. Compression molding The same principle as for short fiber materials.
(continuous fibers) Continuous fibers require special clamping fixtures for the sheets and can
primarily be used for simple geometries.
7. Diaphragm forming A stack of prep reg is placed inbetween two diaphragms (superplastic
aluminium or polymer film). The diaphragms are fixed whereas the
prepreg can move freely. The material is slowly deformed by external
pressure and the mold.
8. Pultrusion Prepreg tape or tape with the resin as fibers or powder is pulled through
a heated die to form beams or similar continuous structures with constant
cross-section geometry. The material is allowed to cool and solidify.
9. Resin injection Dry reinforcing fibers are placed in the mold. Monomers and/ or low
molecular weight polymer with low viscosity are injected, the
reinforcement is impregnated. Polymerization to a high molecular weight
thermoplastic occurs by mixing of reactive components and/or thermal
activation.
 Material forms 117

processing but low stiffness and strength. On Compression molding of glass mat thermo-
the other hand, materials with high fiber con- plastics (GMT) is a wide-spread process of
tent have high stiffness and strength but great significance in the automotive industry
require slow processing and are difficult to (Berglund and Ericson, 1994). Resin injection of
shape into geometrically complicated struc- polymerizing prepolymer molecules of low
tures. For high fiber content materials, the high viscosity is in principle the same process as for
viscosity of a molten thermoplastic usually thermosets although the chemical reactions
requires some kind of prepreg fabrication step lead to increased molecular weight rather than
before final processing. The prepregs may need to cross-linking. Such a process does not pro-
to be combined into the consolidated, semi-fin- vide the advantages of infinite storage life
ished sheets before the final processing step. materials with low toxicity. Diaphragm form-
Regular autoclave processing can be used ing is a processing route where the problem of
for thermoplastic composites. For most high- low extensibility of prepreg-based materials is
performance thermoplastics, however, addressed (Mallon, O'Bradaigh and Pipes,
temperatures have to be higher than the typi- 1989).
cal 177°C used for epoxy-based composites.
Often, the composite manufacturer must pur-
6.3 MATERIAL FORMS
chase a new autoclave if this is the preferred
processing route. Autoclave processing of Thermoplastic composites are usually sup-
thermoplastics has been modeled (Lee and plied as semi-finished materials, with the
Springer, 1987). Consolidation of the prepreg exception of resin injection materials. In Table
layers is an important issue. At a given tem- 6.2, material forms for thermoplastic compos-
perature, sufficient time must be available for ites are presented. Prepregs of high fiber
the polymer molecules to diffuse from one volume fractions (Vf "" 0.6) may be prepared
prepreg layer into the other and form strong by solvent-, melt-, prepolymer- or powder-
physical entanglements (Howes, Loos and impregnation of the reinforcing fibers.
Hinkley, 1989). In addition, the air initially Solvent-impregnation is limited to amorphous
present in the material must be displaced. resins with high solubility. Melt-impregnation
For thermoplastic composites, filament is a technique successfully developed by ICI
winding has demonstrated good economic (Cogswell, Hezzell and Williams, 1981) pro-
potential (Egerton and Gruber, 1988). The ducing high-quality prepreg. The resulting
major problem is in the welding of filaments or prepreg is considered too stiff, for some pro-
the tape onto the underlying composite layers. cessing situations with little drapability in
Heat has been applied by means of a gas flame, comparison with CF fEP (epoxy) prepreg. This
IR, laser beam or simply from a hot metal sur- problem is addressed in prepolymer- and
face. Pultrusion of thermoplastic composites powder-impregnated prepreg. One example is
offers potential for faster processing than with the FIT-technology where small tubes contain-
thermoset composites (Astrom, Larsson and ing reinforcing fibers and polymer powder are
Pipes, 1991), due to the absence of exothermal used (Thiede-Smet, 1989). In addition, com-
heat generation from chemical reactions. mingled weaves (prepregs) are available. The
Profiles may also be produced by roll-forming resin is present in the form of fibers which are
techniques similar to those used in metal- melted during processing to form a matrix.
working. The shape of existing profiles can be Composites produced from commingled
changed. The low-cost folding technique (GE material forms may have a fairly inhomoge-
Plastics, 1990) has been used commercially by neous distribution of fibers (Olson, 1990).
Fokker and TenCate in Holland for quite large Film-stacking is a simple method often
components of fairly simple geometry. used for preparation of laboratory samples
118 Thermoplastic resins

Table 6.2 Material forms for composites based on thermoplastic resins

Material forms Outline of preparation procedure

Prepregs
1. Solvent impregnated Reinforcing fibers are impregnated by a mixture of solvent and
thermoplastic. The solvent is removed by evaporation.
2. Melt impregnated Reinforcing fibers are impregnated by thin molten films to which pressure
is applied.
3. Prepolymer impregnated Low viscosity prepolymers are used to impregnate the fibers.
Polymerization to high molecular weight thermoplastic takes place during
processing.
4. Powder impregnated Fibers are enclosed by thermoplastic powder, either in small tubes
containing reinforcing fibers and powder or by powder particles adhering
to the fibers from partial melting in a fluidized bed.
Other material forms
5. Film stacked composites Stacks with alternating layers of dry fibers and polymer film are heated
and compressed.
6. Fiber hybridized weaves Roving of commingled reinforcing fibers and the matrix in fiber
and roving form. Weaves are produced from the roving.
7. Prepolymer liquid and ReinforCing weaves are impregnated by prepolymer liquid and
dry reinforcement polymerized to a thermoplastic composite.
8. Semi-finished glass mats Supplied as sheets prepared in belt press by extrusion and melting of
(low fiber content) films which impregnate fiber mats. Porous sheets are produced from slurry
of fibers and thermoplastic powder in water, by technology similar to
manufacturing of paper. Recycled material can be used.
9. Pellets Often prepared by pultrusion of unidirectional fibers and matrix
(low fiber content) followed by chopping into pellets. Used in injection molding or in
plasticizing and! or compounding unit combined with compression
moulding. Recycled material is easily incorporated.

(Hartness, 1982) although the technique has On the other hand, the material cost and
also been used commercially. In the category energy consumption is reduced, greater free-
of materials with low fiber volume fraction dom in materials selection is obtained and
(Vf == 0.2), semi-finished sheets of GMT-materi- recycling is facilitated. Suppliers of thermo-
als are available. They usually have random, plastic composites are listed in Table 6.3.
chopped or continuous fiber mat reinforce-
ments. Unidirectional prepreg may be used in
6.4 THERMOPLASTIC RESINS
order to selectively provide additional stiff-
ness, strength and creep resistance. An Thermoplastics have either amorphous or
interesting step forward is provided by extru- semi-crystalline structure (Sperling, 1992). The
sion compounded GMT (Composite Products large, chain-like polymer molecules do not
Inc, 1994; HoechstAG, 1994). No semi-finished show long-range order in amorphous thermo-
sheets are used, instead a special extruder is plastics, which may be viewed as polymer
used to produce a hot, soft 'cake' constituted glasses and, in the absence of color pigments,
of chopped fibers and the polymer matrix, are usually transparent. Thermosets are also
often PP. The cake is placed in a press and amorphous. In contrast, crystalline polymers
molded. The investment in technology is have regions of molecular order. In melt-
higher than for conventional GMT molding. processed crystalline polymers, a spherical
 Thermoplastic resins 119

Table 6.3 Suppliers of thermoplastic composites

Supplier Materials
Baycomp Unidirectional tapes. Matrices PP, HDPE, PA12
Burlington, Ontario, Canada PC, PEl, PBT, PES, PPS, PEEK, ABS, PPO. Fibers: glass, carbon,
aramid and stainless steel.
CYTEC, Anaheim Commingled yarns. Carbon fiber with PEEK, PEKEKK, PA6,6, TPI
CA, USA (Aurum®). GF IPA6,6.
DuPont de Nemours Prepreg based on Avimid®K, thermoplastic polyimide, and
Bad Homburg, Germany and carbon fiber. Sheets laminated of continuous fiber thermoplastic
Newark, DE, USA composites or unidirectional discontinuous fibers. Molding
compounds of lower fiber content. Matrices: PA6,6, PEKK, PET and
others. Fibers: carbon, glass and aramid.
Electrostatic Technology Prepreg fabrication by deposition of polymers in powder fonn
Branford, C, USA on tow and fabrics. Wide variety of resins and fibers.
GE Plastics GMT-materials based on glass fiber mats and PP, PBT, PC and
Amsterdam, Netherlands and blends PC/PBT. Unidirectional GF IPP.
Pittsfield, MA, USA
Hoechst Unidirectional prepreg of GF /PP, GF /PA6, GF /PE, CF /PPS,
Frankfurt, Germany CF/PA6. Pellets> 12 mm for use in plasticating extruder combined
with compression moulding.
Hills GF /PA12 fabric prepreg.
Marl, Gennany
ICI/Fiberite APC-2 (CF /PEEK) prepreg tape and tow and developmental
Monchengladbach, Gennany materials, primarily for high-temperature applications.
and Laguna Hills, CA, USA
Porcher Textile FIT-weaves (Thiede-Smet, 1989). Matrices PA12, PEl, PEEK. Glass
Lyon, France and carbon fibers. Enichem, Milano, Italy reportedly produces
GF /PP, PET, PBT with FIT-technology.
Quadrax Corp Prepreg fabrics and unidirectional tape, consolidated sheets.
Portsmouth, RI, USA Matrices PA6,6, PMMA, PEl, PPS and PEEK. Carbon, glass and
aramid fibers.
Schappe Techniques Spun yarns combining reinforcing and matrix fibers for subsequent
Chamoz, France weaving. Matrices PP, PA6, PA6,6, PPS, PC, PEl, PEEK. Carbon, glass
and aramid fibers.
SymalitAG Glass mat thennoplastic sheets based on GF /PP.
Lenzburg, Schweiz
TenCate Advanced Composites Prepreg fabrics and unidirectional tape, consolidated sheets.
Nijverdal, Netherlands and Matrices PES, PEI and PA12. Carbon, glass and aramid fibers.
Fountain Valley, CA, USA
120 Thermoplastic resins

morphology, termed spherulitic, is often amorphous sample shows a dramatic drop in

observed (Bassett, 1981). Crystalline lamellae modulus at the Tg (glass transition tempera-
are present within the spherulites although ture). The drop in modulus for semi-crystalline
disordered regions exist between and within samples is less dramatic: the higher the crys-
the lamellae. This is because the large size of tallinity, the slower the drop. Above T I

polymer molecules inhibits perfect crystalliza- material modulus is maintained by the cry;-
tion. The crystalline thermoplastics are talline phase (although strength usually
therefore more correctly described as semi- decreases dramatically). Another effect of
crystalline, since the degree of crystallinity reduced degree of crystallinity is reduced
never reaches 100%. Semi-crystalline thermo- chemical resistance.
plastics can be viewed as two-phase materials A disadvantage with semi-crystalline poly-
with a crystalline and an amorphous phase. mers is the high processing temperature, see
To illustrate the difference in behavior of Tables 6.4 and 6.5, compared with the heat
semi-crystalline and amorphous thermoplas- deflection temperature (see next section, Table
tics, polyethylene terephthalate (PET), may be 6.11). The melting temperature, Tm' of the crys-
used as an example. PET is a thermoplastic talline phase must be exceeded during
polyester which crystallizes fairly slowly. processing, although the maximum use tem-
Therefore, upon rapid cooling from the perature, as for amorphous polymers, is still
molten state, crystallization can be sup- below T.g
pressed and an amorphous polymer is Characteristic temperatures of thermoplas-
obtained (similar behaviour is shown by tics used in applications where only moderate
PEEK). Samples of PET with different degrees temperatures are experienced are presented in
of crystallinity can be produced by changing Table 6.4. These materials are available from
the conditions of cooling. The shear modu- many different chemical companies, therefore
lus G' (obtained from dynamic mechanical trade names and suppliers are not listed. In
thermal analysis, DMTA) is plotted against Table 6.5, thermoplastics for applications at
temperature for such samples in Fig. 6.1. The higher temperatures are listed. These polymers

=-:.=:::.==..:s.:~,\.
'\<.:.
'\'"
\ \ ....
\ ',\.
,\
'"
\ ' ............................. '" .

\
".
,.,., 40%

"
"

\ Increasing
0%
crystallinity

50 100 150 200 250

Temperature CC)

Fig. 6.1 Shear modulus (G') compared with temperahrre for PET of different degrees of crystallinity.
 Thermoplastic resins 121

Table 6.4 Characteristic temperatures for thermoplastic resins with Tg < 90°C

Polymer Chemical Structure Tg Tm Processing

type name (OC) (OC) temp. (OC)
Polyolefin Polypropylene Crystalline -10 165 200-240
(PP)

Polyamides Polyamide 6,6 Crystalline 55 265 270--320

(PA6,6)
Polyamide 12 Crystalline 35 180 220-260
(PA12)

Polyesters Polyethylene Crystalline 70 265 280-310

terephthalate
(PET)
Polybutylene Crystalline 20 240 260-290
terephthalate
(PBT)

Table 6.5 Characteristic temperatures for thermoplastic resins with Tg ~O°C

Polymer Chemical Trade name Structure Tg Tm Processing

type name and supplier eC) (OC) temp. (OC)
Polyester Polycarbonate Lexan, GE Amorphous 150 none 280--330
(PC) Makrolon, Bayer
Polyarylene Polyphenylene Ryton, Phillips Crystalline 90 280 300-340
ether or sulfide
sulfide (PPS)
Polyarylene PAS-2, Phillips Amorphous 215 none 330
sulfide
PEEK Victrex PEEK, ICI Crystalline 143 343 380-400
PEEKK Hostatec, Hoechst Crystalline 165 365 390-415
PEKK Declar, DuPont Crystalline 155 340 380--400
PEKEKK Ultrapek, BASF Crystalline 175 375 400-420
Polyketone Victrex HTX, lCI Crystalline 205 385 420-430

Poly- Polysulphone Udel P1700, Amoco Amorphous 190 none 300-350

sulfones (PSU)
Polyether- Victrex 4100G, lCI Amorphous 220 none 300-320
sulfone (PES) Ultrason E, BASF

Polyamide- Polyamide- TorIon C, Amoco Amorphous 275 none 350-400

imides imide (PAl) TorIon AIX-159 Amorphous 290 none 350-400
Amoco

Polyimides Polyetherimide Ultem,GE Amorphous 217 none 335-420

(PEl)
Polyimide (TPI) Avimid KIII, Du Pont Amorphous 250 none 340-360
Polyimide (TPI) Aurum, Mitsui Crystalline 260 390 400--420
Toatsu
122 Thermoplastic resins

are more expensive and are often termed 'high- of static mechanical properties of the resins is
performance' resins. The higher cost of these given in Table 6.6. Resins with low Ts'. such as
materials is due to small material volumes, PP and PA12, have lower modulus and
more expensive monomers and more difficult strength. Their fracture toughness is high and
polymerization procedures. valid data according to linear elastic fracture
Many resins used in injection molding are mechanics are difficult to obtain. Among poly-
so called blends, physical mixtures between mers with Tg well above room temperature, the
two thermoplastics. In the field of commercial modulus is fairly similar. It is controlled by
composite materials, this technology is pri- weak physical forces between the molecules.
marily used for GMT-materials, where VIscoelastic effects such as creep and stress
composites based on PC/PBT blends are relaxation during loading will affect the data.
available (Table 6.3). However, for high-per- Tensile strength varies more widely than mod-
formance resins, blending amorphous with ulus between different resins. As a material
semi-crystalline thermoplastics is an interest- property it is unfortunately not very reliable. It
ing route to improved chemical resistance. is sensitive to loading rate, specimen geometry,
Most polymer mixtures form immiscible two- specimen preparation and the presence of
phase structures although PEEK and PEl may microscopic flaws on the specimen surface. In
be mixed to form a miscible blend (Crevecoeur addition, uniaxial resin tensile strength is dif-
and Groeninckx, 1992). ferent from resin strength in the composite
where the stress state is different.
Thermoplastics have higher fracture toughness
6.5 PROPERTIES AND DESIGN
than epoxy and other thermosets, although
CONSIDERATIONS
epoxy fracture toughness can be improved by
In contrast to thermoset resins, thermoplastics addition of a thermoplastic (Bucknall and
can be dissolved and melted. In general, vis- Gilbert, 1989) or other means. Although not
coelastic and plastic effects are more apparent from the table, epoxy modulus is usu-
pronounced in thermoplastics. A presentation ally slightly higher than for thermoplastics.

Table 6.6 Mechanical properties of thermoplastic resins

Material Tensile Tensile Fracture

modulus strength toughness
E (GPa) 0'0 (MPa) Glc (kJ m-2)
PP 1.1-1.6 30-40
PA6,6 2.5-3.B 50-80
PET 2.7-4.0 50-70
PC 2.3-3.0 60-70
Amorph. PA 0-2) 3.2 100 1.6
PPS (Ryton) 3.5 BO 0.5-0.9
PAS (PAS--2) 3.2 100
PEEK 3.1-3.B 90-100 4.0
PSU (Udel P1700) 2.5 70 2.5
PES (Victrex) 2.6 BO 1.9
PAl (Torlon) 2.8-4.4 90-190 3.4
PEl (Ultem) 3.0 105 3.3
TPI (Avimid K-III) 3.5 100 1.5
TPI (LaRC-TPI) 3.7 120 1.B
EP(thermoset) V~3.5 40-120 0.1-0.5
 Properties and design considerations 123

Many mechanical properties of composites based on LaRC-TPI, J-2, PAS-2 and K-III show
are dominated by the influence of fiber mod- transverse strengths in the range 32--41 MPa.
ulus, fiber strength and fiber volume fraction. Otherwise, typical transverse strengths for
This is usually true for longitudinal tensile thermoplastic composites are in the range
modulus and strength as well as flexural 60-90 MPa. Toughened epoxy composites also
modulus and strength. For thermoplastic show fairly high transverse tensile strengths,
composites based on AS-4 carbon fiber, typi- typically around 75 MPa.
cal tensile data are: longitudinal tensile The use of transverse tension data in failure
modulus EL= 130 GPa, longitudinal tensile criteria will lead to conservative estimates.
strength a L = 1950 MPa. In the present con- Data are higher for transverse plies in multidi-
text, we are more interested in properties rectional laminates (Berglund, Varna and
dominated by the matrix and the fiber/matrix Yuan, 1991). The modulus data for carbon
interface. One such property is the transverse fiber composites in Table 6.7 appear insensi-
tensile strength of unidirectional laminates. tive to small differences in matrix modulus.
When a multidirectional laminate is loaded in Variations in fiber volume fraction and trans-
in-plane tension, the first major damage verse fiber modulus between the materials
mechanism is likely to be matrix cracking in mask any such effect. The detrimental effect of
the plies with transverse orientation to the glass fiber as opposed to carbon fiber is appar-
maximum load direction. This reduces lami- ent from the GF /PA6,6 and CF /PA6,6 data.
nate stiffness and initiates other damage GF /PP shows very poor performance, proba-
mechanisms such as delamination. In Table bly due to poor fiber/matrix interfacial
6.7, transverse strength and modulus are pre- adhesion (note the low Vf ). Interfacial weak-
sented for different thermoplastic composites. ness is also likely to explain the low strength
Fiber volume fractions are high, V£ "" 0.5-0.6. for Kevlar /PEKK. For thermoplastic compos-
For composites based on brittle epoxies, typi- ites based on AS-4 carbon fiber, Table 6.7 can
cal transverse strength is 40 MPa. Composites be used to estimate typical data: transverse

Table 6.7 Transverse tensile properties of thermoplastic composites

Material Transverse modulus Transverse strength

ET (CPa) aT (MPa)
GF /PP Plytron (Vi '" 0.35) 4 11
GF /PA6,6 E-glass/Ultramid 8.6 48
CF/PA6,6 G30-500/UltTamid 7.2 72
CF / Am. PA AS-4/J-2 9 35
CF /PPS AS-4/Ryton 7.6 72
CF /PAS T 650-42/Radel 8320 8.4 61
CF /PAS AS-4/PAS-2 8.3 32
CF/PSU AS-4/UdelP1700 7 59
K49/PEKK Kevlar/PEKK LDpM 6.2 21
CF/PEKEKK G30-500/UltTapek 10.3 90
CF /PEEK AS-4/PEEK (ICI) 8.9 80
CF /PAl T650-42/PAI-696 7.6 67
CF/TPI AS-4/K-III 9 41
CFjTPI G30-500j'NewTPI' 8.3 59
CF/TPI AS-4/LaRC-TPI 33
CF /EP AS-4/3501 (thermoset) 9 52
CF/EP HTA/6376C (thermoset) 9.9 75
124 Thermoplastic resins

tensile modulus ET "" 8.6 GPa, transverse ten- on neat resins, energy is absorbed by yielding
sile strength O"T "" 75 MPa. and other types of damage when the volume
High interlaminar toughness is desirable of material is relatively large. In a composite,
since this suppresses the tendency for delam- the presence of fibers tends to limit this mate-
ination crack formation during loading. rial volume. For shear strength, no
Interlaminar fracture toughness is determined comparable data for different thermoplastic
on double cantilever beam specimens (OCB), composites appear to be available in the liter-
usually unidirectional materials are used ature. The interlaminar fracture toughness in
(Whitney, Browning and Hoogsteden, 1982). mode II shear loading, Gnc' is higher for ther-
In Table 6.8 such data are presented. moplastic than for comparable thermoset
CF /PEEK shows the highest fracture tough- composites (Cantwell and Davies, 1993). This
ness. All thermoplastic composites show also indicates a higher shear strength for the
higher toughness than the thermoset compos- thermoplastic composites. A typical value for
ites. There is a difference between crack the in-plane shear modulus of thermoplastic
initiation and crack propagation data (Davies, composites based on AS-4 carbon fiber is
Benzeggagh and de Charentenay, 1987). The 4.8 GPa which is similar to toughened CF/EP
data presented here are crack propagation systems but slightly lower than for brittle
data; crack initiation data are in general much matrix CF /EP composites.
lower. For tough matrices one may question Compressive strength is lower for thermo-
the applicability of the data to design prob- plastic than for thermoset composites
lems. In DCB experiments, the crack opening (Table 6.9). Most of the thermoplastic compos-
displacement (COD) is very high, whereas the ites are in the range 900-1100 MPa whereas
COD at small central cracks in stiff laminates typical thermoset composite data are
is much smaller. Local stress fields and dam- 1700 MPa. Fiber misalignment, shear stiffness
age mechanisms may therefore be different and strength have been shown to affect com-
and affect the measured fracture toughness. pression strength based on plastic kink band
At present, delamination fracture toughness formation (Budiansky, 1983). Compression
from DCB tests are therefore preferentially modulus data in Table 6.9 are similar, in sup-
used to compare material. It has been pointed port of similar fiber volume fractions for the
out that composite data are significantly materials compared. The composite based on
lower than resin data (Hunston, 1984). In tests PA6,6 has the lowest strength; PA6,6 also has

Table 6.S Interlaminar fracture toughness of thermoplastic composites

Material Trade name Fracture toughness

G1C (kJm-2)
CF / Amorph. PA AS-4/J-2 1.3
CF/PPS AS-4/Ryton 0.9
CF/PEEK AS-4/Victrex PEEK 2.1
CF/PEEK IM7/Victrex PEEK 2.5
CF/PSU A5-4/Udel P1700 1.2
CF/PAI T650/Torlon AIX-159 1.3
CF/PEI T300/Ultem 1000 0.9
CF/TPI AS-4/ Avimid K-ill 1.8
CF/TPI AS-4/LaRC-TPI 0.8
CF/EP AS-4/3501-6 (thermoset) 0.2
CF/EP IM7/8551-7 (thermoset) 0.5
 Properties and design considerations 125

Table 6.9 Compressive properties of unidirectional thermoplastic composites

Material Trade name Compression Modulus

strength (MPa) (MPa)
CF/PA6.6 G30-500 IUltramid 700 110
CF I Amorph. PA AS-4/J-2 1100
CF/PPS AS-4/Ryton 940 130
CF/PAS AS-4/PAS-2 900 120
CF/PEEK AS-4/Victrex PEEK 1100 120
CF/PEEK IM7/Victrex PEEK 1140
CF/PEKEKK AS-4 I Ultrapek 1310 127
CF/PSU AS-4/Udel P1700 1040
CF/PAI C-3000/Torlon C 1380
CF/TPI AS-41 Avimid K-III 1000 110
CF/EP AS-4 I 3501-6 (thermoset) 1720 140
CF/EP HTA/6376C (thermoset) 1720 130

the lowest creep modulus and yield stress of superior performance to first generation ther-
the investigated matrices. It is interesting to moset composites (AS-4/3501-6). This is
note that AS-4/PEEK and IM-7 jPEEK have because the delaminated area due to the
roughly the same strength although the IM-7 impact event is more limited for the thermo-
fiber has higher modulus. The smaller diame- plastic composites. However, toughened
ter of the IM-7 fiber appears to have a negative epoxy resin composites combined with tough-
effect as expected from Euler-buckling consid- ened interlayers between the plies do in
erations. general show as good compression strength
Compression strength after impact, a mea- after impact as thermoplastic composites. In
sure of laminate and material damage fatigue, delamination resistance is higher for
tolerance (Dorey, 1989), is presented in AS-4jPEEK compared with epoxy composites
Table 6.10. A quasi-isotropic laminate of given (Gustafsson, 1988). However, in uniaxial ten-
lay-up and geometry is subjected to impact of sion, brittle CF /EP was found to be superior to
a certain energy. Internal damage mechanisms both toughened CF JEP and the thermoplastic
such as matrix cracking and delamination composite (Curtis, 1987). Claims have been
occur in the laminate. The plate is then sub- made that this observation is due to heating
jected to compressive load and the stress and effects in the thermoplastic composite speci-
strain at failure can be determined. The data mens from testing at high frequency (Moore,
show that thermoplastic composites have 1991).

Table 6.10 Compression strength after impact of thermoplastic composites

Material Trade name Compression strength after impact

(impact energy)
(28 J) (42 J) (57 J)
cr (MPa) cr (MPa) cr (MPa)
CF/PPS AS-4 I Ryton 221 179
CF/PEEK AS-4/Victrex PEEK 331 310 290
CF/PAI C-3000/Torlon C 365 345 317
CF/EP AS-4/3501-6 (thermoset) 179 145 131
CF/EP AS-4/8551-7 (thermoset) 303
126 Thermoplastic resins

Table 6.11 Glass and heat deflection temperatures Table 6.12 Glass melting and heat deflection tem-
for amorphous thermoplastics peratures for semi-crystalline thermoplastics

Material Tg (0C) HDTeC) Material Tg (0C) TmeC) HDTeC)

PC 150 132 PP -10 165 60
Amorph.PA 160 154 PA6.6 55 265 75
PSU 190 175 PET 70 265 41
PES 220 203 PPS 90 280 135
PAl 290 278 PEEK 143 343 160
PEl 217 200 TPI (Aurum) 250 388 238
EP (thermoset) 200 180

Increased market need for polymer compos- is that molecular mobility in the polymer is
ites with good performance at elevated increased dramatically as the temperature
temperature has generated interest in thermo- approaches T .
plastic composites. Materials with continuous In Table g.12, similar data to those in
use temperatures above 150°C are of particular Table 6.11 are presented for semi-crystalline
interest since they perform better than epoxies. thermoplastics. HOT is usually somewhat
One question is how maximum use tempera- higher than T. However, for some semi-
ture relates to Tg• In Table 6.11, Tg and the heat crystalline pol~ers, HDT is below T (as for
deflection temperature (HDT) for amorphous amorphous polymers). Creep effects gwill be
thermoplastics are presented. HOT is deter- very strong close to and above T . For this rea-
mined by subjecting the material to static load son the maximum temperature tor continuous
(typically 1.8 MPa) and slowly increasing the service under significant load is unlikely to
temperature. HDT is determined as the tem- exceed a temperature of 20°C below T for
perature at which a critical deflection of the semi-crystalline thermoplastics. The pruhary
sample is obtained. Table 6.11 shows HDT to advantage of crystallinity is therefore chemical
be 6-20°C below the T of amorphous poly- resistance. This is apparent from Table 6.13,
mers. In comparison w~th room temperature where chemical resistance for different ther-
strength, the strength of the composite is sig- moplastics is indicated in a qualitative way.
nificantly reduced above the HDT. The reason Semi-crystalline thermoplastics have much

Table 6.13 Chemical resistance of thermoplastics

Material Structure Hydraulic Chlorinated Ketones Esters Hp abs.

fluid hydro- (%)
carbons
PA6,6 Crystalline 0 0 0 8
PEEK Crystalline 0 0 0 0 0.5
PPS Crystalline 0 0 0 0 0.5
PEl Amorphous 0 0 A A 1.2
PSU Amorphous A 0 A A 0.9
PAl Amorphous 0 0 0 0 2-4
PES Amorphous A 0 A A 0.3
Am.PA Amorphous A A A 5
PC Amorphous 0 A A

0= no effect, A = is absorbed, 0 = is dissolved

 Applications 127

better chemical resistance than the amorphous induction welding. There are difficulties in con-
polymers. A notable exception is the high trolling the processes and very few methods
water absorption in PA6,6 caused by offer the promise of portable equipment.
hydrophilic groups in its chemical structure. Various bonding technologies for PEEK com-
The solvent resistance of a large selection of posites have been compared (Silverman and
different thermoplastic composites has been Criese, 1989). The study favored a technique for
reported Gohnston, Towell and Hergemother, fusion bonding with a polyetherimide film.
1991). For most thermoplastics, e.g. PEEK, This approach can also be used in repair of
moisture expansion coefficients of the carbon damaged structures, since the temperature
fiber composite may be taken as O. Thermal needed is below the Tm of PEEK.
expansion coefficients have been characterized
(Barnes et aI., 1990) and are similar to epoxy
6.6 APPLICATIONS
composites. For AS-4/PEEK, the composite
density is 1600 kg m-3. Thermoplastic composites can be used in sim-
For polymers with high Tg' exposure to ele- ilar applications to thermoset composites. The
vated temperature may lead to increased following examples will demonstrate some of
density not connected with crystallinity but the reasons for choosing a thermoplastic com-
with the amorphous state. The phenomenon is posite material. In Europe, the automotive
termed physical aging and leads to a more market for GMT composites is significant.
brittle behavior of the polymer (Kemmish and Rapid processing by compression molding,
Hay, 1985). Further work is needed to eluci- cycle times of typically 30 s even for large
date the importance of physical aging to structures, results in cheaper components. In
composite fracture behavior under practical addition, more functions can be integrated
service conditions. For composites processed into each component compared with sheet
at high temperatures, residual stresses will metal structures. Bumper beams dominate the
also affect fracture behavior. The magnitude of automotive market in the US whereas, in
the residual stresses and, consequently, detri- Europe, a wider variety of applications are in
mental effects will increase with increasing commercial production. Typical components
cooling rate (Manson and Seferis, 1992). are subjected to minor loads or impact and
Thermoplastic composites can be joined by surface appearance is not important. Battery
the same methods as thermoset composites. trays, beams supporting the hood, seat sup-
Bolted joint performance has been compared ports, oil trays, engine shields and even the
for thermoset and thermoplastic composites complete front end have been produced for
(Walsh, Vedula and Koczak, 1989) with results Volvo, Volkswagen and others.
in favor of thermoplastic composites. With the In the aerospace market, DuPont has sup-
semi-crystalline thermoplastics, adhesive plied thermoplastic polyimide composites to
bonding requires careful surface preparation the prototype programs for the F-22 fighter
(Kinloch and Taig, 1987). This is because of the aircraft. In the supersonic civil aircraft pro-
good chemical resistance and limited solubil- gram, the same thermoplastic polyimide is
ity of these polymers. However, with careful considered for wing skins. The main reason
surface preparation, as good adhesive bonds for this particular thermoplastic composite is a
are obtained as with thermoset composites. continuous high maximum temperature.
The thermoplastic nature of the matrix offers AS-4/PEKK (unidirectional discontinuous
another possibility: fusion bonding. Different fibers) was used by Bell Helicopter Textron in
methods have been compared (Davies and a V-22 tiltrotor thermoplastic wing rib for bet-
Cantwell, 1993), including hot gas, IR, laser, ter open hole compression behavior than
ultrasonic, vibration, electrical resistance and thermoset composites for high proportions of
128 Thermoplastic resins

Bike helmets are also produced (Fig. 6.2). A

semi-finished sheet is heated by IR, transfered
to heated dies where a selective clamping sys-
tem is used to hold the sheet. Forming and
consolidation pressure is applied and the
mold is cooled before demolding. Hoechst in
Germany present applications made by wind-
ing, a pressure vessel, tubes, sealings and
support rings (Fig. 6.3). Matrices include PA6
and PPS with carbon and glass fibers.
In the field of biomedical applications, such
as hip prostheses, semi-crystalline thermo-
plastic matrices offer good potential due to
their chemical stability (Williams and
Fig. 6.2 Bike helmet, an example of a commercial McNamara, 1987).
application of thermoplastic composites (Courtesy:
DuPont Europe).
REFERENCES
Astrom, BT, P.H. Larsson and R.B. Pipes. 1991.
± 45° ply angles (probably for decreased ten-
Experimental Investigation of a Thermoplastic
dency to delamination). A female steel tool Pultrusion Process. 36th Int. SAMPE Symp., pp.
and matched-die press forming was used 1319-30.
(Chang, 1992). A stretch forming process was Barnes, J.A, Sims LJ., Farrow G. et al. 1990. Thermal
used to fabricate C-section curved fuselage Expansion Behaviour of Thermoplastic
frames for Boeing Helicopter from AS- Composite Materials. J Thermoplastic Composites
4/PEKK (Chang, 1992). TenCate in Holland 3: 66-80.
Bassett, D.C. 1981. Principles of Polymer Morphology.
supplies materials for Fokker Special London: Cambridge University Press.
Products. Landing flap ribs and impact resis- Berglund, L.A and Ericson M.L. 1994. Glass Mat
tant ice-protection plates are produced for the Reinforced Polypropylene. In Polypropylene:
Dornier 328 aircraft. Structure, Blend and Composites. Vol 3:
Composites, ed. J. Karger-Kocsis. London:
Chapman and Hall.
Berglund L.A, Varna J. and Yuan J. 1991. Effect of
Intralaminar Toughness on Transverse Cracking
Strain in Cross-Ply Laminates. Adv. Compo
Mater. 1: 225-34.
Bucknall, C.B. and AH. Gilbert. 1989. Toughening
of a Tetrafunctional Epoxy Resins Using
Polyetherimide. Polymer. 30: 213-18.
Budiansky, B. 1983. Micromechanics. Computers and
Structures. 16: 6-10.
Cantwell, W.J. and Davies P. 1993. Short-Term
Properties of Carbon Fibre PEEK Composites.
In Advanced Thermoplastic Composites, ed.
Kausch, H-H. pp. 173-191. Munich: Hanser.
Carlsson, L.A, ed. 1991. Thermoplastic Composite
Materials. Composite Materials Series, Vol 7.
Fig. 6.3 Thermoplastic composite applications New York: Elsevier.
made by winding: pressure vessel, tubes, sealing Chang, LY. 1992. Thermoplastic Matrix Composites
and support rings. (Courtesy: Hoechst AG) Development Update. 37th Int SAMPE Symp.
 References 129

Available from DuPont as report entitled Composites. In Advances in Thermoplastic

Overview of Thermoplastics Composites Technology. Composite Materials. ed. G.M. Newaz, ASTM
Cogswell, EN. 1992. Thermoplastic Aromatic Polymer STP 1044, pp. 33-49. Philadelphia: American
Composites. Oxford: Butterworth-Heinemann. Society for Testing and Materials.
Cogswell, EN., Hezzell D.J. and Williams P.J. 1981. Johnston, N.J., Towell T.W. and Hergenrother P.M.
US Patent 4 549 920. 1991. Physical and Mechanical Properties of
Composites Products Inc 1994. Brochure of CPI High-Performance Thermoplastic Polymers
process. Winona, Minnesota, USA and their Composites. In Thermoplastic
Crevecoeur, G. and Groeninckx G. 1992. Melt- Composite Materials. ed. L.A Carlsson,
Spinning of in situ Composites of a Composite Materials Series, Vol 7, pp. 27-71.
Thermotropic Liquid Crystalline Polyester New York: Elsevier.
(TLCP) in a Miscible Matrix of Polyether ether- Kausch, H-H, ed. 1993. Advanced Thermoplastic
ketone (PEEK) and Polyether imide (PEl). Composites. Munich: Hanser.
Polymer Composites 13: 244--50. Kemmish, D.J. and Hay J.N. 1985. The Effect of
Curtis, P.T. 1987. In Investigation of the Tensile Physical Ageing on the Properties of
Fatigue Behaviour of Improved Carbon Fibre Amorphous PEEK. Polymer 26: 905--12.
Composite Materials. 6th ICCM and 2nd ECCM Kinloch, AI. and Taig CM. 1987. The Adhesive
Proc., Vol 4. pp. 4.54-4.64. London: Elsevier Bonding of Thermoplastic Composite. J.
Applied Science. Adhesion 21: 291-302.
Davies, P., Benzeggagh M.L. and de Charentenay Lee, w.I. and Springer G.S. 1987. A Model of the
F.X. 1987. The Delamination Behavior of Carbon Manufacturing Process of Thermoplastic Matrix
Fiber Reinforced PPS. SAMPE Quarterly 19: Composites. J. Comp Mater. 21: 1017-55.
19-24. Manson J-A.E. and Seferis J.C 1992. Process
Davies, P. and Cantwell W.J. 1993. Bonding and Simulated Laminate (PSL): A Methodology for
Repair of Thermoplastic Composites. In Internal Stress Characterization in Advanced
Advanced Thermoplastic Composites, ed. Kausch, Composite Materials. J. Comp Mater. 26: 405--31.
H-H, pp. 337-366. Munich: Hanser. Mallon, P.J., O'Bradaigh CM. and Pipes RB. 1989.
Dorey, G. 1989. Damage Tolerance and Damage Polymeric Diaphragm Forming of Continuous
Assessment in Advanced Composites. In Fibre Reinforced Thermoplastic Matrix
Advanced Composites, ed. LK Partridge, Composites. Composites 20: 48-56.
pp. 369-98. London: Elsevier Applied Science. Menges, G. 1966. The reinforcement of plastics.
Egerton, M, and Gruber M. 1988. Thermoplastic Kunststoffe 56: 818--23. (In German).
Filament Wmding Demonstrating Economics Moore, D.R 1991. In Thermoplastic Composite
and Properties Via In Situ Consolidation. 33rd Materials. ed. L.A. Carlsson, Composite
Int. SAMPE Symp., pp. 35-46. Materials Series, Vol 7, pp. 331-69. New York:
GE Plastics, 1990. Technopolymer Structures, Elsevier.
Design and Processing guide. TPS-400A Olson, S.H. 1990. Manufacturing with
(6/90)RTB. Commingeled Yarns, Fabrics and Powder
Gustafsson, C-G. 1988. Initiation and Growth of Prepreg Thermoplastic Composite Materials.
Fatigue Damage in Graphite/Epoxy and 35th Int SAMPE Symp., pp. 1306--19.
Graphite/PEEK Laminates. PhD thesis 88--9. Silverman, E.M. and Criese RA 1989. Joining
Royal Inst of Techn, Stockholm, Sweden. ISSN Methods of Graphite/PEEK Thermoplastic
0280-4646. Composites. SAMPE Journal 25: 34-38.
Hartness, J.T. 1982. Polyether-etherketone Matrix Sperling, L.H. 1992. Introduction to Physical Polymer
Composites. 14th Natl SAMPE Symp., pp. 26-43. Science. New York: John Wiley & Sons.
Hunston, D.L. 1984. Composite Interlaminar Thiede-Smet, M. 1989. Study of Processing
Fracture: Effect of Matrix Fracture Energy. Parameters of PEEK/Graphite Composite
Composites Technology Review 6: 176--80. Fabricated with 'FIT' Prepreg. 33rd Int SAMPE
Hoechst AG. 1994. Compel® product brochure. Symp., pp. 2086--2101.
Frankfurt am Main, Germany. Truckenmueller, F, and Fritz H.G. 1991. Injection
Howes, J.c., Loos AC and Hinkley J.A. The Effect Moulding of Long Fibre-Reinforced
of Processing on Autohesive Strength Thermoplastics: A Comparison of Extruded and
Development in Thermoplastic Resins and Pultruded Materials with Direct Addition of
130 Thermoplastic resins

Roving Strands. Polymer Eng. Sci. 31: 1316-29. Composite Materials. ]. Reinforced Plas. and
Walsh, R, Vedula M. and Koczak M.J. 1989. Camp. 1: 297-313.
Comparative Assessment of Bolted Joints in a Williams, D.E and McNamara A. 1987. Potential of
Graphite Reinforced Thermoset vs Polyetheretherketone (PEEK) and Carbon-
Thermoplastic. SAMPE Quarterly 20: 15-19. Fibre-Reinforced PEEK in Medical
Whitney, J.M., Browning c.E. and Hoogsteden W. Applications. ]. Mater. Sci. Lett. 6: 188-90.
1982. A Double Cantilever Beam Test for
Characterizing Mode I Delamination of
FIBERGLASS REINFORCEMENT* 7
Dennis J. Vaughan

7.1 INTRODUCTION (2300°F). The molten glass then flows directly

The history of glass is ancient, but its engi- to the fiber-drawing furnace in a direct melt
flow process (Fig. 7.1) or into a marble making
neering scientific development is recent.
machine. These marbles can be sorted and can
Glass was first produced some 4000 years
eventually be remelted and drawn into fibers.
ago, probably in Egypt in the furnaces used to
Continuous glass fibers are produced when
produce pottery. Its first application was as a
molten glass from the fiber-drawing furnace is
form of adornment in jewelry and as added
decoration to vases and drinking vessels. ?ravity fe.d through numerous tiny openings
The use of glass in fiber form dates back to m a platinum alloy tank called a bushing
the early seventeenth century when the (Fig. 7.2). The droplets of molten glass that
ex~de from each of the bushing's openings
Venetians utilized it to create specialized
gowns. However, commercial fiberglass did (FIg. 7.3) are gathered together, mechanically
attenuated to the correct dimensions, passed
not become a reality ~til in 1939 the joint
research efforts of Owen-illinois and Corning through a water spray and over a revolving
Glass Works, resulted in the formation of belt .that applies a protective and lubricating
coating known as a size or binder. The fibers
Owens-Corning Fiberglas Corporation.
Textile fiberglass has now grown into a are then gathered together in a suitably
shaped shoe to form a bundle called a strand
multi-million dollar industry. Glass fiber can
be obtained as a continuous fiber on staple or which is wound onto a core at approximately
discontinuous fiber. Both forms are made by 190 km/h (120 mile/h). This package of fibers
the same manufacturing process until the fiber is then dried or conditioned prior to further
drawing operation. processing and eventually sold as a continu-
ous filament yarn.
Staple fibers are produced by passing a jet of
7.2 FIBERGLASS PRODUCTION air across the openings at the base of the bush-
The production of glass fibers starts with the ing, which pulls individual fibers of
dry mixing of silica sand and limestone, boric approximately 20-40 cm (S-15 in) long from the
~olten g1as~ that is extruding from each open-
acid and a number of other products such as
mg. These filaments are collected on a rotating
clay, coal and fluorspar. These materials are
melted in a high-refractory furnace, the temper- vacuum drum, sprayed with size and gathered
ature of the melt being dependent on the glass into a strand. This package of filaments is again
composition, but is generally about 1260°C conditioned or dried prior to processing into a
specific product for further use.

Handbook of Composites. Edited by S.T. Peters. Published .. Significant .portions of this article appeared in Handbook
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 of CompOSItes, 1982, (G. Lubin ed.) Van Nostrand
Reinhold, New York.
 132 Fiberglass reinforcement

CLAY
RAW MATERIALS
LIMESTONE SILICA SAND BORIC ACID
COAL FLUORSPAR o HOPPER
BINDER
FORMULATION
R
U
PLATINUM
BUSHINGS
Ee:":a (
TANK

INSPECTION
AND
WEIGHING

HIGH SPEED
WINDER

[· · ·1~o:~I
" ___ · 4.
-l.
DECORATIVE AND
INDUSTRIAL YARN

Lr. ~ "~,,. "' - r' ~

INSPECTION
AND
WEIGHING

~,~, J
1
'000

L-I~ ~
.-+IW---Q_~.:f.C~~
0 Sj~~:;~J~
r' · · · ·
OVEN HEAT TREATING.~ "c::=j) _____••
STRAND
CHOPPING

L.1§ PACKING
A capsule view of the fiber glass
manufacturing process for yarn,
roving and chopped strand takes
the fiber !rom raw materials batch
stage to finished products.

Fig. 7.1 Direct-melt fiberglass manufacturing process.

 Glass composition 133

Fig. 7.2 Molten glass flows from tiny orifices in Fig. 7.3 Molten glass flows from tiny orifices in
platinum bushings (Courtesy of PPG industries). platinum bushings (Courtesy of Owens-Corning
Fiberglas Industries).
Each individual fiber is drawn from the
bushing opening and must be controlled so that Table 7.1 Fiberglass filament designations
reproducible filaments, strand dimensions and
properties are obtained. This control is Filament Filament diameter
achieved by the regulation of the melt viscosity, designation inxlO- 4
temperature and drawing speed. It is possible, B 1.5 3.8
therefore, to obtain a large number of filament C 1.8 4.5
diameters by varying the number of openings D 2.1 5
in the bushing and the drawing conditions. DE 2.5 6
As demand has dictated over the years, the E 2.9 7
fiberglass industry has established a number G 3.6 9
of standard filament diameters (Table 7.1). H 4.2 10
K 5.1 13

7.3 GLASS COMPOSITION

arrangement is conducive to formation of an
Glass is generally defined as an amorphous intricate three-dimensional network of oxygen
material, being neither solid or liquid. tetrahedra with a silicon atom in the middle,
Chemically glass is made up of elements such bonded to each oxygen atom. Silicon by itself
as silicon, boron and phosphorus that are con- requires extremely high temperatures for lique-
verted into glass when combined with oxygen, faction. Therefore, other elements are added to
sulfur, tellurium and selenium. The molecular the mix to reduce temperatures and to produce
134 Fiberglass reinforcement

a viscosity in the molten glass that will allow improve dielectric loss properties in electri-
easy drawing. cal applications (similar in performance to
A number of glass compositions are avail- D-glass).
able depending on the properties desired from • Hollow fiber: A special glass whose fibers
the resulting fibers (Table 7.2). are tube-like or hollow; the material has
specific applications in reinforced aircraft
• A-glass: A high alkali or soda glass is made
parts where weight could be significant.
into fibers for use in applications where
• Te glass: A Japanese manufactured 5-g1ass,
good chemical resistance is needed.
for higher strength structural application.
• E-glass: A low alkali glass, based on alu-
minum borosilicate. This glass possesses
7.4 FIBERGLASS PROPERTIES
excellent electrical insulation properties
and is the premium fiber used in the major- The composition of the original glass melt
ity of textile fiberglass production. probably plays the biggest role in determining
• C-glass: A material based on soda borosili- the properties of the fiberglass. The continuing
cate that produces a fiber that offers widespread use of fiberglass in numerous and
excellent chemical resistance. diverse applications can be directly related to
• 5-2 glass: This glass is made up of magne- its inherent unique properties (Table 7.3). The
sium, aluminum silicate and offers higher suppliers of these materials in the USA are
physical strength. Fibers produced from shown in Table 7.4.
this glass have an approximate forty per-
• High tensile strength: Fiberglass has an
cent tensile strength improvement over
exceptionally high tensile strength com-
those of E-glass composition.
pared with other textile fibers. Its strength
• D-glass: This fiber made from a low dielec- to weight ratio exceeds steel wire in some
tric composition has dielectric loss
applications.
properties (dielectric constant of 3.8 at
• Heat and fire resistance: Because fiberglass
1 me S-1) superior to that of E-glass (6.0 at is inorganic it does not burn or support
1 me S-l).
combustion.
• R-glass: A special glass composition that • Chemical resistance: Fiberglass has excel-
produces fiber that is alkali resistant and is
lent resistance to most chemicals and is
used in reinforcing concrete.
impervious to fungal, bacterial or insect
• Low K: An experimental fiber produced to attack.
Table 7.2 Fiberglass compositions (wt.%)

Components Grade of glass

A C E S
(high alkali) (chemical) (electrical) (high strength)
Silicon oxide 72.0 64.6 54.3 64.2
Aluminum oxide 0.6 4.1 15.2 24.8
Ferrous oxide 0.21
Calcium oxide 10.0 13.2 17.2 0.01
Magnesium oxide 2.5 3.3 4.7 10.27
Sodium oxide 14.2 7.7 0.6 0.27
Potassium oxide 1.7
Boron oxide 4.7 8.0 0.01
Barium oxide 0.9 0.2
Miscellaneous 0.7
 Fiberglass properties 135

Table 7.3 Properties of fiberglass

Grade of glass
A C E S
Physical properties
Specific gravity 2.50 2.49 2.54 2.48
Mohs hardness 6.5 6.5 6.5
Mechanical properties
Tensile strength, psi X loa
(MPa)
At 72°P (22°C) 440 (3033) 440 (3033) 500 (3448) 665 (4585)
At 7000P (371°C) 380 (2620) 545 (3758)
At 10000 P (538°C) 250 (1724) 350 (2413)
Tensile modulus of elasticity at
72°P (22°C), psi X 106 (GPa) 10.0 (69.0) 10.5 (72.4) 12.4 (85.5)
Yield elongation, % 4.8 4.8 5.7
Elastic recovery, % 100 100 100
Thermal properties
Coefficient of thermal linear
expansion, 0p-1 X 1(}-'i
eC-1) 4.8 (8.6) 4.0 (7.2) 2.8 (5.0) 3.1 (5.6)
Coefficient of thermal
conductivity, Btu in h-1 £t-2 °P-1
(Wm-1K-1) 72 (10.4)
Specific heat at 72°P (22°C) 0.212 0.197 0.176
Softening point, °P (0C) 1340 (727) 1380 (749) 1545 (841)
Electrical properties
Dielectric strength, V I mil 498
Dielectric constant at 72°P (22°C)
At 69 Hz 5.9-6.4 5'(}-5.4
At 106Hz 6.9 7.0 6.3 5.1
Dissipation (power) factor at 72°P (22°C)
At 60 Hz 0.005 0.003
At H)6Hz 0.002 0.003
Volume resistivity at 72°P (22°C)
and 500 V DC, ohm-em 1015 1016
Surface resistivity at 72°P (22°C)
and 500 V IX, ohm-cm 1013 1014
Optical properties
Index of refraction 1.547 1.523
Acoustical properties
Velocity of sound, ftls (m/s) 17500 (5330) 19 200 (5850)

• Moisture resistance: Because fiberglass does • Thermal properties: Due to its low coeffi-
not absorb water, it neither swells, stretches cient of thermal linear expansion and high
nor disintegrates. Fiberglass does not readily coefficient of thermal conductivity, fiber-
rot and continues to maintain its mechanical glass exhibits excellent performance in
strength in humid environments. thermal environments.
136 Fiberglass reinforcement

Table 7.4 Summary of available fibers from manufacturers

Type of glass Supplier

OCF PPG Nitto Boeski Vetrotex NEG
E x x x x x
5 x x
5-2 x
D x
R x x
A x
C x
LowK x
Te x
Hollow fiber x x

• Electrical properties: Fiberglass being non- 7.S FIBERGLASS TYPES

conductive is an ideal choice for electrical
insulation, where designers can make use of 7.5.1 FIBERGLASS ROVING
the high dielectric strength and low dielec- Fiberglass roving is a collection of parallel con-
tric loss properties. tinuous strands or filaments. Conventional
The user may take advantage of one or more rovings are manufactured by winding together
of the above properties in manufacturing rein- the number of single strands necessary to pro-
forced composites. By utilizing both the high duce the required yield (the number of yards of
strength and the excellent electrical properties roving weighing one pound). Single strand
of glass fiber, the aircraft industry has found roving, as the name implies, consist of a single
fiberglass an excellent reinforcement for strand of fiberglass filaments. These filaments
radome applications. are drawn from a bushing that has the correct
The printed circuit board industry has used number of openings, so that the single strand
the combination of electrical properties and will have the correct yield.
superior dimensional stability of fiberglass to Rovings are generally manufactured from
manufacture circuit boards that can be used G or K filaments (see Table 7.1), larger diame-
under the various adverse environmental con- ters are also available. Roving yields can vary
ditions. from 1800 to 225 yd/lb (276 to 222 Tex). The
In numerous applications glass reinforce- use of the word Tex indicates a unit of fiber
ment is chosen because it allows composites to fineness that is assessed by the weight in
retain maximum properties in high moisture grams of 1000 m of yarn, the lower the num-
environments. E-glass fibers are the material ber, the finer the yam.
of choice because of their excellent water resis-
tance. (E-glass fibers exhibit only a 1-7%
7.5.2 WOVEN ROVING
weight loss when exposed to boiling water for
60 min.) This water resistance helps maintain A number of rovings are woven into heavy,
the physical and electrical characteristics of coarse weave fabrics for use in materials that
the composite over prolonged time exposure require a quick build-up of composite thick-
to aqueous atmospheres!. ness over relatively large areas, such as various
marine products and different types of tooling.
Woven rovings are manufactured in
weights ranging from 400-1400 g/m2 (12-40
 Fiberglass types 137

oz/yd2) and thicknesses of from 0.7-1.5 mm wet layup techniques or by impregnating and
(0.02--0.04 in) (see Table 7.5). applying low pressure {RTM} (see Table 7.5).
The fiberglass ravings can be used in con-
junction with polyester resins in hand lay-up 7.5.3 FIBERGLASS MAT
techniques that would typically be used in the
The three basic forms of fiberglass mat, are
manufacture of boats. Woven roving reinforced
chopped strand, continuous and surfacing or
laminates can be made by using conventional
veil.
Table 7.5 Woven roving and polyester laminate data

Woven roving constructions and properties

Count, Weight, Thickness, Weave
perin oz/yd2 in
(per em) (g/m2) (mm)
5x8 18.0 0.031 Plain
(2 X 3.2) (610) (0.787)
5x8 24.0 0.038 Plain
(2 X 3.2) (814) (0.965)
5x6 30.0 0.049 Plain
(2 X 2.4) (1020) (1.24)
5x8 36.0 0.052 Plain
(2 X 3.2) (1220) (1.32)

Laminate properties - 240z (814 g) woven roving reinforced

Resin type" FREpoxy FREpoxy GPEpoxy
Resin content, wt.% 49.0 45.6 42.5
Thickness, in (nun) 0.299 (7.59) 0.265 (6.73) 0.263 (6.68)
Cureb RT P+T P+T
Flexural strength, psi (MPa)
Cond.N 34 200 (235.8) 49 BOO (343.4) 73 100 (504.0)
Condo D2/100d 38 800 (267.5) 44 400 (306.1) 59600 (410.9)
Flexural modulus, psi (GPa)
Cond.A 20.6 X 106 (14.2) 2.41 X 106 (16.6) 2.92 X 106 (20.0)
Condo D2/100 2.12 X 106 (14.6) 2.31 X 106 (15.9) 2.83 X 106 (19.5)
Compressive strength, psi (MPa)
Cond.A 29 700 (204.8) 25 700 (177.2) 47300 (326.1)
Condo D2/100 27300 (188.2) 25 200 (173.8) 40 100 (276.5)
Tensile strength, psi (MPa)
Cond.A 42000 (289.6) 48 900 (337.2) 51 400 (354.4)
Condo D2/100 40600 (279.9) 47 400 (326.8) 50400 (347.5)

"FR = Fire Retardant; GP = General Purpose

bRT =Room Temperature; P + T =Pressure and Elevated Temperature
<cond. A = as-received condition
dCond. D2/100 = immersion for 2 h in water at 100°C.
138 Fiberglass reinforcement

Chopped strand mat is a non-woven mater- the spirals conform in shape to the central
ial which the fiberglass strands are chopped position of the letter '5'. Alternatively, a yam
into 2-5 em (1-2 in) lengths and evenly distrib- is said to have a 'z' twist if the spiral con-
uted at random onto a horizontal plane, bound forms in shape to the central position of the
together with some type of chemical size. The letter 'Z'.
mats so produced weigh from 2.6-12 g/m2 Strands that have a simple twist (greater
(0.7-3.0 oz/ft2) and are available in a variety of than one turn per inch) will kink, corkscrew
widths, 6~230 cm (24-90 in) with 9.6 em (38 and unravel, because their twist is in only one
in) width being the most typical. direction. The plying operation will normally
Continuous strand mat is made of eliminate this problem by countering the twist
unchopped continuous strands of fiberglass in a twisted 'singles' yam with an opposite
deposited and interlocked in spiral fashion. twist in the plied yam. For instance, a 'singles'
This mat is open and flexible, but due to its yams which have a 'Z' twist are plied with an
mechanical interlocking does not require '5' twist that will result in a balanced yam. The
much application of size to achieve adequate twisting and plying operations allow the
handling strength. yam's strength, diameter and flexibility to be
Surface mat, or veil as it is sometimes varied, which in tum allows scope in produc-
called, is a very thin mat of single continuous ing a wide range of suitable woven fabrics.
filaments used frequently as a decorative sur-
face reinforcing layer in hand lay-up or press
7.5.5 TEXTURED YARN
molding processes in order to minimize sur-
face defects and to prevent 'weep' in 'Singles' or plied yam can be textured by using
pressurized wound tanks. a jet of air directed onto the yam's surface
which results in random but controlled break-
age of surface filaments, producing a general
7.5.4 TEXTILE FIBERGLASS YARN increased loft to the fiber surface. The term
A yam can be described as an assembly of used to describe such a process is texturizing
fibers or strands which can be woven into or bulking and the degree to which it occurs
some form of textile material. can be controlled by regulating the air pres-
The continuous individual strand as it sure and yam feed rate. The texturing process
emerges from the bushing opening represents opens the fiber bundles resulting in some
the simplest form of textile fiberglass yam and mechanical damage to the surface filaments.
is designated 'singles' yam. However, to The increased surface area allows higher resin
ensure the correct and efficient utilization of absorption during impregnation and produces
this yam in a weaving operation, additional low glass-to-resin ratios, resulting in more eco-
strand integrity has to be added by twisting nomicallaminates.
the yam slightly (less than one turn per inch).
A number of woven fabrics, however, require
7.5.6 YARN NOMENCLATURE
heavier yams than can be drawn from the
bushing. These heavier yams can be achieved Because of the wide variety of fiberglass yarns
by combining single strands using a twisting produced, it is necessary to have a precise sys-
and plying operation. This simply involves tem for yam identification. The standard
twisting one or more single strands together or fiberglass yam nomenclature is based on both
subsequently plying together two or more of alphabetical and numerical designations.
the twisted strands. The first alphabetical letter identifies the
A yam can be defined as having an '5' glass composition (E-glass), the second letter
twist if when it is held in a vertical position, indicates filament type (C = continuous) and
 Fiberglass types 139

the third and fourth letters identify filament ECG 150 4/2 3.85
diameter (E 7 micron) [for details see Table 7.1]. E = E-glass; C = Continuous filament; G =
The first series of numbers in the numerical Filament diameter
designation represents 1/100th of the basic 1504/2 =Four basic strands of 150 1/0 are
strand yield, the second number series speci- twisted together to form 1504/04
fies the number of single strands twisted 150 4/2 3.85 = Plying two strands of EGG
together and the number of twisted yarns 1504/04 (using '5' twist to create balance)
plied together. The total number of basic
The above yarn contains 8 (4 x 2) basic 150
strands in a plied yam is found by multiplying
strands with a glass yield of 1875 (15 000 -:- 8)
these two figures. The yield of the yarns is
yards/pound.
obtained by dividing the basic strand yield by
The nomenclature and typical properties of
the total number of strands in the yam. A third
commercially available weaving yams can be
number combined with either the letter '5' or
seen in Table 7.6.
'Z' (designating the type of twist) will some-
times also be included. For example:

Table 7.6 Commercially available fiberglass weaving yams

Yarn designation Yield Minimum breaking strength

Glass system rex system yd/lb rex Ib N
E-glass Yams
ECD 1800 1/0 EC52.751 xO 180000 2.75 0.25 1.11
ECD 1800 1/2 EC52.751 X2 90000 5.5 0.5 2.22
ECD900 1/0 EC55.51 xO 90000 5.5 0.5 1.11
ECD9001/2 EC5 5.51 x2 45000 11 1.1 4.89
ECD4501/0 EC5111xO 45000 11 1.1 4.89
ECD4501/2 EC5111 x2 22500 22 2.2 9.79
ECD4501/3 .EC5111x3 15000 33 3.3 14.7
ECD4502/2 EC5112x2 11250 44 4.4 19.6
ECD 450 3/2 EC5113x2 7500 66 6.6 29.4
ECD2251/0 EC5221xO 22500 22 2.2 9.78
ECD2251/2 EC5221x2 11250 44 4.4 19.6
ECD2251/3 EC5221x3 7500 66 6.6 29.4
ECD2252/2 EC5222x2 5625 88 8.8 39.1
ECD 225 2/3 EC5 222 x3 3750 132 14.4 64.1
ECDE 150 1/0 EC6331 xO 15000 33 3.5 15.6
ECDE 150 1/2 EC6331x2 7500 66 7.0 31.1
ECG 150 1/0 EC9331xO 15000 33 3.5 15.6
ECG 150 1/2 EC9331x2 7500 66 7.0 31.1
ECG 150 1/3 EC9331x3 5000 99 9.0 40.0
ECG 150 2/2 EC9332x2 3750 132 12.0 53.4
ECG 1502/3 EC9332x3 2500 196 18.0 80.1
ECG 1502/4 EC9332x4 1875 264 24.0 107
ECG 1503/3 EC9333x3 1667 297 27.0 120
ECG 1504/4 EC9334x4 938 528 48.0 214
Continued on next page
140 Fiberglass reinforcement

Table 7.6 Continued

Yarn designation Yield Minimum breaking strength

Glass system Tex system yd/lb Tex lb N
ECDE 751/0 EC6 661 x 0 7500 66 5.7 25.4
ECG 751/0 EC9 661 xO 7500 66 5.7 25.4
ECG 751/2 EC9661x2 3750 132 11.4 50.7
ECG 751/3 EC9 661 X3 2500 198 17.1 76.1
ECG 752/2 EC9 66 2x2 1875 264 22.8 101
ECG 75 2/3 EC9 66 2 x3 1250 396 34.2 152
ECG 752/4 EC9 66 2x4 938 528 45.6 203
ECH551/0 EClO 901 X 0 5500 90 9.5 42.3
ECDE 371/0 EC61341 X 0 3700 134 11.2 49.8
ECG 371/0 EC91341 X 0 3700 134 11.2 49.8
ECG 371/2 EC91341 X 2 1850 268 22.8 101
ECG 371/3 EC91341 X 3 1230 403 34.2 152
ECH 251/0 EClO 1981 X 0 2500 198 17.0 75.6
ECK 181/0 EC13 2751 X 0 1800 275 23.0 102
S-glass Yarns
SCD4501/2 SC5111x1 22500 22
SCG 150 1/2 SC9331x2 7500 66 8.2 36.5
SCG 150 2/2 SC933 2 x2 3750 132 16.4 72.9
Textured Yarns
ETDE 150 1/0 ET6331 X 0 14200 35 1.24 5.52
ETDE 751/0 ET6 661 X 0 7100 70 2.2 9.79
ETDE371/0 ET61341 X 0 3450 144 4.7 29.0

7.5.7 FIBERGLASS FABRIC There is a variety of weave patterns

(Fig. 7.4) that can be used to combine warp and
The effect that fiberglass fabrics have on the
filling yams to form a fabric. This weave pat-
properties of composite material is largely
tern controls the handling characteristics of the
dependent on the fabric's construction, which
fabric and to a large extent the properties of the
involves fabric count, warp and filling yam
composite that uses it as a reinforcement.
construction, weight and weave pattern.
The fabric's count is made up of the number (a) Plain weave: This fabric is constructed so
of warp yams ('ends') per inch (em) in the that one warp yam passes over and under
width and the number of filling yams ('picks') one filling yam (and vice versa). While the
per inch (em) in the lengthwise direction. resulting fabric exhibits the greatest degree
The warp yam is the yam that lies in the of stability in respect to yam slippage and
lengthwise (machine) direction of the fabric. fabric distortion, this stability is also a func-
The filling yam is the yam lying in the cross- tion of the fabric count and yam content.
wise direction of the fabric. The physical (b) Basket weave: A basket weave fabric has
parameters of the fabric, such as weight, thick- two or more warp yams interlocking over
ness and tensile strength are directly and under two or more filling yams. While
proportional to the types and numbers of yams a basket weave is less stable than a plain
used to weave it (see Table 7.7) weave, it is more pliable and will conform
or drape more readily to simple contours.
 Fiberglass types 141

Table 7.7 Glass-fabric construction and properties: physical properties for loomstate fabrics (without finish)

Style COUllt, Warp yam, Filling yarn, Weave Weight, Thickness, Breakillg
/em Tex system Tex system g/m2 111m strengtl!,
lin US system US system ozjl yeP ill • N/5 CIII
Ib/in
101 29.4 x 29.4 52.751 xO 52.751 x 0 Plain 17.0 0.20 350 x 333
75x75 D1800 1/0 01800 1/0 0.50 0.0008 4Ox38
104 23.6 x 20.5 55.51 x 0 52.751 x 0 Plain 19.7 0.030 350 x 130
60 x52 0900 1/0 D 1800 1/0 0.58 0.0012 40 x 15
106 22.0 x 22.0 D 5.51 x 0 55.51 x 0 Plain 24.7 0.038 394 x 350
56 x56 09001 /0 0900 1/0 0.73 0.0015 45x40
07 23.6 x 13.8 05.51 x 2 55.51 x 0 Plain 35.6 0.043 613 x 175
6Ox35 D 900 1/2 D 900 1/0 1.05 0.0017 70x20
'108 23.6 x 18.5 D 5.51 x2 55.51 x2 Plain 48.8 0.051 613 x 350
6Ox47 09001/2 0900 1/2 1.44 0.0020 70x4O
12 15.7 x 15.4 5111x2 5111x2 Plain 71.9 0.076 718 x 700
4Ox39 04501/2 04501/2 2.12 0.003 82x80
113 23.6 x 25.2 5111x2 55.51 x2 Plain 83.7 0.076 1077 x 525
6Ox64 04501/2 09001/2 2.47 0.003 123 x 60
16 23.6 x 22.8 Dll1x2 5111x2 Plain 107 0.102 1095 x 1050
6Ox57 04501/2 04501/2 3.16 0.004 125 x 120
120 23.6 x 22.8 5111x2 5111x2 Crowfoot 107 0.102 1095 x 1050
60 x 58 04501/2 D4501/2 3.16 0.004 125 x 120
il28 16.6 x 12.6 5221x3 5221x3 Plain 203 0.178 2185 x 1750
42 x 32 E 2251/3 E 2251/3 6.00 0.007 250 x 200
19.3 x 11.8 5223x2 5111x2 Crowfoot 298 0.229 5250 x 525
49 x 30 E 225 3/2 0450 1/2 8.34 0.009 600 x 60
62 11.0 x 3.2 5222x5 5222x5 Plain 414 0.381 3940 x 3065
28 x 16 02252/5 D 225 2/5 12.2 0.015 450 x 350
7.9 x 7.1 5224x3 5224x3 Plain 427 0.406 4375 x 3940
20 x 18 E 225 4/3 02254/3 12.6 0.016 SOOx 450
22.4 x 21.3 5221x3 5221x3 Satin 302 0.229 3065 x 2975
57x54 E2251/3 02251/3 8.90 0.009 350 x 340
82 23.6 x 21.6 7222x2 7222x2 8H Satin 423.8 0.33 3853 x 3502
6Ox56 E 225 2/2 E 225 2/2 12.5 0.0130 440 x 400
84 16.5 7224x3 7224x3 8 H Satin 881.5 0.66 8318 x 7005
42x36 E225 3 E 225 3 26.0 0.0260 950 x 800
220 23.6 x 22.8 7221xO 7221xO Crowfoot 109.2 0.0889 1096 x 1051
60 x 58 E 2251/0 E 2251/0 3.22 0.0035 125 x 120
341 1 1.8 x 19.3 5111x2 5223x2 Crowfoot 298 0.229 438 x 5250
3Ox49 D 4501/2 D 225 3/2 8.78 0.009 50 x 600
1070 23.6 x 13.8 5111xO 55.51 x 0 Plain 35.6 0.043 613 x 175
6Ox35 04501 /0 D 900 1/0 1.05 0.0017 70 x 20
1080 23.6 x 18.5 5111xO 5111xO Plain 48.8 0.051 613 x 350
6Ox47 0450 ItO D 4501/0 1.44 0.002 70 x40
081 27.6 x 23.6 5111xO 5111xO Plain 58.5 0.060 700 x 613
70x6O 0450 1/0 D 4501/0 1.72 0.0024 80 x 70
125 15.7 x 15.4 5111x2 5331xO Plain 88.1 0.089 788 x 1135
9 0450 112 G 150 0 2.60 0.0035 90 x 130
1165 23.6 x 20.5 5111x2 9331 xO Plain 124 0.109 1310 x 1180
60 x52 04501/2 G 150 1/0 3.66 0.0043 150 x 135
Continued on next page
142 Fiberglass reinforcement

Table 7.7 Continued

Style COUllt, Warp yam, Filling yarn, Weave Weight, TIlickness, Brenkillg
/em rex system rex system g/m2 tl/m strength,
lin US system US system ozlyd2 itl • Nl5cm
lb/ill
500 19.3 x 16.5 7511 x 0 7511 x 0 Plain 167.8 0.132 22~
2215
49x42 E 110 I/O E 110 I / O 4.95 0.0052 261 x 253
SOl 1 .2 x 12.7 751 I/O 7511 x 0 Plain 166.1 0.127 2145 x 2102
46x45 E 110 I/O E 110 I/O 4.90 O.OOSO 245 x 240
504 23.6 x 19.7 6331xO 6331xO Plain 147. 0.127 2688 x 2233
DEI DE 1 0 4.36 O.OOSO 307 x 25
1522 9.4 x 8.7 9331x2 9331x2 Plain 125 0.140 1400 x 1180
24x22 G 150 1/2 CG 150 1/2 3.70 0.0055 160 x 135
1523 11.0 x 7.9 9333x2 9333x2 Plain 403 0.356 4595 x 3500
28x20 G 150 3/2 G 150 3/2 11.9 0.014 525 x 400
526 13.4 x 12.6 9331><2 9331x2 Plain 185 0.165 1970 x 1705
34 x 32 G 150 1/2 Gl 1 2 5.45 0.0005 225 x 195
1527 6.7 X 6.7 9333x3 9333x3 Plain 437 0.381 4375 x 4245
17x 17 G 1503/3 G 1503/3 12.9 0.015 500 x 485
1528 17.3 x 12.6 9331x2 9331x2 Plain 203 0.178 2185 x 17SO
44x 32 G ISO 1/2 G 1501/2 6.00 0.007 250 x 200
1543 19.3 x 11.8 9662x2 7221x2 Crowfoot 298 0.229 5250 x 525
49x3O G 752/2 D 2251/2 8.34 0.009 600 x 60
1557 22.4 x 11.8 9331x2 7221xO Crowfoot 184 0.140 3240 x 525
57x3O G 150 1/2 E 2251/0 5.42 0.0055 370 x 60
1564 9241 x 2 9241 x 2 9334x2 Plain 431 0.406 4375 x 3940
G 371/2 G 371/2 G 1504/2 12.7 0.016 500 x 450
1581 22.4 x 21.3 9331x2 9331x2 Satin 302 0.216 2975 x 288S
57x54 G 150 1/2 G 150 1/2 8.92 0.00 5 340 x 330
1582 23.6 x 22.0 9331x3 9331x3 Satin 454 0.356 4595 x 4375
6Ox56 G 150 1/3 G 150 1/3 13.4 0.014 525 x 500
1583 21.3 x 18.9 9332x2 933 2 x2 Satin 570 0.457 5690 x 5165
54x48 G 1502/2 G 1502/2 16.8 0.018 650 x 590
1584 17.3 x 13.8 9684x2 9684x2 Satin 861 0.686 8315 x 7005
44x36 G 75 2/2 G 75 2/2 25.4 0.035 950 x 800
1588 16.5 x 14.2 9334x4 9334x4 Satin 1756 1.27 16 635 x 11 820
42x36 G 150 4/4 G 150 4/4 51.8 0.050 1900 x 1350
1614 11.8 x 5.51 9331 x 0 9681 xO Leno 79.0 0.127 657 x 744
30 x 14 G 150 I/O G 751/0 2.33 0.005 75x8S
1652 20.5 x 20.5 9331xO 9331xO Plain 141.7 0.114 2154 x 1926
52 x52 G 150 I/O G 150 I/O 4.18 0.00454 246 x 220
1665 15.7 x 9.5 9331xO 6681xO Plain 141.7 0.114 2154 x 1926
4Ox24 G 150 I/O DE 751/0 3.50 0.0049 180 x 180
1674 15.7 x 12.6 9331xO 9331xO Plain 95.6 0.107 1225 x 832
4Ox32 G 150 I/O G 150 I/O 2.82 0.0042 140 x95
1675 15.7 x 12.6 6331xO 6331xO Plain 95.6 0.107 1225 x 832
40x32 DE 150 I/O DE 1501/0 2.82 0.0042 140 x 95
676 22.0 x 18.9 6331xO 6331xO 139 0.122 1705 x 1310
56x 48 DE 150 I/O DE 150 I/O Plain 4.10 0.0048 195 x ISO
1677 15.7 x 15.7 6331xO 6331xO Plain 108 0.114 1225 x 1135
4Ox4O DE 150 I/O DE 150 I/O 3.20 0.0045 140 x 130
Continued on next page
 Fiberglass types 143

Table 7.7 Continued

Style COllnt, Warp yam, Filling yam, Weave Weight, Thielen Breaking
/em Tex system rex sy lem g/m2 mm treng'/!,
/ill US system US system oz/yd2 in • N!5em
Ib/in
167 15.7 x 15.7 9331xO 9331xO Plain 108.5 0.109 1839 x 1751
4Ox40 G 150 I/O G 150 I/O 3.20 0.0043 210 x 200
680 72 x 70 DE 150 1/0 DE 150 1/0 Satm 'l.70 U.U06 275 x 225
28.3 x 27.6 6331xO 6331 0 193 0.152 2405 x 1970
681 22.0 x 1 .2 6331xO 6331xO Plain 122 0.127 1705 x 963
56x36 DE 150 I/O DE 150 I/O 3.60 0.005 195 110
698 18.9 x 7.1 6331xO 61341 x 0 Fancy 169.9 0 3 1313 x 700
48 x 18 DE 1501/0 DE 371 0 5.01 0.0135 150 x 80
1800 6.3 x 5.5 132751 x 0 13 2751 x 0 Plain 327 0.356 3940 x 3590
16 x 14 K 181/0 R 181/0 9.66 0.014 450 x410
112 15.7 x 15.4 722)xO 7221 0 Plain 71.2 0 7 x 700
4Ox39 E 2251/0 E 2251/0 2.10 0.0034 9OxSO
113 23.6 x 22.0 7221xO 5111 x 0 Plain SO.7 O. 1 1225 x 525
60x 56 E 2251/0 E 450 110 2.38 0.0032 140 60
2116 23.6 x 22.8 7221xO 7221xO Plain 107 0.102 1095 x 1050
6Ox58 E 2251/0 E 2251/0 3.16 0.004 125 x 120
26 21.6 7221 0 7221 II Plain 20.4 O. 11 2 1009
66xSS E 2251/0 E 2251/0 3.1 0.0037 135 115
23.6 x 22. 5221xO 5221xO Crowfoot 107 0102 1095 x 1050
6Ox58 D2251/0 02251/0 3.16 0.004 125 x 120
15.7 x 15.4 5221xO 9331 xO Plain 0.094 788 x 1135
4Ox39 02251/0 G 150 I/O 2.62 0.0037 90x 130
23.6 x 20.5 5221xO 9331xO Plain 125 0.114 1095 x 1220
60 52 D 225 1/0 G ISO 1/0 3.70 0.0040; 125 140
2313 23.6 x 25.2 7221xO 5111xO Plain SO.7 0.813 1226 x 525
6Ox64 E 2251/0 0450 1/0 2.38 0.0032 140 x 60
2523 11.0 x 7.9 1019 1 x 0 101981xO Plain 403 0.330 5075 x 3370
28 x20 H 251/0 H 251/0 11.9 0.013 580 x 385
2532 6.3 x 5.5 10 1981 x 0 1019 1 x 0 Plain 246 0.254 2625 x 2450
16 x 14 H 251/0 H 251/0 7.25 0.010 300 x2SO
0 276 x 76 6 16.5 1 x 0 6 16.5 I x 0 Plain 92.1 0.07 1226 x 1226
70 70 OE3001/0 DE 300 1/0 2.74 0.0031 140x 140
23.6 x 24.4 616.51 x 0 616.51 x 0 Plain 81.4 0.0076 1050 x 1094
6Ox62 DE 300 l/O DE 300 I/O 2.40 0.0030 120 125
18.9 x 12.6 91341 xO 91341 xO Crowfoot 431 0.340 SOOOx3700
48 x 32 G 371/0 G 371/0 12.7 0.0134 571 x 423
3733 7.1 x 7.1 91341 x 0 91341 xO Plain 197 0.203 2185 x 1750
18 x 18 G 371/0 G 371/0 5.SO 0.008 250 x 200
3743 19.3 x 11.8 91341 x 0 5221xO Cr wfoot 286 0.203 5250 x 525
49x3O G 371/0 D 2251/0 8.45 0.008 600 x 60
3783 21.2 x 18.9 91341 x 0 91341 x 0 8 Ham 556 0.406 4816 x 4290
54x48 G 371/0 G 371/0 16.4 0.0160 550 x 490
4522 9.4 x 8.7 9331x2 9331x2 Plain 123.4 0.130 1226 x 1130
(6522) 24x22 G 150 1/2 G 150 1/2 3.64 0.0051 140 x 129
4533 7.1 x 7.1 9681x2 9681x2 Plain 200.7 0.188 2382 x 2601
(6533) 18 x 18 G 751/2 G 751/2 5.92 0.0071 272 x 297
Continued on next page
144 Fiberglass reinforcement

Table 7.7 Continued

Style COIlllt, Warp yam, Fillin yam, Weave Weight, Thickness, Breakillg
/em Tex system Tex y tem g/ml 11/111 strength,
/ill US ystem US system oz/ydZ ill • !Scm
lb/in
4700 5.5 x5.1 9681xO 9681xO Plain 147.5 0.196 2189 x 1445
14 x 13 G 371/0 G 371/0 4.30 0.0077 250 x 165
6060 23.6 x 23.6 68.271 x 0 68.271 xO Plain 39.0 0.0019 656.7 x 718
6Ox60 DE 600 I/O DE 600 1/0 1.15 0.048 75 x 82
6581 22.4 x 21.3 9331 x 2 9331x2 8HS 301.8 0.228 3520 x 2846
57x54 G 150 1/2 G 150 1/2 8.90 0 ..90 400 x 325
67 1 22.4 x 21.3 9681xO 9681xO 8HS 301.8 0.28 3520 x 2846
57x54 S2C6751/0 S2CG751/0 8.90 0.0090 400 x 325
7500 6.3 x 5.5 9682x2 9682x2 Plain 327 0.356 3940 x 3590
16 x 14 G 752/2 G 75 2/2 9.66 0.014 450 x410
7520 7.1 x 7.1 9681x3 9681x3 Plain 294.3 0.304 2890 x 2890
18 x 18 G 751/3 G 751/3 8.68 0.0120 338 x 330
7532 6.3 x 5.5 9681x3 9681x3 Plain 254 0.254 2930 x 2765
16 x 14 G 751/3 G 751/3 7.50 0.010 335 x 316
7533 7.1 x 7.1 9681x2 9681x2 Plain 193 0.203 2010 x 1925
1 x 18 G 751/2 G 751/2 5.70 0.008 230 x 220
7544 11.0 x 5.5 9682x2 9682x2 2 x I BSK. 610 0.559 6520 x 7265
28x 14 G 75 2/2 G 75 2/2 18.00 0.022 745 x 30
7587 16.5 x 8.3 91341 x 2 91341 x2 Mock 705 0.711 8450 x 4600
lena
42 x21 G 371/2 G 371/2 20.8 0.028 965 x 525
7626 13.4 x 12.6 9681xO 9681xO Plain 183 0.168 1970 x 1750
34x32 G 751/0 G 751/0 5.40 0.0066 225 x 200
762 17.3 x 12.6 9681 x 0 9681 x 0 Plain 203 0.178 2190 x 1750
44 x 32 G 751/0 G 751/0 6.00 0.007 250 x 200
7629 17.3 x 13.3 9681xO 9681xO Plain 211.9 0.178 2190 x 1883
44x34 G 751/0 G 751/0 6.25 0.007 250 x 215
7637 17.3 x 8.7 9681xO 91341 xO Plain 240 0.229 2190 x 2190
44x22 G 751/0 G 371/0 7.rE 0.009 250 x 250
7642 17.3 x7.8 9681xO 9 134 1 x 0 (Tex) Plain 232.2 0.279 2189 x 1051
44x20 G 751/0 G 371/0 (Tex) 6.85 0.011 250 x 120
7743 47.2 x 7.9 6661xO 9331xO Satin 346 0.279 7000 x 525
120 x 20 DE 751/0 G 150 1/0 10.22 0.011 800 x 60
7652 12.6 x 12.6 9491 x 0 9491 x 0 Plain 254.3 0.211 2277 x 2189
32 x 32 G 50 I/O G 50 I/O 7.50 0.0083 260 x 260
7678 17.3 x10.6 968 668 Plain 192.2 0.160 3108 x 2189
45x27 G 751/0 DE 751/0 5.67 0.0063 355 x 250
77 1 23.6 x 21.3 6661xO 6661xO Satin 302 0.216 2980 x 2890
59 x 54 DE 751/0 DE 751/0 9.15 0.010 340 x 330
76281 17.3 x 12.6 9681xO 9681xO Crowfoot 203 0.173 2190 x 1750
44 x 32 G 751/0 G 751/0 6.00 0.0068 250 x 200

Shaded items may be obsolete or not in all vendor current catalogues

 Fiberglass types 145

Notes to Table 7.7

1. All fabrics listed in Table 7.7 which contain 0 225 5 22 yam can also be woven with the original E
2257 22 yam equivalents.
2. Breaking strength of fabrics in SI units: N was determined by testing a specimen of 5 cm width and is
for loomstate fabrics. Finishing reduces fabric strength.
3. This thickness is an estimate for dry ply thickness and is an average of several tests. The manufacturer
usually reports a range of thicknesses. To estimate the cured ply thickness, divide the dry estimate by
the expected fiber volume, i.e. if the dry thickness is 0.1 rom and the fiber volume is 0.5 then the
estimated cured ply thickness is 0.1/0.5 or 0.20 rom. See introduction for methods of converting
between weight and volume percent.
4. Styles 4522 (6522) and 4533 (6533) are S-2 glass; Styles 6581 and 6781 are S-glass.
a Data from Reference 3

Fig.7.4 (a) Plain weave; (b) basket weave; (c) twill weave; (d) crowfoot satin (four harness weave; (e) eight-
harness satin weave.

(c) Twill weave: This fabric consists of one or over seven and under one filling yam in an
more warp yams over and under two or irregular pattern. The resulting fabric is
more filling yarns in a regular pattern. This very pliable and readily conforms to com-
produces either a straight or a broken diag- pound contours. Because this weave
onal line in the fabric, which produces pattern allows comparatively high fabric
greater drapeability and stability. counts, it contributes maximum strength to
(d) Crowfoot satin weave: This weave pattern composites reinforced by it.
has one warp yam interlocking over three (f) Unidirectional fabrics: Fabrics produced
and under one filling yam in an irregular with heavy warp yams and light filling
pattern. The fabric that results is extremely yams, in either crowfoot or long shaft satin
pliable and lends itself to conforming to weaves. The filling yarns can also be com-
complex contours. posed of yarns other than glass. These
(e) Eight-harness satin weave: A fabric con- fabrics offer high strength reinforcements
structed with one warp yam interlocking in the heavy yam direction.
146 Fiberglass reinforcement

(g) Non-woven: Unidirectional fabrics that can ments together, lubricate the roving for contact
be produced by sticking the 'warp' and 'fill- with various processing equipment, and to
ing' yarns together chemically. Although permit the glass filaments to be thoroughly
this chemical bonding contributes to the wetted when combined with other materials.
fabric's stability, these fabrics have a firm Fiberglass roving sizes consist of polyvinyl
hand and do not drape over complex con- acetate, polyvinyl alcohol, or PVA/ starch as a
tours. film former with the addition of such chemi-
cals as chrome complexes, organosilane
antistatic agents and lubricants that impart the
7.5.8 OTHER WOVEN FORMS
desired strand characteristics. The film former
Fiberglass yams can also be woven into tapes, aids in adhering the filaments together and
contoured fabrics, fluted core fabrics and three giving strand integrity that will reduce fila-
dimensional fabrics. ment abrasion during fiber drawing, strand
Tapes are usually narrow fabrics [less than conversion and the end use of the rovings by
30 em (12 in) wide], with a secured selvedge to the fabricator.
prevent unravelling. Contoured fabrics are An organosilane or coupling agent is
woven into a specific geometrical shape. added, that can react with hydroxyl groups on
Fluted core fabrics are two parallel layers of the glass fiber surface and also possesses one
fabric tied together by stringers of woven fab- or more reactive groups to react with other
rics so that the cross-section is triangular or materials, specifically those present in thermo-
rectangular. Three dimensional fabrics are plastic and thermosetting resins. The silane
really planar or fabrics woven with yams in produces a form of chemical bridge between
three distinct directions within the fabric plane. the glass surface and the resin matrix.
That is, yarns are interwoven in: (i) the machine The inclusion of antistatic agents and lubri-
direction, (ii) +45 0 from the machine direction, cants improves the softness and choppability
or (iii) -45 0 from the machine direction. of the roving. The selection of a roving size is
based on its resin compatibility, processing
and the expected end use performance.
7.5.9 MILLED FIBERS
The largest percentage of rovings are used in
Continuous fiberglass strands can be hammer processes that require relatively short fiber
milled into very short fiber lengths [approx. lengths (1-5 cm, 0.5-2.0 in). This type is used in
2-6 mm (X6-X in) long]. The actual fiber processes such as spray-up, preform molding,
lengths are determined by the diameter of the bulk molding compounds and continuous
screen openings through which the fibers pass lamination.
during milling. These milled fibers are usually Rovings suitable for chopping are available
used as inert fillers for thermoplastic and ther- in varying degrees of softness; the softer rov-
moset resin systems. ings are generally more difficult to chop, but
they are recommended for use in applications
that require intricate shapes, sharp comers
7.6 SURFACE TREATMENTS
and difficult radii. Conversely, the harder
types of rovings are generally recommended
7.6.1 ROVINGS
for use where the choppability is of prime
In order for fiberglass rovings to be compatible importance and they are used in simple parts
with processing methods and materials, a with minimum contours and radii.
chemical size is applied during the basic fiber Single strand rovings that are commonly
forming operation. The formulation of this size used in filament winding and pultrusion
is designed to hold the individual glass fila- should have good strand integrity, controlled
 Design considerations 147

level of broken filaments, good wettability and nated Y) and hydrolyzable groups (designated
uniform processability under controlled ten- X) in materials with a generic structure X3SiRY.
sion. The hydrolyzable groups are intermediaries in
S-glass based rovings have been used in the formation of silanol groups that bond to
applications where the composite needs the glass surface, whereas the organofunc-
increased physical properties (e.g. filament tional groups are designed for reactivity or
wound pressure vessels). compatibility with the polymer to be used by
the composite manufacturer. (Table 7.8)
The chemical functionability of the cou-
7.6.2 FIBERGLASS TEXTILE YARNS
pling agent can determine the resistance to
The continuous fiberglass strands that will be varied environments, chemical, physical and
used in weaving are treated at the bushing thermal (Table 7.9). There are many functions
with a starch-oil binder; the general formula that may be attributed to coupling agents at
for such a binder can be a partially or fully the glass-resin interface. They may provide
dextrinized starch or amylose hydrogenated lubrication to protect against abrasion during
vegetable oil, a cationic wetting agent, emulsi- fabrication. The coupling agent mechanism
fying agent and water. These sizes or binders can protect against stress corrosion as a result
are intended to protect the fibers from damage of water incursion and can help improve the
during their formation and subsequent opera- mechanical and electrical properties of rein-
tions of twisting, plying and weaving. forced composites.

7.6.3 FIBERGLASS FABRICS 7.7 DESIGN CONSIDERATIONS

Fabrics directly from the 100m still have the It is only recently that the engineering and
original binders that were applied at the yam design criteria in the combination of such dif-
bushing; the warp yarns have also been treated ferent materials as thermosetting and
with a polyvinyl alcohol solution to help pro- thermoplastic polymers with reinforcing of
tect them during weaving. These binders and glass were properly understood. There are a
sizes are not compatible with the resins used number of basic considerations that are com-
by the composite manufacturers and must be mon to the use of fiberglass as a reinforcing
removed prior to impregnation with other media in the production of composites. The
polymers. This is usually accomplished by type form, weave pattern, weight, permeabil-
treating the fabric to carefully controlled ity and alignment of the fibers have direct
time-temperature cycles, which results in com- bearing on the mechanical properties of the
plete removal of all organic material. However, composite. Also in some instances the chemi-
after this heat treatment no suitable interface cal, thermal and electrical properties may also
exists between the fiberglass surface and the be influenced by the judicious choice of rein-
resin matrices that will eventually be used. An forcing material (Table 7.10).
improved interface with resulting good adhe-
sion even in adverse environment conditions
7.7.1 GLASS COMPOSmON
can be achieved by the application of a cou-
pling agent to the glass fiber surface. The chemical composition of the glass can
The most commonly used coupling agents have a direct bearing on the fiber properties
in current use are those termed silanes. The which can indirectly affect composite perfor-
coupling agent mechanism of these mance.
organofunctional silanes depends on a link It has been noted that 'E' glass composition
between the organofunctional group (desig- fibers can contribute 210 000-225 000 psi to the
148 Fiberglass reinforcement

Table 7.8 Commercial coupling agents

Organofunctional group Chemical structure

Vmyl C~", CHSi(OCHa)3
Chloropropyl ClC~C~C~Si(OCHa)3

/0\
Epoxy CH2CHCH2OCH2CH2CH2Si(OCH3h
CH2
I
Methacrylate (CH2=C-COOCH2CH2CH2Si(OCH3h
Primary amine ~NC~C~CH2Si(OC2H5)3
Diamine ~NCH2C~NHC~C~CH2Si(OCHJ3
Mercapto HSCH2C~C~Si(OCH3)3
Cationic styryl C~=CHC6H4C~NHCH2CH2NH(C~)3Si(OCH3)3'HCl
CH3 Cl
I I +
Cationic methacrylate CH2",C-COOCH2CH2 - N(Me2)CH2CH2CH2Si(OCH3h

CH3
I
CH2=C
I
Chrome complex
/C~
R10H 0 ~O R10H
\/ \/
Cl-Cr Cr-Cl

Itl~o/I~l
H 20 I H 20
H

CH2
I
Titanate (CH2",C-COOh TIOCH(CH3h
Cross-linker (CHP>3 SiC~CH2Si (OCHa)3
Mixed siIanes C6H 5Si(OCH3)3 + F
Formulated Melamine resin + C

tensile strength of a composite which they 7.7.2 FILAMENT DIAMETER

reinforce. A number of different 'E' glass filament diam-
Other glass compositions, namely'S' and eters have been evaluated comparing the
'52' glass, have produced even greater strength influence of filament diameter on resulting
improvements in reinforced composites. laminate properties. The fabrics evaluated
were woven with the same count and weave
pattern as Style 1581, using both plied and sin-
gles yam containing different filament
diameters (Tables 7.11 and 7.12).
 Design considerations 149

Table 7.9 Effect of coupling agents on the mechanical strength properties of laminates

Finish Loomstate 112 Volan Al72 A174, Z6030

Polyester laminates reinforced with style 1581 glass fabric

Flexural strength,
psi X 103 (MPa)
Cond.A 40.4 (278.6) 32.3 (222.7) 73.2 (504.7) 68.8 (474.4) 75.5 (520.6)
Condo D2/100 25.3 (174.4) 31.8 (219.3) 59.0 (406.8) 66.1 (455.8) 77.3 (533.0)
Compressive strength,
psi x 103 (MPa)
Cond.A 23.8 (164.1) 40.4 (278.6) 48.0 (331.0) 47.1 (324.8) 66.2 (456.4)
Condo D2/100 12.5 (86.2) 12.1 (83.4) 38.9 (268.2) 44.9 (309.6) 58.8 (405.4)
Tensile strength,
psi x 103 (MPa)
Cond.A 48.1 (331.6) 20.6 (142.0) 49.9 (344.1) 50.0 (344.8) 52.2 (359.9)
Condo D2/l00 48.8 (336.5) 51.2 (353.7) 50.7 (349.6)
Epoxy laminates reinforced with style 1581 glass fabric
Flexural strength,
psi X 103 (MPa)
Cond.A 78.2 (539.2) 76.0 (524.0) 80.9 (557.8) 71.3 (491.6) 85.7 (590.9)
Condo D2/100 76.7 (528.8) 72.4 (499.2) 83.1 (573.0) 68.9 (475.1) 82.4 (568.1)
Compressive strength,
psi x 103 (MPa)
Cond.A 64.3 (443.3) 63.4 (437.1) 70.7 (487.5) 63.0 (434.4) 67.5 (465.4)
Condo D2/l00 58.5 (403.3) 54.1 (373.0) 64.7 (446.1) 62.9 (433.7) 59.5 (410.3)
Tensile strength,
psi X 103 (MPa)
Cond.A 56.9 (392.3) 48.8 (336.5) 57.8 (398.5) 50.9 (351.0) 58.2 (401.3)
Condo D2/l00 55.8 (384.7) 47.2 (325.4) 56.1 (386.8) 49.4 (340.6) 52.2 (359.9)

Filament diameter does not appear to play The effect of variation in the available sur-
an important role in the determination of the face area for stress transfer is affected by the
mechanical properties of such plied yarn fab- coupling agent-resin interaction. Larger sur-
rics as Style 1581 woven with ECG 150 1/2 face areas with poor coupling agents reduces
(EC 9 33 1 x 2 yarn and style) and Style 1281 specific shear loading on the interface and so
woven with ECB 150 1/2 (EC 3.8 33 1 x 2 produces high laminate flexural strengths. But
yarn). However, in comparison with commer- with improved interfacial bonding the surface
cially available singles yarn Style 7781 woven areas effect disappears because other failure
with ECDE 75 1/0 (EC 6.66 1 x 0) yarn, sin- modes appear first. Singles yarns produce
gles yarn Style B 7781 woven with finer higher composite strengths due to the smaller
filaments [ECB 75 1/0 (EC3 8.66 1 x 0 yarn)] angle between fiber axis and load axis in the
gives laminates with higher flexural lower twist yarns. This produces less tensile
strengths, flexural modulus and tensile loading across the interface; the magnitude of
strengths. Conversely a singles yarn G 7681 this effect will vary with the resin coupling
woven with coarser filaments [ECG 75 1/0 agent efficiency and the function of the twist of
(EC9 66 1 x 0 yarn)] gives composites with the yam or the angle of the spiral of the fila-
improved compressive strengths. ments.
150 Fiberglass reinforcement

Table 7.10 The effect of fabric construction on laminate mechanical properties2

Fabric construction
Fabric no. E1581 51581 YM31A"
Count/in (em) 57 X 54 (22.4 x 21.3) 57 x 54 (22.4 x 21.3) 57 x 54 (22.4 x 21.3)
Warp yam
Glass system ECG 1501/2 SCG 150 1/2 MCG 150 1/2
Texsystem EC9331 x2 SC9 331 x2 MC9 331 x2
Filling yam
Glass system ECG 150 1/2 SCG 150 1/2 MCG 150 1/2
Tex system EC9331x2 SC9331x2 MC9331 x2
Weave Long-shaft satin Long-shaft satin Long-shaft satin

Laminate mechanical properties

Fabric no. E1581 51581 YM31A"
Flexural strength,
psi x 103 (MPa)
Cond.A 88.6 (610.9) 87.1 (600.6) 89.8 (619.2)
Flexural modulus of elasticity,
psi x 106 (GPa)
Cond.A 3.78 (26.1) 3.79 (26.1) 5.00 (34.5)
Compressive strength,
psi x 103 (MPa)
Cond.A 64.8 (446.8) 56.4 (388.9) 61.0 (420.6)
Tensile strength,
psi x 103 (MPa)
Cond.A 60.0 (413.7) 61.4 (423.4) 64.5 (444.7)

• Data from Reference 4

7.7.3 WEAVE PATTERN or more yams lying 90 degrees to it), the

The amount of yam and the pattern of the higher the reinforced composite strength. So
weave often determine the handling character- by extending the length of the float which
istics of a fabric. If a fabric is too tightly woven, reduces the interlocking frequency, the result-
it will not readily drape and conform to vari- ing composite's strength will be increased.
ous contours during molding; a tight weave Changing the weave, for example from a plain
can also adversely affect impregnation of the weave to a crowfoot weave, will also achieve
fabric by the resin. If a fabric has a weave pat- improved composites strength properties
tern that is very open, this can produce a (Table 7.13).
weaker reinforcement due to insufficient fiber,
7.7.4 GLASS-TO-RESIN RATIO
a tendency to distort and in horizontal treater,
difficulty in impregnation. In general, approximation of physical proper-
A general assumption can be made in the ties for fiber glass reinforced composites
longer the float (i.e. the portion of a warp or follows the 'rule of mixtures'. The calculation of
filling yam that will extend unbound over two loads stresses and strains parallel to the fibers
 Design considerations 151

Table 7.11 Composite mechanical properties as a function of glass-filament diameter

Plied-yam fabrics
Fabric style 181 1581 1281
Count/in (em) 57 x 54 (22.4 x 21.3) 57 x 54 (22.4 x 21.3) 57 x 54 (22.4 x 21.3)
Warp yam
Glass system ECE 2251/3 ECG 150 1/2 ECB 150 1/2
Tex system EC7221x3 EC9331x2 EC3.8331 x2
Filling yam
Glass system ECE 2251/3 ECG 150 1/2 ECB 150 1/2
Texsystem EC7221 x3 EC9331 x2 EC3.8331 x2
Weave Long-shaft satin Long-shaft satin Long-shaft satin
Flexural strength,
psi x U)3 (MPa)
Cond.A 93.0 (641.2) 88.6 (610.9) 92.4 (637.1)
Flexural modulus of elasticity,
psi x 1()6 (GPa)
Cond.A 3.88 (26.8) 3.78 (26.1) 4.06 (28.0)
Compressive strength,
psi x 103 (MPa)
Cond.A 58.1 (400.6) 64.8 (446.8) 64.9 (447.5)
Tensile strength,
psi x 103 (MPa)
Cond.A 60.0 (413.7) 60.0 (413.7) 64.8 (446.8)

Singles-yam fabrics
Fabric style G7681 DE7781 B7781
Count/in (cm) 60 x 54 (23.6 x 21.3) 60 x 54 (23.6 x 21.3) 60 x 54 (23.6 x 21.3)
Warp yam
Glass system ECG751/0 ECDE 751/0 ECB751/0
Texsystem EC9661xO EC6661xO EC3.8 661 xO
Filling yam
Glass system ECG 751/0 ECDE 751/0 ECB 751/0
Tex system EC9661xO EC6661 xO EC3.8 661 xO
Weave Long-shaft satin Long-shaft satin Long-shaft satin
Flexural strength,
psi x 103 (MPa)
Cond.A 92.6 (638.5) 94.9 (654.3) 101.1 (696.4)
Flexural modulus of elasticity,
psi x 106 (GPa)
Cond.A 3.78 (26.1) 3.90 (26.9) 4.02 (27.7)
Compressive strength,
psi x 103 (MPa)
Cond.A 76.7 (528.8) 67.7 (466.8) 69.9 (482.0)
Tensile strength,
psi x 103 (MPa)
Cond.A 64.2 (442.7) 66.3 (457.1) 69.4 (478.5)
152 Fiberglass reinforcement

Table 7.12 Mechanical properties of 5-g1ass fabrics-epoxy composites as a function of plied compared
with unplied yam 2

Fabric style 51581 57681 57681

Finish VoianA VolanA C5-31O
Flexural strength,
psi X 103 (MPa)
Cond.A 84.8 (584.7) 120.0 (827.4) 116.0 (799.8)
Flexural modulus of elasticity,
psi x 106 (GPa)
Cond.A 3.10 (21.4) 4.18 (28.8) 5.21 (35.9)
Compressive strength,
psi x 103 (MPa)
Cond.A 61.3 (422.7) 55.8 (384.7) 73.9 (509.5)
Tensile strength,
psi x 103 (MPa)
Cond.A 70.6 (486.8) 87.5 (603.3) 86.4 (595.7)

Table 7.13 Effect of weave pattern on composite mechanical properties

Fabric style 7628" 76281 b 16-149< 7781 d

Finish Volan VolanA VoianA VoianA
Resin Polyester Polyester Polyester Polyester
Plies 18 18 12 12
Resin content, wt. % 37.1 36.7 36.5 37.6
Thickness, in (mm) 0.124 (3.15) 0.121 (3.07) 0.120 (3.05) 0.120 (3.05)
Flexural strength,
psi x 103 (MPa)
Cond.A 53.8 (371.0) 84.7 (584.0) 63.2 (435.8) 87.0 (600.0)
Flexural modulus of elasticity,
psi X 106 (GPa)
Cond.A 3.89 (26.8) 3.41 (23.5) 3.81 (26.3) 3.24 (22.3)
Compressive strength,
psi X 103 (MPa)
Cond.A 25.7 (177.2) 57.0 (393.0) 48.0 (331.0) 64.3 (443.3)
Tensile strength,
psi x 103 (MPa)
Cond.A 45.9 (316.5) 59.2 (408.2) 58.7 (404.7) 60.0 (413.7)

a Plain weave, data from Reference 3 c A 5-shaft satin weave version of Style 7781
b Crowfoot satin weave, data from Reference 3 d An 8-shaft satin weave
 Design considerations 153

and perpendicular to the fibers is quite differ- pattern, weave density and yam diameters. The
ent. Also, the stress transfer across the material optimum strengths of laminates are obtained
boundaries is greatly affected by the degree of with the lowest practical resin content.
adhesion at the reinforcement-resin interface. With coarse plain weave fabrics using hand
A more subtle effect is the possibility that lay-up, laminates can be produced with resin
wall effects at the surface may alter the matrix contents of 36-38%; with other weaves using
orientation and mechanical properties. less twist in the yam and optimum yam spac-
It must also be assumed that there will be ing, excellent physical properties can be
sufficient resin present to fill all of the spaces achieved in the 25% resin content range. Plain
between the fibers. Internal voids have been weave fabrics with higher twist yam will
shown to produce considerable stress concen- require higher resin contents than low twist
trations resulting in premature mechanical singles fabrics of equivalent weave.
failure of the laminate. Woven fiberglass fabrics In filament winding applications parallel
by virtue of the weave interlocking reduce the strands of fiberglass are wound around a cylin-
effective glass-ta-resin ratio below those of fila- drical mandrel, the resulting cross-section of
ment wound structures. The optimum glass to round filaments provides a close packing,
resin ratio then becomes dependent on weave resulting in low resin contents of the final part.

Table 7.14 Laminate mechanical properties as a function of glass-to-resin ratio (Style 181 glass
fabric-epoxy resin composites)

Plies 7 9 11 13 14 16 18 20
Resin content, wt. % 55.0 49.2 44.0 35.6 31.2 28.0 22.5 22.2
Thickness, in 0.117 0.126 0.127 0.126 0.125 0.131 0.132 0.147
mm 2.97 3.20 3.23 3.20 3.18 3.33 3.35 3.73
Flexural strength,
Cond.A
psi x103 45.6 58.8 64.8 83.0 81.2 92.1 87.4 91.9
MPa 314.4 405.4 446.8 572.3 559.9 635.0 602.6 633.7
Flexural modulus
of elasticity,
Cond.A
psi x 1()6 2.26 2.62 2.92 3.33 3.82 4.04 4.37 4.64
GPa 15.6 18.1 20.1 23.0 26.3 27.9 30.1 32.0
Compressive strength,
Cond.A
psi X 103 45.2 47.9 44.7 53.2 56.9 52.4 54.3 67.8
MPa 311.7 330.3 308.2 366.8 392.3 361.3 374.4 467.5
Tensile strength,
Cond.A
psi X 103 30.8 35.7 40.5 51.3 53.7 61.0 64.7 65.1
MPa 212.4 246.2 279.2 353.7 370.3 420.6 446.1 448.9
Tensile load/ply,
lb 514 494 468 498 480 500 476 480
N 2286 2197 2082 2215 2135 2224 2117 2135
154 Fiberglass reinforcement

For example, resin contents of 25-30% with right angles to the other half), the resulting
glass fiber content of 70-75% product final mechanical strengths at either angle is less than
composites with high physical strengths. But, if that of the parallel alignment. As the distribu-
there is any variation from the parallel align- tion of strands varies between 0° to 90°
ment of the glass fibers, this will reduce the alignment to a +45° to a -45°, this further
degree of packing and inevitably the optimum reduces mechanical strength in the primary
glass-ta-resin ratio. The weaving of glass fab- direction, but shows an increase in the +45° and
rics will of course reduce the effective -45° directions. Of course the yam distribution
glass-ta-resin ratios because of the weave inter- can be varied from 0° and 90° directions as part
locking. The correlation between glass-ta-resin of the fabric design. A so called 'balanced' fab-
ratios and composite strength is well known, ric, that is one with equal yam distribution in
the composite's strength increases as resin con- the warp and filling directions, will have com-
tents are reduced (see Table 7.14 and Fig. 7.4). parable composite properties in both directions.
When fiberglass yam is woven with yam
primarily in the warp direction with only the
7.7.5 FillER DISTRIBUTION
minimum amount of filling yam (enough to
Reinforced composites that have the glass give the fabric stability), the result is a unidi-
strands aligned parallel to each other have rectional fabric. Composites reinforced with
their maximum strength and stiffness in the this type of fabric have the maximum mechan-
direction of the alignment. This type of paral- ical strength in the direction of the greatest
lel alignment is conducive to certain filament concentration of yams.
winding and pultrusion operations. The differences between the mechanical
When the reinforcement is aligned at right properties, bidirectional and unidirectional
angles to itself (half of the strands are laid at reinforced composites, can be seen in Table 7.15.

Table 7.15 Effect of fabric yam distribution on laminate mechanical properties

Fabric type (style) Bidirectional (1581) Unidirectional (7743)

Test direction Warp Filling Warp Filling
Yam content, % 52 48 90 10
Flexural strength,
psi x HP (MFa)
Cond.A 84.2 (586.0) 78.2 (539.2) 95.0 (655.0) 23.3 (160.6)
Condo D2/100 75.5 (520.5) 68.5 (472.3) 87.4 (602.6) 24.1 (166.2)
Flexural modulus of elasticity,
psi X 106 (GPa)
Cond.A 4.02 (27.7) 3.65 (25.2) 4.95 (34.1) 2.47 (17.0)
Condo D2/100 3.87 (26.7) 3.41 (23.5) 4.80 (33.1) 1.46 (10.1)
Compressive strength,
psi x 103 (MFa)
Cond.A 62.7 (432.3) 63.4 (437.1) 64.5 (444.7) 28.9 (199.3)
Condo D2/l00 51.8 (357.2) 54.5 (375.8) 51.1 (352.3) 23.6 (162.7)
Tensile strength,
psi x 103 (MFa)
Cond.A 57.9 (399.2) 54.9 (378.5) 94.0 (648.1) 12.1 (83.4)
Condo D2/100 55.7 (384.0) 50.2 (346.1) 91.6 (631.6) 11.7 (80.7)
 References 155

Table 7.16 Mechanical properties of polyester lam- Fiberglass reinforcement aligned in a ran-
inate reinforced with chopped strand mat dom manner within the polymer matrix (e.g.
chopped strand mat) produces composites with
Resin content, wt. % 69.8 fairly uniform mechanical strengths in all direc-
Thickness, in (mm) 0.238 (6.05) tions. However, this also tends to produce
Flexural strength, psi (MPa)
composites that have relatively low physical
Cond.A 26 400 (182.0) properties in all directions (Table 7.16).
Condo 02/100 30400 (209.6)
Flexural modulus of elasticity, REFERENCES
psi (GPa)
1. Pitt, c.P. and Harvey J. 20th Anniversary
Cond.A 0.99 x 1()6 (6.83)
Technical Conference, SPI Reinforced Plastics
Cond.02/100 0.94 x 1()6 (6.48)
Oivision, 1965, Section 9-C.
Compressive strength, psi (MPa) 2. Knox, C.E. Non-Metallic Materials (SAMPE) 4,
Cond.A 27 600 (190.3) 127 1972.
Condo 02/100 23 200 (160.0) 3. Horton, R.c. and Adams, R.G. 21st Annual
Conference, SPI Reinforced Plastics Division,
Tensile strength, psi (MPa) 1966, Section 10-A.
Cond.A 14100 (97.2) 4. Knox, C.E. New Horizons in Materials and
Cond.02/100 13800 (95.1) Processing (SAMPE) 18, 527 1973.
5. Peterson, G.P. Properties of High Modulus
Reinforced Plastics, S.P.E./., 57, 1961 January.
BORON, HIGH SILICA, QUARTZ 8
AND CERAMIC FIBERS
Anthony Marzullo

8.1 INTRODUCTION tinuous. Continuous fiber or whisker is capa-

ble of being manufactured to an indefinite
Boron, high silica, fused quartz and ceramic
length. Discontinuous fiber can be chopped
fibers are used in demanding industrial, auto-
continuous fiber, or a short fiber called staple.
mo~ive, electronic, aircraft and aerospace
Discontinuous whisker is manufactured to a
enVIronments. Boards made of ceramic fibers
definite length in batch type processing, while
such as alumina, alumina-silica or zirconia are
continuous whisker is produced by melt
used as supplemental linings to insulate high
processes such as the laser-heated floating-
temperature furnaces. Ceramic tiles contain-
zone process.
ing alumina and silica fibers protect the
Summarizing, there is continuous fiber
aluminum skin of USA's space shuttle orbiter
(indefinite length), continuous whisker (single
during re-entry, where the tile's surface may
crystal, indefinite length), discontinuous fiber
be aerodynamically heated to 1260°C (Banas,
(chopped continuous fiber or staple fiber) and
McCormick and Creedon, 1991). Other appli-
cations of these fibers involve reinforcing discontinuous whisker (single crystal, definite
length).
polymer-, glass-, ceramic-, or metal-matrix
~omposites. If the composite is well designed,
It wIll be tougher than the matrix material by 8.2 MANUFACTURE
itself. As an example, the fracture toughness of
Melt, vapor deposition and chemical processes
SiC is 1.5 MPa m-2, but 8-15 for SiC reinforced
are used to manufacture fibers and whiskers.
with SiC fibers (Richerson, 1992).
Fibers can be classified according to struc- Fibers and whiskers that are made by melt
processing include continuous quartz fibers
ture, diameter (or cross sectional width) and
by drawing from a fused quartz rod, alu-
length. In general, a material is classified as a
fiber if its diameter or cross sectional width is mina-silica staple by atomizing a molten
ceramic stream and continuous alumina
less than 0.0254 m (0.010 in) and length:diam-
~~skers by drawin? from a melt. Vapor depo-
eter ratio is greater than 10:1. Most commercial
SItIon processes mclude chemical vapor
fibers meet these requirements easily. A fiber is
deposition of silicon carbide or boron onto a
called a whisker if its microstructure is pre-
tungsten core to make continuous fibers and
dominantly single crystal. Fibers or whiskers
can be subclassified as continuous or discon- the vapor-liquid-solid (VLS) process to make
discontinuous whiskers. Chemical processing
called sol-gel technology is used to create alu-
mina-based continuous fiber. Other chemical
Handbook of Composites. Edited by S. T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 processes include the creation of polymer
 Continuous fibers 157

precursor fiber that is later pyrolyzed into sili- fibers with a minimum of surface flaws, the
con, carbon and nitrogen-based continuous rod's surface should be flame polished with an
fiber and the reaction between silica and car- oxy-hydrogen flame, as is done in optical fiber
bon to make discontinuous whiskers. manufacturing (Blyler and DiMarcello, 1991).
The manufacturing processes are summa-
rized in Table 8.1.
8.3.3 CONTINUOUS FillERS BY CHEMICAL
VAPOR DEPOSITION
8.3 CONTINUOUS FIBERS Producing silicon carbide fibers by chemical
8.3.1 HIGH SILICA FillERS vapor deposition (CVD) begins by pulling a
melt-spun carbon fiber through a mercury seal
Amorphous high silica fibers can be manufac- and into a tubular reactor. The carbon fiber is
tured by leaching sodium oxide from sodium heated by coupling the mercury seals on the
silicate glass fibers. The glass fibers, made by top and bottom of the reactor to a source of
using conventional glass fiber production tech- electricity. The melt-spun carbon fiber is called
niques, are composed primarily of 74.5% silica a substrate or a core because silicon carbide is
and 24.2% sodium oxide. The fibers are held in vapor deposited onto it. The deposited silicon
a perforated basket, which is suspended within carbide is produced by the reaction of a gas
a cylindrical tank. A solution of HCI and water mixture of silanes and hydrogen that enters
is pumped from the bottom of the tank up to the reactor through a port. Details of produc-
and over the glass fibers for several hours. The ing silicon fibers by CVD are found in
HCljwater/fiber mixture is stirred periodi- Wawner, Jr. (1988) and Textron Specialty
cally. The HCl and water are drained away, the Materials (1993b). Producing boron fibers by
flbers rinsed and a second leaching and rinsing CVD is a similar process (Tsirlin, 1985;
cycle is performed. The fibers are dried at a low Wawner, Jr., 1988; DeBolt, 1982).
temperature (Price and Kielmeyer, 1980). Other
techniques used to make high silica fibers
include spinning from solutions and spinning 8.3.4 PYROLYSIS OF POLYMERIC PRECURSOR
from a gel (Cooke, 1991). FillER

Several ceramic fibers composed of various

combinations of silicon, carbon, nitrogen,
8.3.2 DRAWING CONTINUOUS FUSED
boron and titanium have been produced by
QUARTZ FIBERS
the pyrolysis of polymeric precursor fiber. The
Amorphous fused quartz fibers are made from process consists of four major steps: synthesize
preforms of fused quartz rods. The end of a a spinnable polymer, spin the polymer into
fused quartz rod is heated with an oxy-hydro- precursor fiber, cross-link (cure) the precursor
gen flame and the fiber is drawn, sized and fiber and pyrolyze the cured precursor fiber
either taken up on a spool or combined with into ceramic fiber.
other simultaneously drawn fibers to form a The synthesis of a polymeric precursor is
strand. Fused quartz rods are made by fusing known as Preceramic Polymer Chemistry.
chemically purified ground quartz crystal or Many spinnable polymers have been synthe-
quartz sand. Fused quartz rods are classified sized, including polycarbosilane as a
as Type I or Type II vitreous silica depending precursor for silicon carbide fiber (Yajima,
on the fusion technique. Type I signifies that 1985), polytitanocarbosilane as a precursor for
fusing was done in an electrically heated cru- Si-Ti-C-O fiber (Yamamura et al., 1988) and
cible, while Type II implies fusing with an hydridopolysilazane as a precursor for silicon
oxy-hydrogen flame. To obtain fused quartz carbonitride fiber (LeGrow et al., 1987). In
 ......
Table 8.1 Manufacturing processes for fibers and whiskers CJl
00
Fiber class and manufacturer Trade name" Description Compositionb Manufacturing process OJ
<:)

Continuous fiber: ~
BP Chemical Inc. Refrasil High silica 97.90% Silica (Si02) Leaching sodium from glass fiber .?t
;::.-
Manville Corp. Q-Fiber High silica 99.7% Silica (Siq) Leaching sodium from glass fiber 0<:;'
IPS Glass Fabrics Astroquartz II Fused quartz 99.95% Silica (Si02) Drawn from a fused quartz rod ;:::t"

Fiber Materials Inc. HPQY Fused quartz >99.9% Silica (Si02) Drawn from a fused quartz rod ~
'"
Textron Specialty Materials Textron Boron ~.
Filament Boron B mantle onto a Chemical vapor deposition -l::)
~
tungsten core $:)

Textron Specialty Materials SC5-2, --6, -8, -9A Silicon carbide SiC mantle onto a Chemical vapor deposition ;:t.
N
carbon core $:)
;:::t
Nippon Carbon Co. Ltd Nicalon, ;:...
Ceramic Gr Silicon carbide 58% Si,31% C,l1% ° Pyrolysis of spun polycarbosilane '"'
Ube Industries Ltd Tyranno Fiber Si-TI-C-O 44.2% Si, 11% TI, Pyrolysis of spun ~
24.5% C, 12.3% ° polytitanocarbosilane ;:!
'";:h
-.'"'
Dow Coming Corp. HPZ 57% Si, 28% N, ~
Silicon carbonitride Pyrolysis of spun hydridopolysilazane
1O%C,4%0 '"<;l
Toa Nenryo Kogyo K K Tonen Silicon nitride Primarily silicon Pyrolysis of spun polysilazane
and nitrogen
Mitsui Mining Co. Ltd Almax Alumina 99.5% min ~03 Sintering of slurry-spun fiber
Sumitomo Chemical Co. Ltd Altex Alumina--silica 85% ~031 15% Si02 Calcination of dry-spun fiber
Minnesota Mining Nextel312 Alumina-boria--silica 62%A12Oy Sintering of solution--spun fiber
and MfgCo. 14% ! f Y 24% Si02
Minnesota Mining Nextel440 Alumina-boria-silica 70% 203' Sintering of solution-spun fiber
and Mfg Co. 14% B20 y 24% Si02
Staple:
The Carborundum Company Fiberfrax Alumina--silica 42.0-48.0% ~031 Atomizing a molten ceramic stream
49.0-55.0% Si02
Imperial Chemical Industries Saffil Alumina 96--97% Alpy Sintering of extruded fiber
3-4% Si02
Continuous whisker:
Sapphikon Inc EFG Sapphire Alumina AIP3 Edge-defined, film-fed growth
Discontinuous whisker:
Tokai Carbon Co. Ltd Tokamax Silicon carbide Near stoichiometric SiC Chemical reaction between Si02 and C
Los Alamos Natl Laboratory - Silicon carbide Near stoichiometric SiC VLS (vapor-liquid-solid)
GTE Valenite Corp. TItanium carbide Near stoichiometric TiC VLS (vapor-liquid-solid)
For footnotes see following page
 Staple 159

general, the polymeric precursor is melt spun. operation. They claimed that the fiber proper-
But other spinning methods, such as dry spin- ties were improved due to uniform heating and
ning, have been used. even fiber tensioning (Koba et al., 1989).
The spun precursor fiber is cross-linked to Sumitomo Chemical Co invented a fiber
prevent melting during the subsequent pyroly- with 85% gamma or delta-alumina and 15% sil-
sis step. Yajima cross-linked his polycarbosilane ica by calcining dry-spun fibers. The spinnable
precursor fiber by heating it in air to 200°C. mixture was formed by mixing an aluminum
Later research showed that silicon carbide fiber compound in a solvent and adding water to
properties were improved after curing the poly- get polyaluminoxane. A silica containing com-
carbosilane precursor fiber with electron beam pound, an organic polymer to improve
radiation (Takeda et al., 1992). spinnability and a compound such as lithium
The pyrolysis step consists of heating the to improve fiber properties were added. The
cured precursor fiber to a temperature gener- solvent was removed by distillation. The mix-
ally above 800°C in order to create a ceramic ture is spun into fiber and the fiber is calcined
fiber. (Kadokura, Harakawa and Matumoto, 1991).
Minnesota Mining and Mfg. Co. (3M)
makes fibers composed of alumina, boria and
8.3.5 ALUMINA BASED FIBER
silica by drying and sintering fiber that is spun
E.I. du Pont de Nemours patented a process to from concentrated aqueous solutions. The
make continuous alumina fibers that involved aqueous solution can be composed of an alu-
spinning a viscous slurry. One slurry composi- mina and boria precursor such as a basic
tion was prepared by dissolving an alumina aluminum acetate, silica sol and water. The
precursor such as aluminum chlorohydroxide 14% boria in 3M's Nexte1312 fiber inhibits the
in water and then adding solid alumina parti- formation of alumina crystals and aluminum
cles. An organic polymer might be added to borosilicate is the predominant crystalline
the slurry to aid its spinnability by modifying species. The 2% boria in Nextel440, however,
its viscosity. The viscous slurry was extruded is not enough to inhibit alumina, with eta-
through a spinneret and the green (unfired) alumina being the predominant crystalline
fibers were wound onto a bobbin. The fibers species (50wman, 1988).
were fired at a temperature below the sinter-
ing temperature (calcining) to remove water
8.4 STAPLE
and volatile material. The calcined fibers were
sintered (densified) to remove porosity and The process to make alumina staple is similar
convert the alumina precursor to the stable to the processes that are used in making alu-
alpha-alumina structure. mina-based continuous fibers. An aqueous
This slurry-spinning process was advanced and viscous solution of an alumina precursor,
by Mitsui Mining Co. where the fiber was spun, silica and an organic polymer is extruded
dried, calcined and sintered in a continuous through apertures into a high velocity gas
stream, creating green (unfired) staple. The
green staple is sintered. The silica is added to
Footnotes for Table 8.1
toughen the resultant staple by inhibiting alu-
mina crystal growth while the green staple is
• Most of the trade names are registered trademarks. Each
trademark belongs to the manufacturer listed to its left.
being sintered (Bunsell et al., 1988).
b All data on composition for the continuous and staple Imperial Chemical Industries (ICI) has a
fibers, except for Tyranno Fiber, are from the commercial alumina staple called 5affil. 5affil
manufacturers' product literature. Tyranno Fiber is 3 /lm in diameter and its length is 1-5 cm in
composition is from Weddell (1990). the mat form. Saffil comes in six grades ranging
160 Boron, high silica, quartz and ceramic fibers

from the catalytic grade with an eta-alumina the end of a polycrystalline feed rod is melted
structure, to an aerospace grade with a delta + with a CO2 laser, a grain free seed crystal is
alpha-alumina or an alpha + delta-alumina + brought in contact with the melt and a whisker
mullite structure. is drawn. The LHFZ process can be used to
Alumina-silica staple, also called alumi- produce whiskers with diameters as low as
nosilicate staple, is made by atomizing a 10 Jlffi or as high as 10 Jlm (Haggerty, Wills and
molten ceramic stream. The raw material for Sheehan, 1991).
the melt is a clay mineral or a blend of alumina
and silica. The melt is atomized by pouring it 8.5.2 DISCONTINUOUS WHISKERS BY THE VLS
into the path of a high velocity air stream or by PROCESS
feeding it onto a rotating wheel that throws
out the melt by centrifugal force. The former is VLS stands for vapor feed gases-liquid cata-
called the blowing process, while the latter is lyst-solid crystalline whisker growth. In a
called centrifugal spinning. One commercial particular VLS process to make single crystal
alumina-silica staple called Fiberfrax ranges SiC whiskers, uniformly sized catalyst particles
in length from 70-1000 Jlm and is made by The are applied to carbon substrates. The carbon
Carborundum Co. substrates and catalyst are placed within a
growth chamber. Also within the chambers are
SiD generators, made by impregnating porous
8.5 WHISKERS brick with carbon and Si02• The chamber is
placed within a furnace and heated to a growth
8.5.1 CONTINUOUS WHISKERS
temperature of 1300-1450°C. During this time,
Single crystal continuous alumina whiskers can gases such as CO, CH4, N2 and ~ are fed into
be grown from a pure melt by using a modifi- the chamber. The catalyst particles melt and
cation of the Czochralski method. The modified become supersaturated with material supplied
method is still classified as a moving crystal from the vapor and carbon from the substrate.
technique, but, unlike Czochralski's technique, Solid SiC precipitates from the liquid catalyst
uses a die to shape the growing whisker. A cap- onto the growth substrate. As precipitation con-
illary tube draws the melt up to and over a die. tinues, the whisker grows, lifting the catalyst
The melt flows to the die's vertical edges and ball from the substrate (petrovic and Hurley
stops; so the die's vertical edges control the 1990; Milewski et al., 1985).
shape of the subsequently drawn whisker. A The time necessary to produce SiC whiskers
grain free seed crystal is brought in contact with by the VLS process can be reduced by under-
the melt surface and a whisker is drawn. A lin- cooling or pre-alloying the catalyst balls. After
ear temperature distribution is maintained reaching the growth temperature, the catalyst
along the axis of the growing crystal to avoid balls are undercooled by 150°C, initiating pre-
defects by lowering thennal stresses (LaBelle, cipitation. The chamber is brought back up to
Jr, 1980; Antonov, 1990). Saphikon Inc. grows the growth temperature and held there until
whiskers in many shapes using a similar whisker growth is completed. Another tech-
process that they call EFG (edge-defined, film- nique to reduce the growth time is to pre-alloy
fed growth). One of the shapes is a hollow tube the catalyst particles with carbon and silicon
that may be used as a waveguide (Harrington so that whisker growth occurs immediately
and Gregory 1990). upon reaching the growth temperature
Another moving crystal technique to make (Shalek, 1987; Shalek, Katz and Hurley, 1988).
continuous alumina whiskers is called the Other whiskers produced by the VLS
laser heated floating zone directional solidifi- process include ~O3' boron, MgO and Si
cation process (LHFZ). In the LHFZ process,
 Forms 161

(Wagner, 1970); TiC by using TiCl4 and CH4 as 8.6 PROPERTIES

the primary feed gases and a nickel catalyst
Boron, high silica, quartz and ceramic fibers
(Pearson and Easley, 1992; Narasimhan and
have a high thermal and chemical stability rela-
Bhat, 1992); Si3N4 by using SizCl6 and ~ as the tive to organic fibers. Maximum use
primary feed gases and ~n iron. cata~yst temperatures of boron, high silica, quartz and
(Motojima et al., 1989); and T1Bz by usmg T1Cl~ ceramic fibers, except for boron fiber in an oxi-
and BCl3as the primary feed gases and Au or S1
dizing atmosphere, all are over 1000°C. These
as the catalyst (Withers, Loutfy and Lee, 1988).
fibers have a good balance of specific strength
(tensile strength/density ratio) and specific
8.5.3 OTHER DISCONTINUOUS WHISKER modulus (elastic modulus/density ratio) com-
GROWTH PROCESSES pared with other materials. Tsirlin's (1985)
graphs of specific strength against specific
An early vapor phase growth process to make
modulus for boron or SiC fibers with cores
alumina whiskers involved hydrogen reduc-
illustrate this point. In Table 8.2 note the low
tion. Wet hydrogen was passed over heated
coefficient of thermal expansion for quartz
aluminum powder. Whisker lengths and
fibers, which allows them to be used in parts
diameters were from 1-30 mm and 3-50 ~m,
that are subjected to high temperatures, such as
respectively. Hydrogen reduction of meth~l­
in high speed printed circuit boards.
trichlorosilane was used to produce P-SlC
whiskers (Milewski, 1991). Other vapor phase
growth techniques are evaporation-condensa- 8.7 FORMS
tion and vapor phase reaction. These and
The form that a fiber can take depends on its
other growth techniques, are discussed in
length, diameter and mechanical properties,
Bracke, Schurmans and Verhoest (1984).
such as elastic modulus and tensile strength.
Some commercial processes to make SiC
In general, discontinuous whiskers are too
whiskers use a chemical reaction between silica
short to be spun into yam, however, VLS
and carbon powders. Tokai Carbon manufac-
whiskers about 75-100 mm long might be
tures P-SiC whiskers by heating a mix of
spun into staple yam (Ko, 1989). A fiber's flex-
pulverized silica gel, carbon black and a metal
ibility is related to its diameter, elastic
catalyst. The mix is heated in a nonoxidizing
modulus and tensile strength. A measure of
atmosphere at 1300-1700°C and the whiskers
flexibility is lj[d(E/cr -1)] where d =: diameter,
grow within the pores of the mix (Y~ma~oto, E =: elastic (Young's) modulus and cr =: tensile
1985). The whiskers are under 1.0 ~m m diame-
strength. A greater value indicates a greater
ter and under 300 ~ in length.
flexibility. Obviously, the greater a fiber's flex-
Other companies have produced SiC
ibility, the easier it is to twist, braid and weave
whiskers by chemical reactions: the J.M.
into complex shapes without breakage.
Huber Corp. passed perforated trays filled
As an example, l/[d(E/cr -1)] is about 126
with ground rice hulls over heated carbon pel-
for continuous SiC fiber with a core (d =:
lets and through a horizontal furnace. Ground
80 ~m), 533 for continuous alumina fiber, 1298
rice hulls were used as the reactant material
for continuous SiC fiber (d =: 12 ~m) and 10 101
because they contain both silica and carbon
for continuous fused quartz fiber. The SiC
(Tanaka, Kawabe and Kobune, 1986).
fiber with a core, having a lower flexibility rel-
Silicon nitride whiskers have been pro-
ative to other fibers, is generally used for
duced by heating a silica and carbon mix or
unidirectional reinforcement. The fiber might
rice hulls in a nitrogen atmosphere (nitriding),
also be used in a material-geometric hybrid to
or by heating silica powder in a gas mixture of
axially reinforce a woven preform of finer
ammonia gas and hydrocarbon gas.
 ~
Table 8.2 Properties of fibers and whiskers" 0'\
N
Fiber class and Trade named Description Diameter, Density, Tensile Elastic Coeffident
manufacturer gcm-3 strength, modulus, of thermal O::i
Jlm c
CPa CPa expansion C§
,?!
Continuous fiber: ;:r
c;q"
~
IPS Glass Fabrics Astroquartz II Fused quartz 9 or 14 2.2 6.0 72 0.54x1~rC
;::,:
'"
Quartz Products Corp. Quartzel Fused quartz 9 or 14 2.2 6.0 72 0.54x1~rC 1=;.
Fiber Materials Inc. HPQY Fused quartz 9 2.2 DNA DNA DNA .f>
-$:)
Textron Specialty Textron Boron Boron 100 or 140 2.57 or 2.50 3.600 400 4.5x1~rC $:!
l::I
Materials Filament ....
Textron Specialty SCS-2, -6, -8 Silicon carbide 140 3.0 3.450 400 DNA N
l::I
Materials ;:::
l::I..
Textron Specialty SCS-9A Silicon carbide 81 c DNA 2.7~2.93e 27~lOC DNA
Materials '~"'
Nippon Carbon Co. Ltd Nicalon, Silicon carbide 12 or 15 2.55 2.96c 193e 4.0x1~rC ~
1=;"
Ceramic Gr '-;:n
Ube Industries Ltd Tyranno Fiber Si-TI-C-O 8.5 or 10.5 2.37 2.9 190 3.1 x l~rC <:3"'
'I>
Dow Corning Corp. HPZ Silicon 1~12 2.3-2.5 2.4-3.1c 152-193c 4.1 x l~rC ;rl
carbonitride
Mitsui Mining Co. Ltd Almax Alumina 10 3.6 1.8 320 7.0x1~rC
Sumitomo Altex Alumina- 15 3.25 1.8 210 DNA
Chemical Co. Ltd silica
Minnesota Mining Nextel312 Alumina-boria-silica 1~12 2.7 1.725 138 3.0x1~rC
and MfgCo.
Minnesota Mining Nextel440 Alumina-boria-silica 1~12 3.05 2.070 186.3 3.7x1~rC
and Mfg Co.
Staple:
The Carborundum Fiberfrax, Alumina-silica 2-3 2.73 1.72e 9~103c DNA
Company Chopped
Imperial Chemical Saff11., RF Grade Alumina 3 3.3 2.0 300 DNA
Industries
Discontinuous whisker:
Los Alamos National Silicon carbide 5.9 DNA 8.4 580 DNA
Laboratoryi'
For footnotes see following page
 Applications 163

diameter SiC fiber (Ko, 1989). Fine diameter and two diagonal directions. Figure 8.1 shows
(less than 20 11m) continuous ceramic fiber is the 3D and 50 fabrics. See Ngai (1990) for dia-
more flexible, can be formed into nontwisted grams of 4D, 7D and llD fabrics.
strands (sometimes called nontwisted yam) Continuous boron fibers are supplied in the
and braided or woven into two- or three- form of boron! epoxy prepreg tape and
dimensional shapes. The quartz fibers, being boron! aluminum preform sheets. The
the most flexible, can be twisted into filament boron! aluminum preform sheets are made by
yam and braided or woven into two- or three- resin bonding the boron fibers to the side of
dimensional shapes. aluminum foil. The fibers can be aligned in
Figure 8.1 shows some of the many forms any direction. These preform sheets are used
that fibers can take. Each strand of the fine in solid state (low temperature, high pressure)
diameter ceramic fibers can contain anywhere diffusion bonding processes (Textron Specialty
from 250 to 1000 filaments per strand, depend- Materials, 1993a).
ing on the product. The strands may be Continuous SiC fibers with a core are
bundled to form roving. Fused quartz formed into plasma-sprayed aluminum or
untwisted strands are formed into roving that titanium preform sheets for hot molding (a
contains either 2880 or 4800 filaments per rov- low pressure, hot pressing process) and into
ing. As mentioned earlier, fused quartz fibers 'woven fabric' for investment casting or diffu-
may be twisted into yam at 4 tums per inch, sion bonding. In diffusion bonding, the fabric
with 120, 240, 480, 960, or 1920 filaments per is placed between two metal foils prior to con-
yam and in a variety of yam constructions. solidation. The 'woven fabric' is similar to
Fine diameter ceramic and fused quartz fibers nonwoven roving because the fibers are
can be formed into many multidimensional aligned in one direction and held together by a
structures, such as orthogonal nonwoven fab- cross-weave of metallic ribbons (Textron
rics (Fiber Materials, Inc., 1993; Ko, 1989). Specialty Materials, 1993b; Mittnick 1990).
Orthogonal nonwoven fabrics are so-called Continuous high silica fibers are processed
because there is no interlacing of fibers. Other into a variety of forms such as cloth, tape,
names for these fabrics are xyz fabrics, multi- sleeving, yam, cordage, rope, nonwoven mat,
directional preforms and 3D, 4D, 5D, 7D or blanket, felt, rigid tile and sewing thread.
llD woven blocks. The 5D, 7D and llD blocks Alumina-silica staple fibers are made into
have strands going in the three orthogonal paper, rigid board and cylinders, other rigid
directions and in diagonal directions. The 4D three-dimensional shapes and into most of the
blocks have strands going in two orthogonal high silica fiber forms. Alumina staple fibers
are processed into paper, mat, felt, blanket,
rigid board and cylinders and other rigid
Footnotes for Table 8.2 three-dimensional shapes.
a All values, except those for silicon carbide whiskers, are
from the manufacturers' product literature.
Test methods may vary between manufacturers. 8.8 APPLICATIONS
All values are averages, are for reference only and are
not to be taken as a specification. Table 8.3 provides a sampling of the many
Most fibers are under continual development. Contact applications for boron, high silica, fused
the manufacturers for the most up to date property values.
DNA =Data Not Available. quartz and ceramic fibers. For other applica-
b The silicon carbide whisker values are from Petrovic and tions, see Bracke, Schurmans and Verhoest
Hurley (1990). (1984), Mortensen and Koczak (1993),
, This value in metric units was converted from a value
that was given in English units.
Schwartz (1992) and the other chapters within
d Most of the tradenames are registered trademarks. Each this handbook.
trademark belongs to the manufacturer listed to its left.
~ f-'
....
qq 0\
>i'>
00
j... Q:j
<:)
w
F{
d
.?
;:s-o
~ OQ'
~. ;::-0

a, o 1f ~
~.
SINGLE FIBER- STRAND OR TOW- FILAMENT YARN-
~
~
AlSO CALl.ED
MONOFILAMENT
UNlWISTED OR SUGHTLY TWISTED
COLl.ECnON OF CONTINUOUS FIBERS.
AlSO CAlLED MULTIFILAMENT TOW
2 OR MORE TWISTED STRANDS

SPUN YARN-
rmROVING-
BUNDLE OF STRANDS OR YARN
...l::>
;;::
;::,
....
~ TWISTED STAPLE FIBERS OR N
LDNG WHISKERS ;::,
;:::
:::...
J ~~
F)'
';.:n
[

NONWOVEN AOVlNG- WOVEN AOVING- BRAlD- TAPE-

LOOSE COLl.ECTION OF ROVING BUNDLES. PLAIN WEAVE IS SHOWN CAN BE TUBULAR, 2-D, OR 3-D THIN, UNIDIRECTIONAl PREPREG
HELD TOGETHER BY RESIN pMPREGNATED ABERS OR STRANDS).
OR A PERIODIC STRAND TAPE AlSO REFERS TO A NARROW
STRIP OF WOVEN FABRIC

II MAT-
CHOPPED OR SWIRLED FIBERS
WITH OR WITHOUT A BINDER.
OTHER RELATED PRODUCTS INCLUDE 3D ORTHOGONAL NONWOVEN FABRIC 50 ISOTROPIC NONWOVEN FABRIC
PAPER, FELT, BLANKET AND BOARDS
Table 8.3 Some applications of boron, high silica, fused quartz and ceramic fibers
Fiber type Matrix Manufacturing method Application Reference
Alumina-silica staple Vacuum formed Boards for thermal Literature from various
insulation manufacturers
High silica fiber Water depOSition process Felt for thermal Schuller marketing literature
insulation
High silica fiber Wet felting and interlocking Felt for thermal Carborundum marketing
of loose fibers insulation literature
Alumina whisker Edge-defined, film-fed Laser waveguide Harrington and Gregory,
growth process 1990
Titanium carbide whisker Alumina Cold and hot pressed Cutting tool Narasimhan and Bhat, 1992
Silicon carbide whisker Alumina Hot pressed Cutting tool Tech Trends, 1990; Xiao, Ai
and Yang, 1993
Silicon carbide whisker Silicon nitride Injection molded, HIP Axial turbine blade Buljan et al., 1992
Silicon carbide whisker Aluminum Powder metallurgy Extruded tubing and Mohn and Vukobratovich,
bar for space telescope 1988
Alumina whisker Aluminum Squeeze cast Piston insert Schwartz, 1992
Fused quartz fiber Polyimide Laminated construction Radome Tricoles, 1988
Fused quartz fiber Woven fabric Flexible heat shield Covington and Sawko, 1986
Boron fiber Epoxy Laminated construction Stabilizer for F-14 Schwartz, 1992
Boron fiber Aluminum Diffusion bonding Heat sink for electronic Schwartz, 1992
package
Silicon carbide fiber Titanium Diffusion bonding Bladed ring for turbine Brown, 1993
engine compressor
Silicon carbide fiber Silicon nitride Layup of tape, followed Inter-turbine duct Naik, et al., 1991
by hot pressing
Silicon carbide fiber Silicon carbide Chemical vapor infiltration Turbine engine exhaust Dupont marketing literature
nozzle flap, combuster
elements
Alumina-silica-boria Silicon carbide Chemical vapor infiltration Radiant burner tubes, Lasday, 1993
fiber and other applications

t2"
.....
~.

.....
8i
166 Boron, high silica, quartz and ceramic fibers

8.8.1 NONCOMPOSITES Inorganic Fibres and Composite Materials. New

York: Pergamon Press.
Alumina and alumina-silica staple fibers and Brown, AS. 1993. High-temperature materials
the continuous high silica fibers are used in warm up for debut. Aerospace America 31 (3):
high temperature applications in many indus- 18-27.
tries. Forms of these fibers are used for Buljan, S.T., et al. 1992. 'Development of Ceramic
thermal insulation, personnel protection and Matrix Composites for Application in Ceramic
gasketing. See Carborundum (1991) for the Technology for Advanced Heat Engine
Program', Contract DE-AC05-840R21400, U.s.
many applications of high silica fibers and Department of Energy, April 1992.
Carborundum (1989) for the many applica- Bunsell, AR et al. 1988. Ceramic fibers. In Fibre
tions of alumina-silica fibers. Reinforcements for Composite Materials, ed. AR
Bunsell, pp. 427-78. Vol. 2 of Composite
Materials Series, Series ed. RB. Pipes.
8.8.2 COMPOSITES Amsterdam: Elsevier Science Publishers B.v.
Fibers and whiskers have been used to rein- Carborundum 1989. 'Fiberfrax Ceramic Fiber'. The
Carborundum Company. Niagara Falls, NY,
force polymers, epoxies, glasses, metals, or USA
ceramics. The reinforced materials are called Carborundum 1991. 'Refrasil Products. Product
composites in general and, more specifically, Specifications'. The Carborundum Company.
polymer, epoxy, glass, metal, or ceramic Niagara Falls, NY, USA
matrix composites. The reinforcements are Cooke, T.E 1991. Inorganic Fibers - A Literature
either discontinuous whiskers or continuous Review. J. Am Ceram. Soc. 74 (12): 2959-78.
fibers for the applications in Table 8.3, but sta- Covington, M.A. and Sawko, P.M. 1986. Optical
Properties of Woven Ceramic Fabrics for
ple fiber and continuous whisker may be used Flexible Heat Shields. AIAA/ ASME 4th Joint
as a reinforcement material as well. Usually Thermophysics and Heat Transfer Conference,
the fiber or whisker is coated to cover its sur- 2-4 June 1986, Boston, Mass.
face flaws, provide a diffusional barrier DeBolt, H.E. 1982. Boron and Other High-Strength,
between it and the matrix material and protect High Modulus, Low Density Filamentary
the fiber or whisker from oxidation. In addi- Reinforcing Agents. In Handbook of Composites,
tion, the coating provides a weak interface that ed. George Lubin, pp. 171-95. New York: Van
Nostrand Reinhold.
deflects cracks propagating through the Fiber Materials, Inc. 1993. '3-D Orthogonal Woven
matrix in ceramic matrix composites Blocks' and 'Multi-Directional Woven
Preforms'. Fiber Materials, Inc. Biddeford,
Maine, USA
REFERENCES Haggerty, J.s., Wills, K.c. and Sheehan, J.E. 1991.
Antonov, P.l. 1990. Review of Factors Controlling Growth and Properties of Single Crystal Oxide
the Growth of Shaped Crystals. J. Cryst. Growth Fibers. Ceram. Eng. Sci. Proc. 12 (9-10):
104: 39-46. 1785-1801.
Banas, R, McCormick, M. and Creedon, J. 1991. Harrington, J.A. and Gregory, Christopher C. 1990.
HTP-6.5-22 and HTP-1-22, Reusable Rigid Hollow sapphire fibers for delivery of CO2 laser
Fibrous Ceramic Materials for Heat Shields on energy. Optics Letters 15 (10): 54103.
Future Reentry Vehicles. In Ceramics Today - Kadokura, H., Harakawa, M. and Matumoto, T. 1991.
Tomorrow's Ceramics, ed. P. Vincenzini, pp. Process for Producing Alumina-Based Fiber. US
2889-98. Amsterdam: Elsevier Science Patent 5002750. Mar. 26, 1991. Assigned to
Publishers B.Y. Sumitomo Chemical Co Ltd.
Blyler, Jr., L.L. and DiMarcello, EY. 1991. Optical Ko, EK. 1989. Preform Fiber Architecture for
Fibers, Drawing and Coating. In Encyclopedia of Cerami-Matrix Composites. Ceramic Bulletin 68
Lasers and Optical Technology, ed. Robert A. (2): 401-14.
Meyers, pp. 457-67. New York: Academic Press Koba, K., et al. 1989. Continuous Process for
Bracke, P., Schurmans, H. and Verhoest, J. 1984. Producing Long a-Alumina Fibers. US Patent
 References 167

4812271. Mar. 14, 1989. Assigned to Mitsui Price, G.B. and Kielmeyer, W.H. 1980. Method for
Mining Co Ltd. Making High Purity, Devitrification Resistant,
LaBelle, Jr., H.E.1980. EFG, The Invention and Amorphous Silica Fibers. US Patent 4200485.
Application to Sapphire Growth. J. Cryst. Apr. 29, 1980. Assigned to Johns-Manville Corp.
Growth 50 (1980): 8-17. Richerson, D.W. 1992. Modern Ceramic Engineering,
Lasday, S.B. 1993. Ceramic/Ceramic Composite 2nd edn. New York: Marcel Dekker, Inc.
Components Advance Furnace Systems and Schwartz, M.M. 1992. Composite Materials Handbook,
Processes. Industrial Heating LX (4): 26-9. 2nd edn. ed. M.M. Schwartz. New York:
LeGrow, G.E., et al. 1987. Ceramics from McGraw-Hill, Inc.
Hydridopolysilazane. Am. Ceram. Soc. Bull. 66 Shalek, P.D. 1987. Process for Growing Silicon
(2): 363-67. Carbide Whiskers by Undercooling. US Patent
Milewski, J.V., et al. 1985. Growth of Beta-silicon 4 702 901. Oct. 27, 1987. Assigned to The United
Whiskers by the VLS Process. J. Mater. Sci. 20 States of America.
(1985): 1160 66. Shalek, P.D., Katz, J.D., Hurley, G.F. 1988.
Milewski, JV 1991. Whiskers. In Concise Prealloyed Catalyst for Growing Silicon
Encyclopedia of Advanced Ceramic Materials, ed. Carbide Whiskers. US Patent 4789537. Dec. 6,
R.J. Brook, pp. 516-19. Oxford: Pergamon Press. 1988. Assigned to The United States of America.
Mittnick, M.A. 1990. Continuous SiC Fiber Schramm, Dale E. 1988. Process for Producing
Reinforced Metals. SME Metal Matrix Clinic, Silicon Carbide Whiskers. US Patent 4 789 536.
13-14 November 1990, at Anaheim, Ca. Dec. 6, 1988: Assigned to J.M. Huber Corp.
Mohn, W.K and Vukobratovich, 0.1988. Engineered Sowman, H.G. 1988. Alumina-Boria-Silica Ceramic
Metal Matrix Composites for Precision Optical Fibers from the Sol-Gel Process. In Sol-Gel
Systems. SAMPE J. 24 (1): 26-35. Technology for Thin Films. Fibers. Preforms.
Mortensen A. and Koczak, M.J. 1993. The Status of Electronics. and Specialty Shapes, ed. Lisa C.
Metal-Matrix Composite Research and Klein, pp. 162-83. New Jersey: Noyes
Development in Japan. JOM 45 (3): 10-18. Publications.
Motojimi, S., et al. 1989. Preparation of Whiskers Takeda, M. et al. 1992. Thermal Stability of the Low
and Spring-like Fibres of Si3N4 By Impurity- Oxygen Silicon Carbide Fiber Derived From
Activated Chemical Vapor Deposition. J. Cryst. Polycarbosilane. Ceram. Eng. Sci. Proc. 13 (7-8):
Growth 96: 383-89. 209-17.
Naik, S.K, et al. 1991. Ceramic Matrix Composites Tanaka, M., Kawabe, T. and Kobune, M. 1986.
and its Application in Gas Turbine Engines. Method of Manufacturing Crystalline Silicon
Presented at the International Gas Turbine and Carbide Employing Acid Pretreated Rice
Aeroengine Congress and Exposition, 3-6 June Husks. US Patent 4591492. May 27, 1986.
1991, in Orlando, Fl. Assigned to Tateho Kagaku Kogyo Kabushiki.
Narasimhan, K and Bhat, D.G. 1992. Process for Tech Trends 1990. Ceramic Matrix Composites:
Producing Single Crystal Titanium Carbide Technology and Industrial Applications. Paris:
Whiskers. US Patent 5094711. Mar. 10, 1992. Innovation 128.
Assigned to GTE Valenite Corporation. Textron Specialty Materials 1993a. Boron
Ngai, T. 1990. Carbon-Carbon Composites. In Aluminum Preform Sheets. Textron Specialty
International Encyclopedia of Composites, ed. Materials. Lowell, MA, USA.
Stuart M. Lee, pp. 158-87. New York: VCH Textron Specialty Materials 1993b. Continuous
Publishers. Silicon Carbide Metal Matrix Composites.
Pearson, A. and Easley, M.A. 1992. Process For Textron Specialty Materials. Lowell, MA, USA.
Production of Small Diameter Titanium Carbide Tricoles, G.P. 1988. Radome Electromagnetic
Whiskers. US Patent 5160574. Nov. 3, 1992. Design. In Antenna Handbook, ed. Y. T. Lo and
Assigned to Aluminum Company of America. S.w. Lee, pp. 31-1 to 31-31. New York: Van
Petrovic, J.J. and Hurley, G.F. 1990. Nostrand Reinhold.
Vapor-Liquid-Solid (VLS) SiC Whiskers: Tsirlin, A.M. 1985. Boron Filaments. In Strong Fibres,
Synthesis and Mechanical Properties. In Fiber ed. W. Watt and B.Y. Perov, pp.155-99. Vol.1 of
Reinforced Ceramic Composites, ed. KS. Handbook of Composites, Series ed. A. Kelly and
Mazdiyasni, pp. 93-121. New Jersey: Noyes Yu. N. Rabotnov. Amsterdam: Elsevier Science
Publications. Publishers B.Y.
168 Boron, high silica, quartz and ceramic fibers

Wagner, KS. 1970. VLS Mechanism of Crystal Xiao, H., Ai, X and Yang, H.S. 1993. Effect of
Growth. In Whisker Technology. ed. Albert P. whisker orientation on toughening behavior
Levitt, pp. 47-119. New York: John Wiley & and cutting performance of SiCW-A1203 com-
Sons. posite. Mat. Sci. Technol. Vol. 9:21-25.
Wawner, EE. 1988. Boron and Silicon Yajima, S. 1985. Silicon Carbide Fibers. In Strong
Carbide/Carbon Fibers. In Fibre Reinforcements Fibres, ed. W. Watt and B.v. Perov, pp. 201-37.
for Composite Materials, ed. AK Bunsell, pp. Vol.1 of Handbook of Composites, Series ed. A
371-425. Vol. 2 of Composite Materials Series, Kelly and Yu. N. Rabotnov. Amsterdam:
Series ed. RB. Pipes. Amsterdam: Elsevier Elsevier Science Publishers B.V.
Science Publishers B.V. Yamamoto, A 1985. Process for Preparing Silicon
Weddell, J.K. 1990. Continuous Ceramic Fibers. J. Carbide Whiskers. US Patent 4 500 504. Feb. 19,
Text. Inst. 81 (4): 333-59. 1985. Assigned to Tokai Carbon Co Ltd.
Withers, J.C, Loutfy, R.o. and Lee, CT. 1988. Yamamura,T., et al. 1988. Development of a New
Process to Produce Titanium Diboride Whiskers Continuous Si-Ti-C-O Fibre Using an
as Reinforcement for Metal and Ceramic Organometallic Polymer Precursor. J. Mater. Sci.
Composites. NSF Grant ISI-8760300. National 23: 2589-94.
Science Foundation, Washington D.C October
1988.
CARBON FIBERS 9
Khalid Lafdi and Maurice A. Wright

9.1 INTRODUCTION ~crostructural features developed in the spin-

rung process that also influence the stiffness
Carbon fibers exhibit truly outstanding prop-
erties. As shown in Table 9.1, their strength, 0, s~ngth ~d the thermal and electrical proper~
ties of pitch can be optimized. Nevertheless,
competes with the strongest steels; they can
the tensile strength of PAN-based fibers has
have stiffness, E, greater than any metal,
always been superior to pitch-based fibers'
ceramic or polymer; and they can exhibit ther-
mal and electrical conductivities that greatly
s~ce PAN fibers were developed before pitch
fibers, most structural materials and compo-
exceed those of competing materials. If the
nents use PAN-based fibers.
strength or stiffness values are divided by the
The most important mechanical and physi-
low density, p, - 1800-2200 kg m-3, then their
cal properties exhibited by carbon fibers are the
huge specific properties (o/p, E/p) make this
elastic modulus, tensile strength and the electri-
class of materials quite unique.
cal and thermal conductivities. These
. All carbon fibers sold commercially are fab-
properties are sensitive to the crystallite size
ncated from polyacrylonitrile (PAN) or from a
and perfection of the graphene layers devel-
coal, petroleum or synthetic pitch. PAN-based
oped within the carbon fiber and depend for
fibers are produced from a solubilized mixture
the most part on the degree of molecular align-
that is wet or dry spun to produce a fiber,
ment with respect to the fiber axis. Growth and
ostensibly for use in the textile industry. This
alignment of such layers occur within the pre-
fiber is stabilized and carbonized to produce a
cursor and within the solid carbon fiber when it
carbon fiber. Aerospace grade material can be
is heated to high temperature. However, it can
obtained in tows that contain between 3000
be argued that the most extensive structural
and 12 000 fibers. Lower performance materi-
rearrangement occurs when the basic structural
als are usually formed using larger tows that
units (BSUs) present in the original precursor
contain up to 320 000 fibers. PAN-based car-
are large and plate-like so that the shear stresses
bon fibers are cheaper when produced from
generated during spinning can align these large
larger tows.
areas more easily. Heating the fiber to high tem-
Pitch fibers are melt spun products obtained
perature after spinning relaxes the spinning
in small tow sizes varying from 2000 to 4000
stresses and allows the oriented regions to
fibers. They are usually of larger diameter
grow. Compared to PAN, the basic structural
(-10-15 \llll.) than fibers formed from PAN.
units in an original mesophasic pitch are much
. The spinning process is controlled by the
larger in area and length and are not twisted to
pItch-based carbon fiber producer; thus,
the same degree. The resulting ease of graphiti-
zation and alignment of the graphene layers
and the development of large crystallites, pro-
Handbook of Composites. Edited by S.T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7
duces a large elastic modulus and electrical and
170 Carbon fibers

thermal conductivity parallel to the fiber axes. properties (a/p, E/p) tend to remain relatively
Unfortunatel)" large graphitized regions tend to constant; structures designed to resist a given
produce high local stress concentrations, espe- set of loads tend to weigh the same irrespective
cially if they are misaligned (Reynolds and of whether they are made from aluminum, tita-
Sharp, 1974; Reynolds and Moreton, 1980). nium or steel. These mechanical property I
Thus, pitch-based fibers tend to exhibit high density relationships are not followed by car-
modulus and high electrical and thermal con- bon fibers; so, compared with metals, similar
ductivities, but low strength. PAN-based fibers structures made from carbon fiber reinforced
tend to be of intermediate modulus and rela- composites will be lighter.
tively high strength. The most successful use of carbon fiber rein-
The mechanical properties of structural met- forced materials has been for military
als generally vary in direct proportion to purposes, especially aeronautical structures
density. Thus, as shown in Table 9.1, the specific (Schwartz, 1984; Hadcock, 1982). Composite

Table 9.1 Physical and mechanical properties of typical carbon fibers

Supplier/Fiber name Tensile Elastic Fracture Density, alp, E/p,

strength, modulus, elongation, p (kg/m-3) 106 m2 S-2 106 m2 5-2
a (GPa) E (GPa) %
Pitch Type
Amoco/Thomel P100 2.2 690 0.3 ~2150 1.02 321
P120 2.4 830 0.3 -2150' 1.11 386
Petoca/Carbonic HM50 2.8 490 0.6 -2000' 1.40 245
HM60 3.0 590 0.5
HM80 3.5 790 0.4 ~2000a 1.75 395
PAN Type
Amoco /Thomel T-300 3.45 231 NA 1760 1.96 131
T-50 2.9 390 NA 1810 1.60 215
T-40 5.65 290 NA 1810 3.12 160
Akzo/Fortafil F-5(c) 2.76 345 NA 1800 1.53 192
F-3(c) 3.80 227 NA 1700 2.24 133
RK Fibers/RK RK30 >3.0 -230 NA 1780 1.68 129
RK25 >2.5 ~230 NA 1780 1.40 129
Toray /Torayca M46J 4.2 436 0.5
T300 3.5 230 1.2 1750 1.71 143
T800 5.5 294 2.0
Toho /Besfight ST-1 3.6 240 1.5 -1800b 2.00 160
ST-2 4.0 240 1.7 ~1800b 2.20 141
ST-3 4.4 240 1.8 ~1800b 2.38 133
Metals
AlUITIillumC 0.172 73 NA 2720 0.063 27
1itaniumc 0.324 110 NA 4500 0.072 24
SteelC 0.414 199 NA 7860 0.052 25
aEstimated from Tanabe et al. 1987; bEstimated from Johnson 1987; cEshbach and Souders, 1974;
NA, Not Available.
 Overview 171

use in civilian aviation has been more limited, a precursor material is prepared, spun into a
largely because of cost. However, in 1985 the fibrous shape, stabilized to change it from a
European Airbus consortia used carboni epoxy thermoplastic to a thermoset, then heated until
vertical stabilizers on their A310 and, since all unwanted elements are expelled. Depending
1993, delivered their A340 and A330 models on the final heat treatment temperature, a fiber
equipped with composite tail sections, floor is produced that is primarily carbon. A strik-
panels, landing gear doors and carbon-carbon ingly similar production technique is used to
brakes. The new Boeing 777 and the projected produce either type of fiber; however, the initial
McDonnell-Douglas MD-12 contains similar pretreatment and the chemical reactions that
composite structures (High-Performance occur within either PAN or pitch during stabi-
Composites, 1994). lization and carbonization are markedly
different. These differences are discussed in the
following sections since they determine the
9.2 OVERVIEW
ultimate properties and the cost of the carbon
In this chapter we have chosen to combine, fibers that are produced.
rather than separate, the discussion of the
preparation of carbon fibers from PAN or from
9.2.1 POLYACRYLONITRILE
pitch. This style of presentation is attempted
since the methods used to fabricate carbon The chemical composition of PAN-based pre-
fibers from either of these precursor materials cursors tends to be proprietary; however, in
seem to be almost identical, as shown in general it consists of small diameter linear
Fig. 9.1; for example, both methods involve the molecules that are made up from nitrogen,
preparation, spinning and subsequent thermal hydrogen and carbon (Fig. 9.2). Spinning
degradation of organic precursors. Essentially, tends to orient these molecules parallel to the

Pitch Polyacrylonitrile
precursor precursor
I
Particulate removal
I
Pretreatment
(chemical, thermal, mechanical)

I
Fiber spinning
Wet or dry spinning
I
Solvent extraction
I I
Stabilization
at 200-260°C
I
Carbonization
to 1500°C
I
Graphitization
to 2500°C
Fig. 9.1 Schematic illustration of process to produce carbon fibers from polyacrylonitrile and pitch.
172 Carbon fibers

"'" /CH2 "", /CH 2 "", /CH2 "", /CH2 "", /

CH CH CH CH CH
6 6 666
C'%N C'%N C'%N C'%N C'%N

Fig. 9.2 Structure of the ideal PAN molecule (Riggs, 1976).

fiber axis, but they continue to be randomly of carbon fibers from PAN is relatively minor;
oriented transverse to that direction. Thus, a more important is the chemical makeup of the
fiber with a twisted fibrillar structure is pro- PAN and the presence of small amounts of
duced. This fiber is supplied by the textile other constituents that influence the complex
industry to the carbon fiber fabricator who sta- chemical reactions that occur during stabiliza-
bilizes it under tension before converting it tion and carbonization. Stabilization involves
into a carbon fiber using a controlled heat cydization of the oriented molecules and
treatment process. Apart from the obvious results in the release of most of the hydrogen
generation of the fibrous shape, the impor- and part of the nitrogen as NH3 and other
tance of the spinning process to the fabrication nitrogen compounds (Fig. 9.3). The role of the

xxr
Initiation Cyclization

Abnormal structures
end structures
impurities
N N
H intra- or
n = 0, 1, 2 ... Abstraction intermolecular
m = 0, 1, 2. transfer

CN~:
. .:
: 1 ~
:ccr~
I
; ; .CN CN
:. N ·:n NH N NH
UrH
L\~ 2 CHbCH
I m 2 CN

j j
CN 2. Transfer
re-initiation
1. Cyclization
end group chain
reactions scission

x:c
CN
NH2J HCN fragments N

CH2
Fig. 9.3 Release of HCN and NH3 during cyclization of PAN fiber (Grassia and McGuchan, 1971b).
 Pitch precursor treatments 173

retained nitrogen is very important to both the Mochida et al., 1988; Lewis and Nazem, 1987a).
crosslinking process and to the development Today, fibers are produced commercially from
of optimum properties during carbonization. rayon, PAN and the various pitches. However,
the process to produce fibers from rayon is
very expensive because it involves stretching
9.2.2 PITCH
at very high temperatures and the yield of car-
A pitch precursor taken from a petroleum or bon after carbonization is small. Rayon-based
coal tar feedstock initially contains individual fibers are therefore fabricated in such small
molecules that exhibit appreciably different amounts that they are really of no commercial
molecular weights. These untreated precur- importance. Additional information can be
sors have been used to produce fibers; obtained from: Yanagida et aI., 1991; Bacon,
however, they are isotropic and exhibit rela- 1973; US Patent 3 107152, 1963; Yoneshoga
tively poor mechanical and physical and Teranishi,1970.
properties. Conversely, the carbon fiber pro-
ducer can pretreat the pitch to develop a
9.3 PITCH PRECURSOR TREATMENTS
continuous anisotropic phase (similar to a
mesophasic liquid crystal) or a two phase mix- A typical pitch precursor material is obtained
ture that becomes highly oriented during the from either the distillation products involved
subsequent spinning process. In contrast to the in the chemical treatment of decanted oil or as
chemical changes occurring in PAN, physical a by-product of the production of metallurgi-
changes are responsible for the final properties cal coke from coal. In a conventional
of pitch-based carbon fibers. Essentially, the as-received pitch, basic structural units (BSUs)
precursor isotropic pitch is pretreated to pro- are already present to a degree that pitch can
duce a two phase mixture that is be considered a random suspension of highly
predominantly anisotropic. During spinning aromatic molecules similar to coronene (mole-
and drawdown, this mixture is very strongly cular weights between 600 and 900)
oriented both parallel and transverse to the surrounded by a liquid of smaller molecular
fiber axis. Oxygen added during stabilization weight. Heating such a system initially
tends to crosslink these large molecules in a reduces the viscosity. Eventually, however, the
simple way before being released on car- viscosity increases as the BSUs grow and coa-
bonization as CO, CO2 and H 20. More recent lesce to form larger entities (molecular weights
interest has centered on producing fibers from between 1000 and 4000) with a specific long
synthetic pitches. These require no extensive range anisotropy called 'mesophase' (Brooks
pretreatment and they stabilize faster at a and Taylor, 1965). Continued heating eventu-
given temperature. Fibers made from these ally causes an inversion in which the
pitches should cost less to produce. mesophase spheres become the continuous
Precursor materials that have been used to anisotropic phase within which are suspended
produce carbon fibers include: polyamides, spheres of the isotropic low molecular weight
polyesters, polyvinyl alcohol, polyvinylidene material. The rate of viscosity increase is very
chloride, poly-p-phenylene, phenolic, naptha- slow at low temperatures, but accelerates as
lene, naphthalene-phenanthrene, alkylbenzenes, the temperature is increased. At temperatures
rayon, polyacrylonitrile (PAN) and various greater than 350°C, the pitch begins to form
petroleum, coal tar and synthetic pitches coke by a process of thermal degradation and
(Ezekiel, 1969; US Patent 3533741, 1970; gas evolution. The variation in viscosity with
Shindo, Nakanishi and Soma, 1969; Boncher, temperature for numerous pitch fractions has
Cooper and Everett, 1970; French Patent 1 535 been documented (Bathia, Fitzer and
800, 1968; Kawamura and Jenkins, 1970; Kompalik, 1984) and is shown in Fig. 9.4. As
174 Carbon fibers

TS CTP A~40 1:1 1:4 11 TI 1:4

I 1:1
I CTP TS
I

Temperature °C
Fig. 9.4 The variations of apparent viscosity with temperature of various pitch fractions (Bathia, Fitzer and
Kompalik,1984).

can be noted, at a given temperature, the vis- A 1980 patent (US Patent 4208 267, 1980)
cosity of any pitch is greater the more discussed a different method of producing
anisotropic phase it contains. By cooling after isotropic/ anisotropic pitch mixtures. The
partial reaction, it is possible to produce a two method essentially consists of dissolving part
phase pitch with a viscosity appropriate for of the original isotropic pitch in an organic sol-
spinning or infiltration. vent such as benzene, toluene, xylene, etc. The
A number of experiments have been carried material that is insoluble can then be con-
out designed to accelerate the process of pro- verted, by heating, into a material that is
ducing two phase pitch mixtures. These greater than 75% anisotropic. The efficiency of
involved heating, stirring, bubbling an inert this process is quite poor however, since only
gas through the liquid (sparging), or the com- a very small amount of this 'neomesophase'
bination of stirring and sparging (US Patent can be produced from a given pitch. For
3629379, 1971; US Patent 3919383, 1975; US instance, using Ashland A260 pitch, about
Patent 3 974 264, 1976; US Patent 4017327, 75-90% of the initial pitch will dissolve.
1977). In this way, an appropriate anisotropic Using Ashland A240, 80-90% of the pitch dis-
concentration can be produced in much shorter solves.
time periods (hours instead of days) with Variations of the gas-sparge process (possi-
improved spinning characteristics. This latter bly associated with a chemical fractionation)
characteristic seems to be associated with the can be made to change the characteristics of a
smaller molecular size existing within the two resultant pitch. Lafdi, Bonnamy and Oberlin
phase mixtures formed by stirring and sparg- (1991a; 1991b; 1991c; 1992) and Lafdi and
ing and the smaller variation of molecular size. Oberlin (1994a;1994b) have indicated that
 Spinning conditions 175

some pitches exposed to a nitrogen sparge at acetamide before extruding it into a 'coagula-
atmospheric pressure or a hydrogenation tion' bath through a spinneret. In the
treatment at high pressure produce a continu- wet-spinning process, the fiber is solidified by
ous strongly anisotropic material that contains using a coagulant (such as ethylene glycol)
small particles of a weakly anisotropic mater- which extracts the solvent from the polymeric
ial. During spinning, the second phase fiber. In a manner similar to the dry spinning
becomes completely absorbed (or trans- process, the rate at which the solvent is
formed) to produce a uniformly anisotropic extracted from the polymer as it passes
fiber. They believe that sparging disturbs the through the coagulation bath governs the final
formation of an anisotropic phase that shape of the fiber. The temperature, concentra-
exhibits large differences in molecular weight. tion and circulation rate of the fluid in the
Indeed, they suggest that the spheres of coagulation bath are known to affect the struc-
anisotropic material contain BSUs that are ture and hence the physical and mechanical
only weakly associated. In contrast to the properties of as-spun fiber. Many companies
strong molecular orientation exhibited by the have added a supplemental stage to the spin-
Brooks and Taylor type of 'mesophase' ning process that is designed to reduce the
(Brooks and Taylor, 1968), the common orien- water content. This additional step tends to
tation of this new anisotropic material results increase the molecular orientation within the
from the statistical orientation of small units. fiber (US Patent 3 846 833, 1975; US Patent
The pitch then behaves as a two component 3841079, 1974).
gel which exhibits a long range anisotropy in A typical acrylonitrile-based precursor con-
the bulk. The short range change in orientation tains several percent of various co-monomers
of the carbon units produces sharp changes in such as methyl acrylate or vinyl acetate which
orientation similar to grain boundaries. Such improve the precursor's spinnability or fabric
regions produce a zig-zag nanotexture in the properties. Though not added to aid car-
resultant carbon fiber that prevents inter-sheet bonization specifically, they have been found
gliding and provides a crack inhibiting func- to influence the properties of the resulting car-
tion that contributes to relatively high ultimate bon fiber. Many modified PAN polymers such
strength values. as acrylonitrile-hydroxyethylene, acryloni-
trile-vinyl chloride-itaconic acid (French
Patent 2 328 723), polyacrylomidoxium (US
9.4 SPINNING CONDITIONS Patent 3 767 773, 1973) have been investigated
to obtain a suitable as-spun fiber capable of
9.4.1 POLYACRYLONITRILE
producing a fiber with a large carbon yield
Dry and wet spinning of polyacrylonitrile pre- after carbonization.
cursors have been used. In the dry process the
polymer is solubilized and spun into a current
9.4.2 PITCH
of hot air that removes the solvent.
Unfortunately, solvent removal tends to be According to Singer (US Patent 3919383
quite rapid and can cause the outer portion of (1975», in order to spin a fiber, pitch must be
the fiber to solidify before the solvent can dif- heated to produce a viscosity between 10 and
fuse from the fiber's center. The large diffusion 200 poise (1-20 Pa s). However, temperatures
gradient that develops can seriously affect the greater than about 350°C cannot be used to
final shape of the fiber (Edie and Diefendorf, obtain the required viscosity because thermal
1993). The more common wet spinning decomposition of the pitch will occur. In addi-
method involves solubilizing the polymer tion, spinning should be carried out above a
with a polar solvent such as dimethyl minimum temperature of about 200°C since
176 Carbon fibers

this determines the maximum temperature The spinning of two phase mixtures is not
that can be used in the subsequent stabiliza- an easy commercial operation since the phases
tion step. These temperature requirements exhibit different viscosities and densities. The
define a processing window into which suit- strongly anisotropic continuous phase con-
able pitches must fall. Using literature data tains within it a less anisotropic (or isotropic)
(Lewis and Nazem, 1987a; Mochida et al., 1988; phase which exists as spheres. If the diameter
Yanagida et al., 1991), White (1992) has shown of the spheres is large relative to the spinning
that the smallest window exists for 100% orifice, localized weak sections of extrudent
anisotropic pitches processed from coal or can be produced that can break and make it
petroleum. A larger window exists for material difficult to maintain a continuous fiber thread.
partially transformed from the same precur- In addition, since the stabilization rates of each
sors; however, as shown in Fig. 9.5, the largest phase differ, one phase may be over stabilized
window exists for synthetic pitches. relative to the other and it becomes difficult to

Temperature (0C)
450 400 350 300 250 200
5

4 Window Boundaries 104

-8!. 3 103
-
--
I/)
Softening I /)

-
point as
>- a..
'in 2 102
0 ~
U)
0
I/) 0
:; 101 0

-
1 f/)

1: :>
....
(I)
I::
as
c. 1 !!?
c. 0 as
« A
c.
a.
0)
0 -1 10-1
«
-I B Lowest viscosity pitch
that can be stabilized
-2 10-2
' " High temperature limit, to
avoid pitch decomposition
-3 10--3
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2

Inverse Temperature, I/Tx 103 (K-1)

Fig. 9.5 Processing window for injection of mesophase pitch (White, 1992).
AA': The lowest viscosity reported for mesophase pitch prepared from petroleum or coal-tar pitch by
pyrolysis to 100% mesophase.
BB': The lowest viscosity reported for a petroleum- or coal-tar-based mesophase pitch only partially
transformed but with the mesophase acting as the continuous phase.
CC': The lowest viscosity reported for a chemically-derived fully-transformed mesophase pitch.
 Stabilization of polyacrylonitrile 177

generate optimum properties in the resulting

carbon fiber. This fact seems to have been rec-
ognized early, since Singer (US Patent
3919383, 1975) describes methods to remove
the isotropic component from the two phase
fiber immediately after spinning. (a) (b) (c)
Due to the proprietary nature of the various
procedures, little is known of the spinning
process itself, however, a number of patents
have been awarded (US Patent 4331 620,1982;
Matsumoto 1985; US Patent 4504454, 1985).
These have been discussed in other publica-
tions (Wright, 1989; Wright and Palmer, 1994); Fig. 9.6 Typical pitch-based carbon fiber
most deal with attempts to combat the two microstructures (Edie, 1990): (a) radial; (b) onion-
phase nature and non-Newtonian behavior by skin; (c) random; (d) flat-layer; (e) radial-folded; (f)
stirring the pitch in the spinneret. Apparently, radial with wedge.
agitation of mesophase within the spinneret is
part of the technique used by the Petoca Oil at temperatures lower than those necessary to
Company to randomize the cross sectional spin anisotropic-isotropic mixtures obtained
microstructure and produce the folded from petroleum or coal tar pitch; thermal
graphene layers that are thought responsible decomposition during spinning is therefore
for the relatively high strengths exhibited by not a problem. Stabilization of such pitches
their 'carbonic' carbon fiber (Guigon, Oberlin must be performed at correspondingly low
and Desarmot,1984a; 1984b). As seen previ- temperatures; however, when they are com-
ously in Table 9.1, for similar modulus pared to petroleum and coal tar mesophasic
materials, the strength of Petoca's carbonic pitches, stabilization at a given temperature
fibers (HM-series), while still not as good as occurs at a significantly faster rate. This fea-
PAN based fibers (T-series), is appreciably bet- ture, when combined with the fact that no
ter than other commercially available pitch pitch pretreatment is necessary, should
material (P-series). enable fibers to be produced at significantly
The flow of pitch through the spinneret lower cost.
(Matsumoto, 1985), the geometry of the die The result of all spinning processes is a PAN
hole (Singer, 1978; Yamada et al., 1984; US or pitch fiber that can be changed into a carbon
Patent 4504454, 1985), pitch viscosity, die fiber by heating at a very slow rate to a tem-
swell and drawdown are all known to affect perature of about 1500°C. A higher rate of
the microstructure of the resulting pitch-based heating will usually melt the fiber unless it is
fiber. Indeed, the above authors illustrate their first converted from a thermoplastic to a
comments with fiber microstructures that vary thermoset. This conversion or stabilization
from circumferential through random, radial treatment is done by heating to a relatively low
or radial with a crack (Fig. 9.6). temperature for an extensive time period in an
Spinning experiments using 100% atmosphere containing oxygen (usually air).
'mesophasic' synthetic pitches produced
from napthalene (Ohtsuka, 1988),
9.5 STABILIZATION OF POLYACRYLONITRILE
napthalene-phenanthrene (Lewis and
Nazem, 1987b) and alkylbenzenes (Yanagida This process converts the thermoplastic,
et al., 1991) have been reported. The viscosity as-spun polymer into a thermoset that is capa-
of these materials is appropriate for spinning ble of maintaining its shape during
178 Carbon fibers

carbonization (Lavin, 1992; Yooh, Korai and 260°C for a period of time that varies between
Mochida, 1994). The operation is identical for thirty minutes and several hours. During sta-
both PAN and pitch based fibers. However, bilization, several interdependent chemical
the stabilization of anisotropic pitches reactions occur. The reaction that dominates is
involves simple cross linking of plate-like primarily determined by the chemical compo-
molecules whereas the stabilization of PAN sition of the initial precursor, the spinning
involves many different chemical reactions. history, the final composition of the as-spun
Stabilization of both PAN and pitch is an fiber and the stabilization heating schedule.
exothermic process, so great care must be A PAN polymer mainly consists of acry-
taken to control the rate of reaction and to lonitrile entities -CH2CH(CN)- which are able
avoid thermal runaway which melts the fiber to cyclize (Johnson et al., 1972) with the help of
and is a fire hazard. an initiator into a presumably linear 'ladder
Commercial stabilization is carried out by polymer' similar to that shown in Fig. 9.7. In
heating the PAN fiber in air between 200 and general, the pendant nitrile groups of PAN

l
.~ I
'~'
I I ~.
-------------, . .

CYCLIZATION I eN p~NN eN DEH~OGENATI~~"" ______ _

I \ HOMOPOLYM
"~" /~" \ PAN
, _,~",AuA.. ,)" . . I I I . \ (not heat treated)
.' N N N " eN eN eN ~-------------

DEHYDROGENATION L~'~ CYCLIZATION

o OXIDATION 0 OH -------------'"
"-
.....' ........ ',---------
STABILIZED
H40% 30% 20% 10% PAN
CARBONIZATION
."~,,,

/~N~NAN)'"
"')~~~CHXCH)/'
CARBONIZED
/l....NAN N'" '"
PAN
'" (tht)
.-' /' I
I ------------------
I
I
" I
I -------------------

.-' / / CARBONIZED
", / / PAN
/ / (hht)
N ":'_====~_I . . . . -----------------
Fig. 9.7 The process of PAN stabilization and subsequent carbonization (Fitzer and Heine, 1988).
 Stabilization of polyacrylonitrile 179

first become crosslinked to form a ladder poly- A balance should be kept during stabiliza-
mer. Initiation of this process is catalyzed in tion as to hydrogenation degree. A large
some cases by the presence of a small amount hydrogen content can result in a small N/C
of reactive copolymer such as itaconic acid. ratio which increases the temperature at which
Oxygen is then incorporated into the ladder the local molecular ordering occurs.
polymer under a number of possible schemes Conversely, increasing the available oxygen
which have been described by Watt and decreases the size of and the temperature at
Johnson (1975) and by Clarke and Bailey which the units of local molecular order
(1973) and are shown in Fig. 9.8. Cyclization (LMOs) are formed. In addition, since the vis-
and stabilization induce tremendous shrink- cosity increases as crosslinking (stabilization)
age into the polymer. Longitudinal shrinkage proceeds, the mobility and growth rate of the
is resisted mechanically; however, the diame- LMOs decrease; hence their final size remains
ter of the fiber is allowed to decrease. small. The smaller the size of the LMOs, the

la

Ib

II

III
180 Carbon fibers

less graphitizable is the carbon and the lower mal product, i.e. forming the LMOs at the
the properties of the fiber. In order to ensure smallest reasonable temperature, retaining the
appropriate N/C ratios at reasonable temper- largest nitrogen content beyond the tempera-
atures, stabilization should result in only a ture of LMO formation and incorporating an
moderate degree of crosslinking. In addition, optimum amount of oxygen during stabiliza-
slow heating rates during precarbonization tion to prevent polymer melting without
should permit hydrogen and delay nitrogen inducing too small a LMO size.
emissions; both of these effects lower the tem- It is known that the overall oxygen content
perature at which extensive formation and should be between 8 and 12 wt. % in order to
growth of LMOs occur. completely stabilize PAN fibers (US Patent
A commercially acceptable rate of stabiliza- 4069297, 1978). Less than 8 wt.% oxygen
tion requires the use of as high a temperature gives a large weight loss on carbonization due
as possible. However, since the reactions that to excessive evolution of gases from the
occur during stabilization are exothermic, it is incompletely stabilized central core; more than
most important to limit the oxidation rate and 12 wt.% oxygen degrades surface layers and
to prevent uncontrollable temperature the properties of the final fiber ijohnson, Rose
increases. These conflicting requirements have and Scott, 1970). Exactly what an average
resulted in the development of alternative value of 8 wt.% translates into for the specific
methods of stabilization. These include stabi- oxygen content of the surface layers and the
lization in: hydroxylamine solution (US Patent core region is unknown, but it would obvi-
3 767 773, 1973), aminophenoquinones (US ously depend on fiber diameter and the
Patent 4004 053, 1976), aminosiloxanes (US kinetics of the stabilization process (diffusion
Patent 4009248,1977), amine salts (US Patent or reaction controlled). A large diameter PAN
4009 248, 1977; US Patent 4024227, 1978; US fiber containing an average of 8 wt. % oxygen
Patent 4031288, 1978) or stabilization in gas exhibiting diffusion controlled stabilization
phases such as mixtures of NO and N02, Br2 kinetics would probably be composed of
and O2, or HCI and 02. Other stabilization highly degraded surface layers with perhaps
processes have been proposed that are an under-stabilized central core. Conversely, a
designed to reduce the cost and/or decrease very thin fiber exhibiting reaction controlled
the stabilization time. It seems possible to stabilization kinetics might be completely and
reduce the time by stabilizing the fibers in per- homogeneously stabilized with an average
sulphate (US Patent 3 650 668, 1972), cobalt oxygen content of less than 8 wt. %.
salt (US Patent 3 656 882, 1972), nitric acid (US Presumably a similar statement can be made
Patent 3814377, 1974; US Patent 3 656 883, for the stabilization of pitch fibers. In the sta-
1972), or to control the final quality of the fiber bilization process, the effect of fiber diameter
by stabilizing in carboxylic acid (US Patent on the rate of oxygen uptake is important. The
3814377, 1974; US Patent 3 656 883, 1972), or three curves shown in Fig. 9.9 illustrate the
nitrophenol. Processes designed to produce slower rate of oxygen uptake exhibited by fat-
fibers of larger modulus have also been devel- ter fibers. In addition, larger diameter pitch
oped. These involve stretching the precursor fibers tend to exhibit a diffusion controlled sta-
during stabilization (US Patent 3917776,1975; bilization that produces an under-stabilized
US Patent 3 677 70S, 1976). central core and an over-stabilized fiber skin.
Fiber manufacturers attempt to fit the Smaller diameter pitch fibers appear to exhibit
physico-chemical conditions of the various reaction controlled stabilization since no skin-
operations cited above (nature and proportion core type microstructures are observed.
of the co-monomers, cyc1ization, stretching,
stabilization, carbonization) to obtain an opti-
 Chemical changes during carbonization 181

10.
9.00

B.00
240 Deg C

,,:,.
7.00 ·················rsS"ii";g·c···········
iF 6.00
'-'
5.00
~

~
4.00
.............. - ..... _--_.. --_......... __ .............. _.... _---_ ......... _-----_ ..
3.00

2.00

~ 1.00
z•••••••••••• ••••••••••••••••••

~ 0.00
-1.00
Petroleum pitch Fiber, 235 Deg C
X Fiber Diameter 100~m
y Fiber Diamerer 55~
z Fiber Diameter 38~m
-2.e0
Solid line - mesophasic particles
-3.00

-4.e0 (O.tdtzed Oxygen,

';;"---'-'""--""""1r:ee-'--~---lnr--'---'---'-~iii"""---'---'---~'"=-"----'--
-S . ee0.00 5.00 le.00 15.00 ..............,j
25.00
Time (hours)
Fig. 9.9 Eff~ct of diameter on the stabilization of mesophasic particles and pitch-based fib (K b 1
Wapner, Wnght, 1989). ers owe,

9.6 CHEMICAL CHANGES DURING rate of emittance are important control para-
CARBONIZATION meters since they affect the strength of the
resultant carbon fibers.
!he carbonization of stabilized PAN and pitch
A stabilized polyacrylonitrile fiber which
mvolves controlled heating to a temperature
contains about 11 wt.% oxygen can be ther-
of about 1500°C. The majority of gases emitted
from either the PAN or the pitch are emitted mally degraded by heating at a slow heating
rate (Riggs, Shuford and Lewis, 1982) in an
before a temperature of 1000°C is reached and
inert atmosphere such as nitrogen or a reactive
the emission is primarily from unstabilized
environment where nitrogen gas is bubbled
regions Gain and Alhiraman, 1987; Lewis,
through acid (US Patent 3 972 984, 1976) or
1982). Indeed, the quantity of gases emitted
water (US Patent 3 677 70S, 1976; US 3 656 903,
from an unstabilized central core of either
1972; US Patent 4 039 341, 1976). As the tem-
PAN or pitch can be so large that the fiber can
perature increases, many complex reactions
disintegrate. Great care should therefore be
take place resulting in the evolution of volatile
taken to. ~etermine the optimum heating rate
products. For example, when the fiber is ini-
for stabilized or under-stabilized fibers. In
tially heated, cydization occurs with the
some cases, hold times should be incorporated
release of large amounts of HCN and NH3• Up
into the heating cycle. Both materials emit a
to 450°C, HCN, acrylonitrile, propionitrile,
variety of gas molecules containing oxygen,
NH3 and Hp are emitted. Subsequently, at
hydrogen and carbon; however, a major differ-
around 500°C and 700°C copious quantities of
ence between PAN and pitch involves
},lCN and water vapor are emitted, respec-
nitrogen containing compounds which are
only emitted from PAN. The temperature and tively. All of these emissions are believed to
182 Carbon fibers

come from reactions involving crosslinking of <4.0.------,,---------,

individual molecules. Evolution of nitrogen
starts near 700°C; so fibers produced after
being heated to 1000°C retain only about
5.8 wt.% nitrogen and have lost about SO wt.% 3.S
soo
of the original PAN precursor fiber.
Results obtained from experiments involv-
ing slow pyrolysis at 4°C/min indicate that 3.0
optimum mechanical and physical properties
are unobtainable unless high nitrogen contents
are retained within the precursor until the later
stages of carbonization (Deurberque, 1990;
Deurberque and Oberlin, 1991). Therefore, a

,--i
/"\ ...
large nitrogen content (large N I C ratio) should
be present when local molecular ordering I
I \
300=z CII

(LMO) begins and the carbon skeleton is being I ~

I
formed. Since the N/C atomic ratio depends I
I
inversely upon the H/C ratio, LMO should I.S J
I '\ 200
occur at large N/C and small H/C ratios.
I
Essentially, small amounts of aromatic hydro- J
I
gen and a relatively large amount of nitrogen 1.0 I
present during the LMO stage allow the carbon
skeleton to remain flexible enough at high tem- 2000 3000
1000
peratures that molecular rearrangement is TNT (·C)
easy. Within this overall fibrous texture, the
nanotextural features of the carbonized fibers Fig. 9.10 Effect of heat treatment temperature on
are the consequence of the variations in the tensile strength of carbon fibers (Bright and
cyclization, stabilization, carbonization and Singer, 1979).
graphitization conditions. If the original pre-
cursor is CH- and NH-rich but oxygen-poor, authors argue that if adequate nitrogen exists
the corresponding carbonized fibers will have within the fiber after LMO occurs, then a
low porosity, high compactibility and stacking greater potential for bonding exists and an
order and a relatively high strength. Likewise, improved rather than a diminished strength
if two nitrogen atoms are present in two aro- will result. This conclusion is based on the
matic rings contained within adjacent sheets, observation that, in contrast to other fibers, the
they are able to promote bonding of contacting strength of (nitrogen containing) commercially
BSUs together with a N2 release (Watt, 1972). available Toray T-300 fibers increased from 2.2
The ultimate strength value of the fibers to 3.2 GPa after heat treatment to 2800°C.
increases as the compactibility and the avail- Tension also increases cyclization and nitrogen
ability of 'efficient' nitrogen (Le. the nitrogen elimination which increases the tensile
remaining at the moment of LMO occurrence) strength of the final carbon fiber (Watt, 1972).
increases (Oberlin and Guigon, 1988; Guigon, Exactly what influence nitrogen has on the
1985). Bright and Singer (1979) agreed with development of high tensile strength is still
others when they found that the tensile being debated. However, the texture of the
strength of most heat treated fibers tends to graphitic layers, the amount of nitrogen pre-
decrease with higher temperature exposures sent in the original precursor and the
and release of nitrogen (Fig. 9.10). The same temperature at which misshapen layers touch-
 Microstructural changes during carbonization 183

and hence are available to bond and emit nitro-

gen - are obviously important.
Up to 1000°C, carbonization leads to an
effluent loss and increasing aromatization. As a
result, the solid residue transforms from a vis-
coelastic into a brittle solid material. The
stabilized PAN normally carbonizes into a sta-
tistically isotropic but nanoporous carbon
material which, because of the small dimen-
sions of the initial LMOs, is inherently
non-graphitizable (Joseph and Oberlin, 1983a).
Continuing pyrolysis up to 1500°C eliminates
most of the residual nitrogen and completes the
conversion of the PAN molecules into sheets of
carbon that are appreciably anisotropic (Mair
and Mansfield, 1987). Continued heating elimi-
nates the remaining nitrogen and, since the
material is then only made of pure carbon, fur-
ther modifications are only structural.
Graphitization is the name of the process
that involves heating the carbonized fiber to
approximately 2500°C in times as short as a
minute (US Patent 4005 183, 1977). Graphitized
pitch fibers exhibit a larger, more graphitic and
better oriented crystal structure than
PAN-based carbon fibers which are inherently
non-graphitizable. Parallel to the fiber axes,
pitch fibers have higher stiffness and thermal
conductivity values and a reduced thermal
expansion coefficient. These changes due to Fig. 9.11 Model of crumpled sheet-like structure
graphitization do not produce any significant (Guigon, Oberlin and Desarmot, 1984a).
increase in relative strength values. As a result
of the extreme temperatures required to temperatures, for the bonding of adjacent
process them, graphitized pitch-based carbon sheets. These sheets contain numerous vacancy
fibers are more expensive and are fabricated imperfections and are folded to enclose pen-
for specialized applications. cil-shaped voids oriented in the general
direction of the fiber axis. The lengths of each
block or sheet are relatively short, with each
9.7 MICROSTRUCTURAL CHANGES DURING
succeeding block misoriented with respect to its
CARBONIZATION
neighbor. A schematic illustration of the
The initial heating of stabilized PAN fibers microstructure as it exists within the actual
causes growth of graphite-like ribbons by a fiber is shown in Fig. 9.12. This structure is typ-
dehydrogenation mechanism. Denitrogenation, ically exhibited by high strength PAN-based
which occurs as the temperature is increased, is carbon fibers.
responsible for the growth in area and the Further temperature increases tend to
transformation of these ribbons into thin sheet- decrease the void space by joining sequentially
like structures (Fig. 9.11) and, at higher oriented and touching graphite like layers and
184 Carbon fibers

J4........l&----+-r.ex t e r nal
sur face
t
fibre ax i s

Fig. 9.12 Model of microstructure within a high

strength PAN-based carbon fiber (Guigon, Oberlin
and Desarmot, 1984a).

aligning them more parallel to the fiber axis.

The distorted sheets of BSUs associated with external
tilt and twist boundaries are bonded to each ~~----~~~~~r---surface
other wherever the boundaries of adjacent
sheets touch. The more compact the fiber, the
larger the number of contact areas and the Fig. 9.13 Model of PAN-based high modulus car-
greater the chance for adjacent sheets to bond. bon fiber (Guigon, Oberlin and Desarmot, 1984b).
The lateral cohesion thus formed causes the
strength of the fiber to increase. Specifically,
the distortions within the polyaromatic centrations and, hence, weaker fibers
graphene layers or sheets tend to be removed (Reynolds and Sharp, 1974; Reynolds and
by accumulating any structural defects at their Moreton, 1980).
boundaries. This induces a progressive
increase of the width and the radius of curva-
9.8 ELECTRICAL AND THERMAL
ture of the aromatic layers which can be
PROPERTIES
correlated with the stiffness, stacking order
and the diameter of the graphitic layers Studies of microstructural features have been
(Oberlin, 1984). A schematic of this microstruc- carried out using techniques that include:
ture is shown in Fig. 9.13. The longer, better X-ray and electron diffraction, electron spin
oriented and more graphitic microstructures resonance, thermoelectric power, magneto
exhibit both higher values of modulus and resistance, lattice fringe imaging, etc. All
thermal and electrical conductivities; unfortu- results indicated that carbon fibers are com-
nately, the misalignment of the larger posed of turbostratic layers of graphite
microstructural units causes large stress con- oriented preferentially at some angle to the
 Electrical and thermal properties 185

fiber axis. Increasing the heat treatment tem- PAN-based fibers is the lower maximum val-
perature results in a reduction of the ues of the elastic modulus and the electrical
interlayer spacing, a decrease in void space, a and thermal conductivities. Typical results
growth in thickness and area of the graphitic which compare the effect of heat treatment on
crystallites and an increase in the preferred the electrical and thermal conductivities are
orientation of the microstructure. All of these shown in Figs. 9.15 and 9.16.
changes increase the elastic modulus and the
electrical and thermal conductance. A corre-
sponding reduction of the tensile strength 1~nM-------------------------'
also occurs by mechanisms that depend on
local defects as discussed in the previous
section.
A comparison of the g-value anisotropy of
pitch and PAN in Fig. 9.14 indicates that the
degree of anisotropy changes for both fibers
after heating to about 1700°C. Although pitch E
based fibers become more anisotropic when ....c:a.
the temperature is increased further, the
anisotropicity of PAN seems to saturate at a ~
>
level which is comparable to that of a pitch i=
CI)
fiber heated only to about 2000°C. A simple Ci5
W
consequence of this inability to fully graphitize 0::

140

-
.; 120 ----------------
Single Crystal Graphite

1000 1500 2000 2500 3000 3500

HEAT TREATMENT TEMPERATURE COC)

Fig. 9.15 Schematic variation of the room tempera-

ture electrical resistivity against THT for: (0)
ex-rayon; (e) hot stretched rayon; (.6. and V) ex-PAN
fibers (Robson et al. 1972); (0) Ex-pitch fibers (Bright
and Singer, 1979) and (solid curve) benzene-derived
fibers (ehieu et al. 1983). The scatter of typical data
points about the mean give an indication of the
2600 3000
uncertainty. The dashed line indicates the decrease
Heat Treatment Temperature, deg C
in resistivity produced by hot stretching the
ex-rayon fibers.
Fig. 9.14 Variation in g-value anisotropy of
pitch-base and PAN-base carbon fibers as a func-
tion of heat-treatment temperature (HTT)
(Aggarival, 1977).
186 Carbon fibers

1011 300
r---~--~--~r---~--~~-

Pitch
,. x.... x ----
,
,,
I

,
.
'l'
51
VI
::t: I
/
/
0
::E
,•
I

,I
.i:
-go. 1
1!
B
I
•
I

Fig. 9.16 Condu~vity of pitch-base, PAN-base and rayon-base carbon fibers as a function of the tensile
modulus of elastiClty. (Courtesy of R. Gray, NSWC, Dahlgren, Virginia.)

9.9 MECHANICAL PROPERTIES OF FIBERS observation, have used electron diffraction

The microstructural changes discussed above (Guigon, 1985) and lattice fringe imaging
have been deduced using X-ray diffraction (Oberlin and Guigon, 1988). This latter tech-
nique is particularly powerful, because in
techniques. In addition, the mean length of
dark field, the parallel orientations (with
the graphite sheets oriented in the fiber direc-
respect to the imaging beam) of the convo-
tion La and their thickness L may be
luted graphite layers can be imaged. Sketches
computed using such techniques. It has been
found that both of these parameters increase of the possible layer convolutions, the images
they can produce and the measurements that
with increasing temperature. The orientation
can be made are shown in Fig. 9.18. Most
of these layers also becomes increasingly
important is the realization that it may be
aligned with the fiber axis. The net effect is to
increase the tensile modulus continuously as possible to classify the properties of commer-
shown in Fig. 9.17. Conversely, the tensile cial samples in terms of decreasing either
strength tends to decrease (Fig. 9.10). transverse (rt ) or longitudinal (rt ) radius of
More recent studies, which allow direct curvature of the sheets. Such a classification,
based on these and other numerical values
 Mechanical properties of fibers 187

700 shown in Table 9.2, could form the basis of

,- ",""
Pan empirical relationships between micro-tex-
/
ture and mechanical properties.

/r
600 At increasingly higher heat treatment tem-
/ peratures, the scattering domains within high
1500 modulus fibers become large and well
defined so that the length of the graphitic
-"
Q.
400
//
/ / sheets in the fiber direction, Lan Cr can be mea-
sured directly (see Table 9.2) from the
CJ
§/ /
>-
III / .' observed Moire fringes. Correspondingly, the
Pitch / I radii rt and r1 of the sheets are also measur-
300 / I able from 002 lattice fringes. The lateral
( I
I cohesion of the fiber is also ensured by bond-
200
/
i ing between adjacent distorted sheets of
carbon wherever two grain boundaries are in
/' contact. The chances of such bonding increase
100 ./ as rt and r1 decrease, but decrease as La
1000 2000 3000 3500 increases. Hence the extent of lateral cohesion
THT t·e) can be defined by a variable S = La [(lIr t )
(lIrt + 1/r1 )]. As shown in Fig. 9.19, a linear
correlation has been observed between O'c and
Fig. 9.17 Effect of heat treatment tetnperature on
the elastic modulus of PAN and pitch-based carbon l/S indicating that long, relatively unbonded
fibers. (Data from Johnson, 1969 and Aggarival, graphitic layers result in weak fibers.
1977). Conversely, Young's modulus and electrical

2 3

b b
II I ..
2r!!!!'·
I
n
I "·1111 11111 II n

~ ~

d ,·11111111111111111111111111111 d '·11 111111 1IIIIilili 111111111 d '·1 ~

Fig. 9.18 Sketches of the possible dark-field images of a longitudinal section (lamellar model of Fig. 9.15)
as r decreases and the fold develops (Guigon, Oberlin and Desannot, 1984b).
188 Carbon fibers

Table 9.2 Quantitative measurements which are suggested for use to classify rnicrotexture and mechani-
cal properties

SAD patterns Half width of 004 reflections

Half width of 11[0] ring
002DF Thickness of elementary bright domain [BSU]
Length of elementary bright domain
002LF Number of fringes in a stack
Length of a perfect fringe
Length of a distorted fringe
11DF Length of turbostratic Moire fringes
Diameter of a domain showing
rotational Moire fringes
ODP of002LF f3 Arc opening
LlDOO2 Interfringe spacing spreading

a'c GPa
,,
\
\
• HTS (HTT 28OO"C)
\
\ .HM (R.w)
\ \
\
3 \
\
\
\
\

.,\
\
\
\

\
\
2

115
o 0.1 0.2

Fig. 9.19 Numerical relations between tensile strength and the rnicrotexture G c = /(1/5). High modulus
fibers (full line). High tensile strength fibers heat treated at 2800°C (dashed line) (Oberlin and Guigon,
1988).

conductivity correlate well with La' as seen in According to the pioneering work of
Fig. 9.20. Griffith (1920), the following expression
describes the strength, OF' of a brittle solid,
containing a crack of length 2a,
9.9.1 MICROSTRUCTURAL CONSIDERATIONS
0F=..J(Ey/4a) (9.1)
Carbon is a brittle solid; thus, since no plastic
deformation can occur, very high local stresses where E is the elastic modulus and y is the sur-
will develop at stress raisers, such as disconti- face energy. Inspection of equation (9.1)
nuities, changes of section size, cracks, etc. indicates that longer cracks are more effective
 Mechanical properties of fibers 189

G Pa E

(a)
400

300

T
o 2 3 4 5
J,Iilm p

10

(b)
5

o 1 2 3 4

Fig. 9.20 Numerical relations between the microtexture and (a) Young's modulus Em and (b) electrical
resistivity p (Guigon, 1985).

as stress raisers than short cracks. Also, since 9.9,2 BUNDLE THEORY
longer cracks generate higher tip stresses, it
In a composite, many fibers are arranged more
can be inferred that once a crack begins to or less parallel to one another and functio~ as
move it will continue to move (accelerate) a load bearing bundle. A number of pubhca-
until it reaches the geometric boundaries of tions have appeared (Herring, 1966; Wright
the material. As a consequence of this, the fail- and Iannuzzi, 1973; Wright and Wills, 1974)
ure of brittle solids is abrupt and depends on that discuss the distribution of strengths
the probability that a crack of some critical exhibited by brittle fibers and how these dis-
length is present. If such a theory can be tributions can be used to compute a mean
applied to carbon fibers, then it can be a~~ed strength, O'm and a corresponding bundle
that the maximum length of cracks (or sundar
strength,O'b ,
microscopic stress raisers) that can be con- For example, the above authors argued that
tained is limited by the fiber diameter; so their individual fiber strength data tended to
small diameter fibers will be stronger than
obey a Weibull distribution characterized by
large diameter fibers. In addition, since the
the expression,
chance of a crack being present is greater, long
fibers will tend to be weaker than short fibers. G(O') = 1- exp{- a(O'/O')ro} (9.2)
where G(O') is the probability of failure of a
fiber subjected to stress 0', 0'0 is the distribution
190 Carbon fibers

scale factor, 00 is the distribution shape factor tion of the same fiber by a shear process. The
and a is a function of the length/ diameter length of matrix required to do this defines a
(Lj d) ratio of the fiber (Corten, 1967). bundle of short fibers (or segment of compos-
If many fibers are tested of different length ite) which must break in order to break the
I, then a graphical method can be used to composite. Such bundles can be modeled as
deduce 00, a and a o and, the mean strength and analogous to a link within a chain; failure of
the strength of a bundle of fibers can be com- the weakest link defines the failure load of the
puted from composite. Nevertheless, since shorter fiber
bundles are stronger than longer, shorter links
O'm= O'o(l/d)-I/Olr(l + 1/00), (9.3)
are stronger than longer links.
where r is the gamma function and The link length, 8, has been discussed by
Rosen (1964) and, for purposes of this discus-
O'b=O'o(aroe)-l!Ol. (9.4)
sion Can be approximated by,
Bundles of twisted fibers would exhibit lower
8 = O'd/2't (9.5)
strengths.
where 0' = O'b the stress in the fibers at failure of
the composite, d is the diameter of the fiber
9.10 COMPOSITES FABRICATED FROM
and 't is the shear strength of the matrix or
CARBON FIBERS
matrix-fiber bond as it exists in the composite.
Carbon fibers are very strong, stiff and light- For a given fiber strength distribution, the
weight materials. In addition, their small stronger composites will all exhibit smaller
diameter (8-12 pm) makes them extremely ineffective (link) lengths. This is accomplished
flexible. Unfortunately, they exhibit little com- by using small diameter fibers, well bonded
pressive strength and they exhibit a poor using high strength glue.
abrasion resistance. A solution to these prob-
lems, and to the problem of brittleness, is to
9.10.1 SURFACE TREATMENT
bond large numbers of fibers together to form
a composite solid. In this case, the glue or A freshly prepared fiber does not bond well to
bonding agent forms a continuous phase that a polymeric glue (or to anything else for that
is usually defined as the matrix. The matter); however, the tendency to bond can be
matrix-fiber mixture is called a composite significantly increased by subjecting the fiber
material. The function of the matrix is to sup- surface to a controlled oxidation. As discussed
port and separate the fibers, to protect them by Riggs, Shuford and Lewis (1982) in the ear-
from reaction with the environment and to lier edition of this handbook, this treatment
transfer load. In a composite, the tensile and essentially etches the surface, cleans it,
compression properties parallel to the fibers increases its surface area and produces polar
are much better than those measured on bun- hydrophilic oxygen-containing groups which
dles. The transverse properties are also bond to it. The process can be carried out in a
optimized, since the matrix serves to improve liquid or gaseous environment; for example,
the fiber-matrix connectivity. This function is heating in air or oxygen-nitrogen mixtures,
important since it affects the mechanical and CO2, CI, N02-NO, NH3 and plasma-ionized
thermal properties in the transverse direction. inert gases (Johnson 1969; US Patent 3 754957,
Fortunately, although the strength of individ- 1973a; US Patent 3 723 ISO, 1973b; British
ual fibers exhibits a pronounced size effect, no Patent 1 341161, 1973; U.S Patent 4374114,
size effect is exhibited by composites. In effect, 1983; US Patent 3 627 466, 1971; US Patent
load is transferred around fiber breaks into 3767774, 1973; US Patent 3780255, 1973).
adjacent fibers and back into the unbroken sec- Direct wet chemical oxidation has been tried
 Mechanical properties of unidirectional composites 191

using aqueous nitric acid, hypochlorite, chlo- crack by interaction with low strength
rate and dichromate in sulfuric acid. fiber-matrix bonds; strengthening depends on
Treatments have also been investigated using load being transferred from the matrix to the
electrolytes of hypochlorite, ammonium fibers through a strong fiber-matrix bond.
hydroxide, sodium hydroxide and ammonium
sulfate (US Patent 3660140, 1972; US Patent
3746506, 1973; US Patent 3894884, 1975a; US 9.11 MECHANICAL PROPERTIES OF
Patent 3 859 187, 1975b; US Patent 3746450, UNIDIRECTIONAL COMPOSITES
1973; US Patent 3 989 802, 1976; US Patent
The properties of unidirectionally reinforced
3832297, 1974; US Patent 3671422, 1972;
composites are strongly orthotropic.
British Patent 1371621,1974; British Patent
Specifically, properties measured parallel to the
2071 702, 1981) .
fibers are quite different from those measured
During oxidation, the strong carbon-oxy-
at right angles to them. More importantly per-
gen complexes which are formed bond
haps is the sensitivity of property
tenaciously to the fiber surface and will subse-
determination with respect to the direction of
quently react with a matrix resin. In order to
the measurement. For example, Fig. 9.21 illus-
preserve this reactivity a thin layer of the final
trates the change in elastic properties exhibited
matrix resin is applied to the surface of the
when a unidirectionally reinforced composite
fibers as a finish or size. This layer does dou-
is loaded at some angle to the fiber axis. In this
ble duty in both protecting the fiber surface
case, it can be observed that the stiffnesses of
against damage during transportation, further
the composite (Qu' etc.) begin to change signif-
processing and handling and in promoting
icantly when the load is misaligned only fifteen
wetting when the sized fibers are bonded
degrees to the fiber axis.
together with the matrix resin.
Poor bonding is a sensitive function of the
surface morphology, anisotropicity, hetero-
9.11.1 MICROMECHANICS
geneity and the nature of the interphase layer
between the fiber and the matrix. For example, Using micromechanics, various equations can
it has been found that the greater the degree of be used to estimate the properties that might
graphitization and the better the alignment of be exhibited by well bonded composites. For
the microstructure with respect to the fiber example, parallel to the fibers, the modulus
axis, the poorer the fiber will bond. (E 1), strength (0) and Poisson's ratio U12 are
Essentially, the higher modulus carbon fibers given by;
will not bond easily in the absence of a surface
treatment. In some composites where the fail-
E1 =Vf Ef + Vm Em (9.6)
ure strain of the matrix is smaller than the 0"1=Vf O"f+Vm O"m (9.7)
failure strain of the fibers (as it is for ceramics
U 12 = Vf U12 + VmUm (9.8)
or carbon), poor bonding is an asset since the
largest fiber-matrix bond strength is not 0"2 == O"m (9.9)
required. Conversely, if the failure strain of the
matrix is larger than that of the fibers, a strong where V is the volume fraction of fibers (f) or
bond is desired. The reason for this apparent matrix (m) respectively, O"f is the strength of a
dichotomy involves the fact that fibers are bundle of fibers with length equal to the inef-
added to brittle matrices primarily to toughen fective length (usually, due to lack of statistical
them; only tough matrices (large failure strain) data O"f is taken as the mean strength supplied
can be strengthened. The mechanism of tough- by the fiber processor). Also,
ening depends on the blunting of a running
192 Carbon fibers

E2=(EfEm)/(VfE~+VmEf) (9.10) mechanics of composites; thus, the reader is

directed to these for a more complete discus-
= (GI2f Gm)/(Vf Gm+ VmGl2f) (9.11)
G12 sion of this subject (Ashton, Halpin and Petit,
where E~ = Em! (l - Urn) and Gl2 is the shear 1969; Jones, 1975; Tsai and Hahn, 1980; Daniel
modulus. and Ishai, 1994).
When only plane stress conditions exist,
(e.g. 0'3= 0 and E3 is related to El and E2), then it
9.11.2 MACROMECHANICS OF LAMINAE is possible to relate stress to strain along the
There are a number of excellent books and principal axes of an orthotropic lamina,i.e. par-
monographs that have been written on macro- allel (1) and perpendicular (2) to the fiber axis:

200 n GPo .." x 200

"'II I ~

~
/9
I ~I
150 150

100 100

50 50

-+----~--~--~--·9 O~--~30~--~60~--90~--tJ
o 30 60 90

60 60

9 0 30 60
tJ
90

q6 60 °26

40
20
9 IJ
90

Fig. 9.21 Transformed, off-axis modulus of T300/5208. The angle is the ply orientation and is positive for
counterclockwise rotation (Tsai and Hahn, 1980).
 Mechanical properties of unidirectional composites 193

the components of the [Q] has already been

cr1 Qn Q12 0 El illustrated in Fig. 9.21. xy
cr2 = Q12 Q22 0 E2 (9.12)
t12 0 0 Q66 'Y2 Ex Sxx Sxy Sxs crx
E = S S S cry (9.15)
or y 5YY 5YY 5YS
'Ys sx sy 55 ty
E1 5n 512 0 cr1
= 512 522 0 (9.13) In addition to the variation of elastic proper-
Ez cr2
ties, the strength of unidirectionally reinforced
'Y2 0 0 566 t12
composites has been found to be sensitively
where
dependent on the angle of loading. A typical
511 = lIEl failure curve for tension and compression is
522 = 1/E2 shown in Fig. 9.22. This type of curve can be
512 = -'\)121EJ = -'\)2/ E2 obtained by transforming the applied stresses
566 = 1/G12 to directions parallel and perpendicular to the
fiber axes and then equating those stresses to
and
the failure strengths actually measured (or
Qn = E/(I-U12'\)21) computed using micromechanics) along those
Q22= E/(1-U12'\)21) directions. Failure of a composite is then con-
Q12= U21E/(1-U12'\)21) sidered to occur when the transformed stress
= '\)12E/(I-'\)12U21) exceeds the failure stress actually measured in
Q66 = Gl2 that direction, i.e. when
Inspection of equations (9.12) and (9.13) indi- Fxle = Fl/ (cosze); F = F2t I (sin2 9);
xl
cates that parallel and perpendicular to the F = F/ (sin9 cos9)
xl
fiber axes, tensile (compressive) stress produces
and
tensile (compressive) strain and shear stress
produces shear strain. There is no coupling Fxte = Fl/(cos2 9); Fxe = F/(sin I} cos 9)
between shear stress and tensile (compressive)
where F11(e)' F2t(e) and F6 are the measured failure
strain, e.g. Q16 =0 = Q26 as shown previously in
tensile (t) or compressive (c) loads measured
Fig. 9.21 and the shear coupling coefficient
parallel (1) or perpendicular (2) to the fiber axis
which is the ratio of tensile (compressive) stress
and F6 is the measured shear strength. These
to shear strain, is zero. The above conclusion is
failure criteria are collectively described as the
not valid when loads are applied at some angle
maximum stress failure criterion.
to the principal axes since appreciable shear
There are other failure criteria such as max-
coupling can occur; Q16 and Q26 are not zero.
imum strain, deviatorial strain energy
This means that tensile (compressive) stresses
(Tsai-Hill) and Interactive Tensor Polynomial
produce shear strains in addition to the more
(Tsai-Wu) that can be used, some of which
normal tensile (compressive strains). The
allow for interaction between the stresses.
stress-strain relationships are then written,
However, for typical composite structures, the
crx Qxx Qxy Qxs Ex maximum stress criterion give reasonably con-
cry = Qyy Qyy Qys Ey (9.14) servative estimates of expected failure
t. Qsx Qsy Q.. 'Ys stresses.
where x and y are the orthogonal axes of the
composite test specimen that are oriented at
9.11.3 MACROMECHANICS OF LAMINATES
some angle to the fiber axes and where [Q] or
[5]xy are the transformed matrixes of equati6ns In order to eliminate coupling and to reduce the
(9.12) and (9.13). The variation with angle of very strong change of medtanical properties
194 Carbon fibers

Fiber failure 160

120
~
(.!) o.
>c
lL 0.4

~.... 0.2i~~~~~~~iI~8i~~~~ii~
lB 0
Ci5 -0.2
-0.4 compressive
failure -80

10 20 30 40 50 60 70 80 90
Fiber orientation, e, deg.
Fig. 9.22 Uniaxial strength of off-axis E-glass/epoxy unidirectional lamina as a function of fiber orienta-
tion (Daniel and Ishai, 1994).

with the direction of applied load, composites Of particular interest is the quasi-isotropic
are usually fabricated from multiple layers, composite which is often specified for com-
each arranged at some angle to their neighbor. mercial structures. The in-plane engineering
The angles that are used can be arranged to elastic constants of these materials are identi-
bring the fiber axes at some optimum angle to cal in all directions and there is no shear
the expected loads; however, there is a require- coupling. Examples of such composites are
ment for each lamina to be oriented very symmetric arrangements of [0/60/-60] or
precisely with respect to one another in order [0/±4S/90] layers.
to avoid tensile-shear effects and coupling
between in-plane loading and out-of-plane
9.12 TESTING TECHNIQUES
deformation (tensile loads can be made to pro-
duce bending and twisting deformation, for A knowledge of the behavior of the constituent
example). Various rules of angle-ply materials phases allows the mechanical properties of the
have been worked out in order to avoid the resulting composites to be computed. These
cross-coupling terms. For example, symmetric calculations are made with the use of expres-
laminates exhibit no coupling between sions (9.6) to (9.11). The results of the following
in-plane loading and out-of-plane deforma- tests should provide input to and in some cases
tion. (Symmetric laminates define a composite confirm these calculations. The following was
in which for every lamina oriented at some adapted from the excellent (and much more
angle there is another layer of identical thick- extensive) discussion contained in Chapter 8 of
ness and orientation placed at an equal the book Engineering Mechanics of Composite
distance from the mid-plane of the composite.) Materials (Daniel and O. Ishai, 1994).
And, composites which exhibit no shear cou-
pling are 'balanced laminates' (pairs of
9.12.1 PROPERTIES OF FIBERS AND MATRICES
identical layers oriented at a positive and an
equal negative angle with respect to the lami- Determination of the elastic and failure prop-
nate reference axes). erties of the fibers is described in ASTM
 Testing techniques 195

specification 03379. The difficulty is mostly verse Young's modulus, tensile strengths and
involved in determining the elastic displace- strains and the major and minor Poisson's
ment parallel to the fiber axis, since no ratio can all be determined from this type of
measurement device can be attached directly coupon test specimen. Similar properties can
to the fragile fiber. Since the fiber tends to shat- also be measured in compression; usually
ter and disintegrate when failure occurs, the however, short, thick specimens are used in
mean diameter measured before the test has to order to avoid buckling failure. ASTM 0-3410
be used to calculate the failure stress. describes this test method.
The method involves attaching the fiber
across a slot cut into a paper tab. The compos-
Flexure testing
ite specimen is aligned coincident with the
load axis of the testing machine, the tab ends A far more expensive test is described in ASTM
are gripped and the sides of the tab are cut to C393. This requires a rather large
allow only the fiber to be loaded by the (22 in/1 in/l.5 in) sandwich flexure specimen
machine. The measured compliance Cm is the which is tested in four point bending. The
sum of the compliance of the loading system dimension of the honeycomb core and the
C, and the compliance of the fiber Cf' Thus composite face sheets are adjusted to cause fail-
ure in the approximate face sheet. Good results
Cm = C,+ Cf = C1 + , (AE f) can be obtained in both tension and compres-
where 1 is the fiber length and Ef is the fiber sion; however, since failure might occur in the
modulus. A plot of measured compliance core or in either of the outer skins or at the
against fiber length allows calculation of the skin/ core interlace, care should be exercised in
loading system compliance and the fiber both determining the exact failure mode and in
modulus. reporting the appropriate failure stresses.
Polymer matrices are evaluated using In the above paragraphs the reader has
coupons cut from thin sheets. Typical geome- been cautioned about making sure that the
tries are described in ASTM specifications loading axis of the testing machine and the
0638, 0638 and 0882. symmetry axis of the specimen are coincident.
Indeed, consistent measurements will only be
achieved by eliminating any tendency to
9.12.2 PROPERTIES OF COMPOSITES develop complex stresses in any region of the
specimen. An additional effect not significant
Coupon tests
in designing or testing ductile materials is the
The determination of the tensile longitudinal tendency for composites to exhibit poor shear
and transverse properties of unidirectionally properties. Indeed it is common for engineers
reinforced composites can be obtained by test- designing metallic components to assume that
ing relatively long coupon specimens. These failure in shear will not occur if failure in ten-
specimens, as described in ASTM specifica- sion or compression is designed against.
tions 03039-76, are 9 in long, 0.5 in wide and Unfortunately, shear resistance of composites
from 0.02 to 0.10 in thick. The apparently is not directly related to tensile or compressive
excessive length requirement is an attempt to properties; thus, a separate shear failure crite-
minimize the effect of specimen misalignment rion must be used when designing with
with respect to the loading axis. Glass/epoxy composites. Shear failure can occur in-plane or
tabs are bonded to the specimen ends to shear interlamina.
the load into the specimen and to avoid dam-
age and failure of the specimen within the
gripped length. The longitudinal and trans-
196 Carbon fibers

In-plane shear testing D3846. This specimen contains two square

notches, each cut into an opposite face of a ten-
A reasonably simple test method which utilizes
sile or compressive specimen. If the slots are
a coupon with an eight layer symmetric ± 45
cut a reasonable distance, 1, apart in order to
deg. layup is described in AS1M standard D35
avoid interactions between the stress fields of
18. If the specimen has strain gages oriented
the cut slots, then the interlaminar shear
parallel to the x and y axes of the specimen
strength is given by
(not the fibers), then the failure stress in shear
is given by (o)max/2, the shear failure strain is F31 = P/wl
(£x - £y)max and the shear modulus is given by
G12=E/2(£x - £y)
A rail shear test has been described in 9.13 CONCLUSION AND PERSPECTIVES
AS1M D4255-83. The shear stress and shear
Low performance isotropic non-continuous
strain can be obtained at intermediate and
pitch-based fibers and anisotropic PAN-based
maximum loads. Thus, G12 and the appropri-
large tow fibers have been available for a num-
ate failure parameters can be obtained.
ber of years in the USA priced at $20/kg.
These fibers have been used to produce elec-
trically conducting flexible heating materials
Interlaminar shear strength
to control the temperature of pipelines, to heat
The interlaminar shear strength is a measure human dwellings, hot-houses, etc. (Karpinos
of the strength of the bond that exists between and Izmalkov, 1982; Glushchenko and Griffen,
the various layers within the composite. It is 1982; Levit, 1986; Jakubowski and
important to know such a value since bending Subramanian, 1979). In addition, low modulus
of beams can cause appreciable shear stresses, carbon fiber materials and their composites
which while not large enough to cause failure have been used as abrasive, anti-friction, seal-
of a traditional metallic structural material can ing and heat-insulating materials. Other
fail a composite. This occurs either within the nonstructural uses include activated carbon
matrix or at the fiber-matrix bond line. The fibers which can be used directly for the purl-
test specimen, as discussed in AS1M D2344, is fication of nitrogen oxide-containing fumes,
a rather short, thick beam which is tested in chimney smoke, automobile air conditioners,
three-point bending. Basically, the test speci- various respiratory, water purification
men is sized such that the ratio of the shear (Chupolov et al. 1983; Richter, Knoblauch and
stress generated at the midplane is maximized Juntgen, 1984). In recent years carbon fiber
with respect to the tensile or compressive adsorbents have been used in medicine to
stresses generated in the outer fibers. If inspec- remove toxic substances from body fluids
tion of the failed specimen indicates that (Temovoi et al., 1985).
failure by shear has occurred, then beam the- The high modulus/high strength continu-
ory indicates that a reasonably accurate ous carbon fibers have been available at a cost
estimate of the interlaminar shear strength F31 of $60--80/kg. Thus, reinforced epoxies, poly-
can be calculated from esters and other polymers have demonstrated
properties that have enabled them to find use
F31=3P/4wh
in the aerospace and aeronautical fields.
where P is the maximum load and w and hare Smaller quantities of fibers have been used in
the width and depth of the beam. high quality sports gear such as golf clubs,
Another way to measure the same maxi- fishing rods, tennis rackets, marine sports, ski
mum shear stress is by using the equipment, bicycle equipment, etc. Fibers
double-notched specimen as described AS1M have also been used in many civil engineering
 Conclusion 197

structures. These include short fiber reinforce- can be made with modulus values much larger
ment of cement mortar and continuous than can be obtained from PAN precursors.
reinforcement of concrete. Some attempts have Unfortunately, rather sophisticated and expen-
been used to provide earthquake resistant sive pitch pretreaments must be applied to a
structures and fiber reinforced ropes and some petroleum or coal tar pitch in order to produce
successes have been reported in fabricating a high performance fiber. The pitch softening
structures with carbon fiber reinforced alu- temperature is much higher than PANi thus
minum and copper. Carbon fiber reinforced stabilization can potentially be carried out
carbon has been considered for high tempera- quickly at higher temperatures. Nevertheless,
ture load bearing structures and heat shields stabilization of both PAN and pitch materials
in spacecraft and supersonic aircraft. And, is exothermic and, in order to avoid overheat-
finally, similar materials are considered useful ing, thermal runaway and decomposition of
in nuclear applications and disc brake materi- the precursor, a less than advantageous tem-
als for aircraft, high speed trains and racing perature of oxidation must be used. Future
cars. At the present time, there is a growing developments in these areas should therefore
interest in the use of very high thermal con- involve methods to increase the stabilization
duction properties to manage local rate and the development of new precursor
temperatures in sensitive electronic equip- materials. Specific topics might include stabi-
ment. Nevertheless, despite all these apparent lization in thermally stable environments
successes, it is vitally important to realize that (fluidized beds, liquids, etc.) and the develop-
the expanded use of high performance carbon ment of alternative synthetic precursor
fibers depends very sensitively on the lifetime materials (polymers and/or pitches).
costs involved in substituting carbon and its Cost of the final component, while very sen-
composites for competing metals, ceramics sitive to the cost of raw materials, also
and polymers. involves all of the design, fabrication and test-
It has been recognized for many years that ing costs. All of these component costs must be
market penetration of carbon fibers will tightly controlled if economically viable per-
always be limited to rather sophisticated formance increases are to be realized. When
structures if the cost remains at the present using composites, it has been found to be
high leveL However, the precursor PAN mate- vitally important to pay strict attention to
rial presently costs about $5/kgi thus, taking detail design if a maximum expected weight
into account the weight loss and processing saving is to be realized. Designers must opti-
costs involved in converting PAN to carbon, it mize the total vehicle weight and not simply
is unlikely that large amounts of carbon fiber substitute a carbon composite for a metal one.
made from such material will ever be less than It is rumored for instance that, due to overly
$10/kg. Conversely, the cost of the pitch pre- conservative design (i.e., the use of metallic
cursor material is almost insignificant, since it joining and fabrication techniques etc.), the
is the byproduct of a commercial process resulting weight of some composite structures
established to produce other end products: has, in the past, turned out to be as heavy as
gasoline, metallurgical coke, etc. In addition, similar components built from aluminum.
since the fiber is fabricated using a melt spin-
ning process, the production rate can be much
9.14 CONCLUSION
faster than the wet or dry process used to pro-
duce PAN based fibers. The carbon yield from Many people appear to believe that despite
pitch precursors can average up to 85%, nearly thirty years of development, carbon
whereas the carbon yield produced from PAN fibers are still an evolving space age material.
averages about 65%. Pitch based carbon fibers Until very recently, there was a production
198 Carbon fibers

over-capacity in the carbon fiber industry. For At the present time, downsizing of the
example, in 1995 it was estimated that industry, increasing use of low cost fibers and
10 000 000 kg of carbon fibers were sold from the resurgence of orders for new commercial
an estimated capacity of 16 000 000 kg. Exactly aircraft that now use increased quantities of
how accurate these estimates were is difficult carbon fiber has brought industrial capacity
to assess; however, it can be concluded that the and market requirements closer together.
market is small and cannot accommodate Indeed, some fiber types are now difficult to
many producers. For this reason, many pro- obtain. Nevertheless, any major growth of the
ducers have seen fit to evaluate their position carbon fiber industry depends on the discov-
in the industry. For example, Table 9.3 is a list- ery of a method to produce fibers for one-half
ing of carbon fiber processors taken from or one-third of the present projected large vol-
Chapman and Hall's directory, 'Carbon and ume price and the development of new
High Performance Fibers' which was pub- inexpensive fabrication methods for struc-
lished in 1991. In 1996, only the first eight of tures. These developments will initiate major
these were still producing significant commer- new transportation based markets for the
cial quantities of PAN-based carbon fibers. material. At the present price, however, the
Table 9.4 is a similar listing for pitch-based use of carbon fibers will always be limited to
fibers. In this case only the first five appear to competitive performance driven applications.
be active.

Table 9.3 PAN-based tow manufacturers

Company Country Trade name

Akzo Carbon Fibers Inc. (Fortafil Fibers Inc.) USA Fortafil
Amoco Performance Products Inc. USA Thomel
RK Carbon Fibres Limited UK RK
Mitsubishi Rayon Co. Ltd. Japan Pyrofil
Soficar SA (Toray Industries Inc.) France Torayca F
Toho Rayon Co. Ltd Japan Besfight
Toray Industries Inc. Japan Torayca
Zoltek Corporation USA Panex
Akzo NV (Fibres and Polymers Division)(Enka AG) Holland Tenax
Anglo-Soviet Materials Ltd USSR Sapem
Asani Kasel Carbon Fiber Co. Ltd Japan Hi-Carbolon
BASF Structural Materials Inc.1 USA Cellon
Formosa Plastics" Taiwan ?
Hercules Advanced Materials and Systems Company2 USA Magnamite
Korea Steel Chemical Co. Ltd Korea Kosca
Nikkiso Co. Ltd" Japan ?
N.W. Chemical Power Co.* China ?
SigriGmbH Germany Sigrafil
Textron Specialty Materials USA Avcarb
* Not available.
1 Now Hexcel; 2 Now Toho.
 References 199

Table 9.4 Pitch-based tow manufacturers

Company Country Trade name

Amoco Performance Products Inc. USA Thomel
Mitsubishi Kasei Corporation Japan Dialead
Petoca Ltd (Kashima Oil Co. Ltd) Japan Carbonic
Tonen Corporation Japan Forca
Nippon Petrochemicals Co. Ltd Japan Granoc
Kawazaki Steel Co. Ltd Japan KMFC
Kobe Steel Ltd" Japan ?
Mitsubishi Oil Co. Ltd" Japan ?
Nippon Carbon Co. Ltd" Japan ?
Nippon Steel Co. Ltd" Japan ?
Osaka Gas Co. Ltd (Donac Ltd) Japan Donacarbo-F
Showa Shell Sekiyu Japan Carbonexel
• Not available.

REFERENCES Deurberque, A., Ph.D. 1990. Thesis, Universite de

Pau et des pays de l'adour.
Aggarival, RK. 1977, Carbon 15:291. . Deurberque, A. and Oberlin, A. 1991. Carbon 29: 621.
Ashton, J.E., Halpin J.e. and Petit, P.H. 1969. Pnmer Donnet, J.B. and Bansal, R 1984. Carbon Fibers, Vol.
on Composite Materials: Analysis, Westport, 3. New York: Marcel Dekker, Inc.
Conn.: Technomic Publishing Co. Edie, D.D. 1990. Carbon Fibers, Filaments and
Bacon, R 1973. Chemistry and Physics of Carbon 9:2. Composites. (eds. Figneiredo, J. Bernard, c.A.
New York: Marcel Dekker, Inc.
Baker. RT.K. and Hiittenger, K.J.) Kluwer
Bathia, G., Fitzer, E. and Kompalik, D. 1984. Academic, pp. 647-655.
International Carbon Conference, Bordeaux, France, Edie, D.o. and Diefendorf, RJ. 1993. Carbon Fiber
Paris: Group Francais d'etude des carbon. Manufacturing. Park Ridge, NJ: Noyes
Boncher, E.A., Cooper, RN. and Everett, D.H. 1970, Publications.
Carbon 8:597. Erlemenko, LN., Plyublines I. and Gulko, N.v. 1990,
Bright, A.A. and Singer, L.S. 1979. Carbon 17: 59.
Chemical Modified Carbon Fibers. New York: VCH
British Patent 1 341161, 1973. Publishers, Inc.
British Patent 1 371 621, 1974. White. Eshback, O.W. and Souders, M., Handbook of
British Patent 2 071 702,1981. Toho Beslon. Engineering Fundamentals, New York: John
Brooks, J.D. and Taylor, G.H. 1965. Nature 206:697. Wiley.
Brooks, J.D. and Taylor, G.H. 1968. Chem. Phys. Ezekiel, H.M. 1969. Appl. Polym. Symp. 9: 315.
Carbon 4:243. Fitzer, E., Muller, K. and Schaeffer, W. 1971.
Carbon and High Performance Fibers Directory, 1991. Chemistry and Physics of Carbon, pp. 237-383.
Chapman and Hall. New York: Marcel Dekker, Inc.
Chieu, T.e., Timp, G., Dresselhaus, M.S., Endo, M. Fitzer, E. and Heine, M. 1988. Fibre Reinforcements for
and Moore, A.w. 1983. Phys. Rev. B 27:3686. Composite Materials Vol. 2; pp. 73-148.
Chupalov, Y.S., Migunov, E.I., Panov, V.P. and Amsterdam: Elsevier. (ed. A. R. Bunsen).
Tereshchenko, L.E. 1983. Zh. Prikl. Khim. (USSR) French Patent 2 328 723. National Research
56: 2595-2597. Development Corp.
Corten, H.T. 1967. Modern Composite Materials (ed. French Patent 1 535 800, 1968. North American
Broughtman L.J. and Krock, RH.). Reading, Aviation, Inc.
MA: Addison-Wesley. Glushchenko, Y.M. and Griffen, L.A. 1982. Gibkie
Clarke, A.J. and Bailey, J.E. 1973. Nature 243: Elektroprovodnye Materialy I .
146-150.
Ustroistva na Osnove dlya Obogreva Lyudel I
Daniel, I.M. and Ishai, O. 1994. Engineering Tekhniki, pp. 23--29.
Mechanics of Composite Materials, Oxford Grassia, N. and McGuchan, R. 1971a. Eur. Polymer J.
University Press. 7:1091-1104.
200 Carbon fibers

Grassia, N. and McGuchan, R. 1971b. Eur. Polymer J. Carbon 29: 849.

7: 1357-1371. Lafdi, K, Bonnamy, S. and Oberlin, A 1991c.
Griffith, AA P. 1920. Trans. R Soc. A221:163. Mechanism of anisotropy occurrence in a pitch
Guigon, M. 1985. Relations entre la microtexture et les precursor of carbon fibres, Part III - Hot stage
proprietes mecaniques des fibres decarbone ex-PAN. microscopy of pitch B and e. Carbon 29: 857.
D.Sc. Thesis (These d'Etat), Universite de Lafdi, K, Bonnamy, S. and Oberlin, A 1992.
Technologie de Compiegne, France. Textures and structures in heterogeneous
Guigon, M. and Oberlin, A 1986a. Composites Sci. pitch-based carbon fibres (as-spun, oxidized,
Technol. 25: 231. carbonized and graphitized); comparison with
Guigon, M. and Oberlin, A 1986b. Composites Sci. homogeneous fibres. Carbon 31: 29.
Technol. 27: 1. Lafdi, K and Oberlin, A 1994a. A tentative to char-
Guigon, M., Oberlin, A and Desarmot, G. 1984a. acterize and elaborate anisotropic pitches and
Fibre Sci. Techno!. 20: 177. derived carbon fibres. Part I: preparation by
Guigon, M., Oberlin,A and Desarmot, G. 1984b. separation. Carbon 32: 11.
Fibre Sci. Techno!. 20: 55. Lafdi, K and Oberlin, A 1994b. A tentative to char-
Hadcock, R.N. 1982. Design and Analysis of acterize and elaborate anisotropic pitches and
Composite Structures. In Handbook of derived carbon fibres. Part II: preparation by
Composites. New York: Van Nostrand Reinhold bubbling. Carbon 32: 61.
Co. Lavin, J.G. 1992. Carbon 30: 351.
Hamada, T., Nishida, S., Matsumoto, Y. and Endo, Levit, R.M. 1986. Conducting Synthetic Fibers,
M. 1987. J. Mater. Res. 2: 850. Khimiya, Moscow.
Herring, H. W. 1966. NASA Rep. No. IND-3202. Lewis, I.e. 1982. Carbon 20: 519.
High Performance Composites. 1994. July / August. Lewis, I.e. and Nazem, EE 1987a. 18th Conference
Jain, M.K. and Alhiraman, AS. 1987. J. Mater. Sci. 22: Carbon, Extended Abstracts, American Carbon
278. Society, p. 190.
Jakubowski, J.J. and Subramanian, RY. 1979. Chern. Lewis, I.e. and Nazem, EE 1987b. 18th Conference
Abstr. 1981,95: 98874. Carbon, Extended Abstracts, American Carbon
Johnson D.J. 1987. Chemistry and Physics of Carbon Society, p. 290.
20: 1. New York: Marcel Dekker, Inc. Mair, W.N. and Mansfield, E.H. 1987. William Watt
Johnson, J.w. 1969. App. Polymer Symp. 9: 229. 1912-1985. Biographical Memoirs of Fellows of the
Johnson, J.W., Potter, W., Rose, P.G. and Scott, G. Royal Society, 33: 643-667.
1972. Brit. Polymer J., 4: 527-540. Matsumoto, T. 1985. Pure Appl. Sci. 57: 1553.
Johnson, J.w., Rose, P.G. and Scott, G. 1970. Proc. 3rd Mochida, I., Shimizu, K, Korai, Y., Ohtsuka, H. and
Con! Industrial Carbon and Graphite, London: Fujiyama, S.1988. Carbon 26: 843.
Academic Press, p. 443. Nazem, EE 1982. Carbon 20: 345.
Jones, RM. 1975. Mechanics of Composite Materials, Nazem, EE and Lewis, I.e. 1986. Mol. Cryst., Liq.
Washington, D.C: Scripta Book Co. Cryst. 139: 195.
Joseph, D. and Oberlin, A 1983a. Carbon 21: 559. Oberlin, A and Guigon, M. 1988. Fibre Reinforcement
Joseph, D. and Oberlin, A 1983b. Carbon 21: 565. for Composite Materials, (ed. AR. Bunsell).
Karpinos, D.M. and Izmalkov, a.M. 1982. Gibkie Amsterdam: Elsevier, p. 149.
Elektroprovodnye Materialy I Ustroistva na Oberlin, A. and Oberlin, M. 1981. Revue Chim.
Osnove dlya Obogreva Lyudei I Tekhniki. Miner. 18: 442.
Kawamura, K and Jenkins, G.M. 1970. A New Oberlin A. 1984. Carbon 22: 521.
Glassy Carbon Fiber. J. Mater. Sci. 5: 262. Oberlin, A. and Guigon, M. 1984. Science and New
Kowbel, W., Wapner, P. G. and Wright, M. A 1988. Applications of Carbon Fibers, Toyohashi
J. Phys. Chem. Solid 49: 11. University of Technology, Japan.
Lafdi, K, Bonnamy, S. and Oberlin, A. 1991a. Ohtsuka, H. 1988. Mitsubishi Gas-Chemical Co.
Mechanism of anisotropy occurrence in a pitch Kurashiki, Okayama 712, Japan.
precursor of carbon fibres, Part I - Pitch A and Reynolds, W.N. and Moreton, R. 1980. Philos. Trans.
B. Carbon 29: 831. Roy. Soc. A294: 451.
Lafdi, K, Bonnamy, S. and Oberlin, A 1991b. Reynolds, W.N. and Sharpe, J,Y. 1974. Carbon 12: 103.
Mechanism of anisotropy occurrence in a pitch Richter, E., Knoblauch, K, Juntgen, H. Deutsche
precursor of carbon fibres, Part II - Pitch e. Offen. Patent 3 412 761, 1984.
 References 201

Riggs, D.M., Shuford, R. and Lewis, R. 1982. US Patent 3 894 884, 1975. Druin.
Handbook of Composites. New York: Van US Patent 3 917 776, 1975. Mitsubishi Rayon.
Nostrand Reinhold. US Patent 3 919 383, 1975. Singer.
Riggs, D.M., 1979. Doctoral Thesis. Rensselaer US Patent 3 972 984, 1976. Nippon Carbon Co.
Polytechnic Inst., Troy, New York. US Patent (8) 3 974 264,1976. McHenry.
Robson, D., Assabghy , EY.I. and Ingram, D.J.E. US Patent (8) 3976729, 1976. Lewis, McHenry,
1972. ]. Phys. D, 5: 169. Singer.
Rosen, B.w. 1964. AlAA 2: 1985. US Patent 3 989 802, 1976. Loo.
Schwartz, M.M. 1984. Composite Materials Handbook. US Patent 4 004 053, 1976.
New York: McGraw-Hill. US Patent 4 005183, 1977. Singer
Shindo, A, Nakanishi, Y. and Soma, 1. 1969. Appl. US Patent 4 009 248, 1977.
Polym. Symp. 9: 305. US Patent (4) 4 017 327, 1977. Lewis, McHe~
Singer, L.S. 1978. Carbon 16: 409. Singer.
Tanabe, Y., Yasuda, E., Machino, H. and Kimura. S. US Patent 4 024 227, 1978.
1987. Ann. Mtg Jpn Ceramic Society/Nagoya, US Patent 4031 288, 1978. Minnesota Mining and
77. Manufacture Co.
Ternovoi, K. S., Zemskov, V. S., Kolesnikov, E. B. US Patent 4039341, 1976. National Research
and Mashkov, O. A 1985. Sorbitsionnaya Development Corp.
Detoksikatsiya v Khirurgicheskoi Klinike US Patent 4 069 297, 1978. Toho Beslon Co. Ltd.
(Detoxification Sorption in Surgery) Kishinev US Patent (6) 4 208 267, 1980. Diefendorf and Riggs.
(USSR): Shtiintsa. US Patent (5) 4 331620,1982. Diefendorf and Riggs.
Tsai, S.w. and Hahn, H.T. 1980. Introduction to US Patent 4 374 114, 1983. Kim.
Composite Materials, Technomic Publishing Co., US Patent (3) 4 376 747, 1982. Nazem.
Inc., Westport, CT. US Patent 4 504 454, 1985. Riggs.
US Patent 3 107 152, 1963. Ford and Mitchell. Watt, W. 1972. Carbon 10: 121.
US Patent 3 533 741, 1970. Courtaulds Limited. Watt, W. and Johnson, W. 1975. Mechanism of oxi-
US Patent 3 627 466,1971. Steingiser. dation of polyacrylonitrile fibres. Nature 257:
US Patent (12) 3 629 379, 1971. Otani. 210-212.
US Patent 3 650 668, 1972. Celanese. White, J.L. 1992. ONR Report for Contract No.
US Patent 3 656 882, 1972. Celanese. 88-K-0424 and 89-J-3056.
US Patent 3 656 883, 1972. Celanese. Wright, M.A., 1989. NASA Conference Publication
US Patent 3 656 903, 1972. Celanese. 3054: 17.
US Patent 3 660140,1972. Scola. Wright, M.A. and Iannuzzi, EA 1973. J. Camp. Mat.,
US Patent 3 671 411, 1972. Ray. 7: 430.
US Patent 3 677 705, 1976. Celanese. Wright, M.A. and Wills, J.L. 1974. J. Mech. Phys. Sol.
US Patent 3 723 150, 1973b. Druin. 22: 161.
US Patent 3 746450, 1973. Goan. Wright, M.A. and Palmer, K.R. 1994. Research into
US Patent 3 746 506, 1973. Aitken. Structural Carbons, Materials Technology
US Patent 3 754 957, 1973a. Druin. Center Publication, SIVC, Carbondale, illinois,
US Patent 3 767 773, 1973. Turner. 62901.
US Patent 3 767 774, 1973. Hou. Yanagida, K., Noda, M., Sasaki, T. and Tate, K. 1991.
US Patent 3 780255, 1973. Boom. 20th Conf Carbon, Extended Abstracts, American
US Patent 3 814377, 1974. Monsanto. Carbon Society, p. 160.
US Patent 3 832 297,1974. Paul, Jr. Yoneshoga, 1. and Teranishi, H. 1970. Japanese
US Patent 3 841 079, 1974. Celanese Patent Specification 2774/70.
US Patent 3 846 833, 1975. Celanese. Yooh, S.H., Korai, Y. and Mochida, I. 1994. Carbon
US Patent 3 859 187, 1975. Druin. 32: 281.
ORGANIC FIBERS 10
Linda L. Clements

10.1 INTRODUCTION A different type of high performance organic

fiber, extended chain polyethylene fibers, was
Before the first aramid fibers were introduced
added in the 1970s. While inferior to inorganic
in the 1960s and 1970s, organic fibers were rel-
fibers in some properties, organic fibers provide
atively low performance materials, primarily
used in textile applications. Now several dif- combinations of properties not available with
inorganic fibers and so have made possible new
ferent types of high performance organic
fibers exist, all competitive with inorganic designs and applications.
In this chapter, only high performance
fibers in some or even most of their properties.
organic fibers which are commercially avail-
The market demand for these fibers exceeds
one billion dollars (Adams and Farrow, 1993a). able will be discussed in detail, although fibers
which are nearing commercialization will be
The main applications for high perfor-
discussed briefly. For a more complete review
mance organic fibers today are in asbestos
of both commercially available and experi-
replacement, ballistics, rubber reinforcementI
ropes and cables and composites. Most of the mental high performance organic fibers, see
Yang (1989, 1992).
usage is of aramid fibers, with over 18 000
metric tons used each year. Both usage and
existing capacity for other organic fibers are 10.2 ARAMID FIBERS
only a fraction of this value (Adams and
Farrow, 1993a). 10.2.1 OVERVIEW
This broad market for organic fibers is a
direct outgrowth of applying the basic princi- Aramid fiber is the generic term for a specific
ples of polymer science to produce a new and type of 'aromatic polyamide fiber.' The US
exceptional engineering material. In the 1950s it Federal Trade Commission defines an aramid
was recognized that if a means could be found fiber as 'a manufactured fiber in which the
to form certain intractable polymers into fiber-forming substance is a long-chain syn-
extended chain fibers, very high stiffnesses, thetic polyamide in which at least 85% of the
strengths and use temperatures could be amide linkages are attached directly to two
achieved. The difficulty of producing such aromatic rings.'
fibers was solved in the 1960s by spinning from Thus, in an aramid, most of the amide
liquid crystalline solutions. The first fibers pro- groups are directly connected to two aromatic
duced by this process were the aramids, which rings, with nothing else intervening. It should
have since been followed by other such fibers. not be surprising that aramids have quite dif-
ferent properties from nylons and other
conventional polyamides since the latter poly-
Handbook of Composites. Edited by S.T. Peters. Published
mers contain few if any aromatic groups in the
in 1998 by Chapman & Hall, London. ISBN 0 412 540207 main chain of the polymer.
 Aramid fibers 203

Aramid fibers can be separated into two include DuPont's Kevlar®, Akzo's Twaron®,
types: the para- aramids and the meta-aramids. Teijin's Technora® and Kaiser VIAM's Armos®
In para-aramids, the chain-extending bonds are and SVM@ fibers, while meta-aramids include
in the para-position on the aromatic ring, as in DuPont's NomeX® and Teijin's Teijinconex®
poly-p-phenylene terephthalamide (PPTA) (Fig. fibers. Hoechst AG also markets a para-aramid
lO.l(a», co-poly-p-phenylene/3,4'-oxydipheny- fiber in Europe. The para-aramids are the fibers
lene terephthalamide (Fig. 10.1(b» and used in high performance applications and thus
poly-p-phenylene-benzimidazole-terephthala- will be emphasized in this chapter.
mide (Fig. 10.1(c». In meta-aramids, on the
other hand, the chain-extending bonds are in
10.2.2 MANUFACTURE
the meta-position on the aromatic ring, as in
poly-m-phenylene isophthalamide (MPIA) (Fig. Historically, meta-aramid fibers were the first
1O.1(d». Commercially available para-aramids to be produced, with DuPont's Nomex fiber

o
--fi-
-0- 1
(a) "C ~ H

oII~" N N
H1 "

(b)

1-0-
H

~-O-( 0 ~
o N N

o m H n

Fig. 10.1 Structural formulae of (a) the para-aramid poly-p-phenylene terephthalamide (PPTA), (b) the
para-aramid co-poly-p-phenylene/3,4'-oxydiphenylene terephthalamide, (e) the para-aramid poly-p-
phenylene-benzimidazole-terephthalamide (PBIA), and (d) the meta-aramid poly-m-phenylene isophthal-
amide (MPIA).
204 Organic fibers

being introduced in the 1960s. The first para- 10.2.3 STRUCTURE

aramids were synthesized in 1965 by S.L.
The excellent properties of para-aramids result
Kwolek of DuPont (Kwolek, 1971; Kwolek,
from both chemistry and physical microstruc-
1972; Kwolek, 1974). Forming these into usable
tures. In both meta- and para-aramids, the
fibers is very difficult because para-aramids
aromatic rings in the backbone chain produce
show no melting point and are soluble in a
high thermal resistance. In addition, in para-
limited number of solvents.
aramids the orientation of the chain-extending
The problem of spinning the polymer into
bonds produces a polymer which is an
fibers was solved for PPTA following the dis-
extended-chain rigid rod. Spinning produces a
covery that the polymer would dissolve in
fiber made up of extended-chain crystallites
strong acids to form a liquid crystalline solu-
which are almost completely aligned parallel
tion. Undiluted sulfuric acid is the solvent
to the draw direction and to each other. The
usually used. Blades (1973, 1974) devised a crystallites have a very high length-to-diame-
special manufacturing process - known. as ter ratio and extensive interconnection of
continuous dry jet wet spinning - for forrrung
molecules between crystallites. Thus, an
the liquid crystalline solution into filaments.
unbroken 'infinite' filament can be formed.
The polymer solution is extruded through
Within the crystallite the chains are bonded
spinnerets at elevated temperature through an
to one another by hydrogen bonds, as shown
air layer into a coagulating water bath. The
in Fig. 10.2. Although these bonds are not
cold water bath also contains a base to neu-
nearly as strong as the covalent bonds which
tralize and remove the retained acid.
occur within the molecules, hundreds or even
Continuous dry jet wet spinning is the man-
thousands of such bonds form between adja-
ufacturing technique used for most
cent para-aramid molecules. Since the
para-aramid fibers. Teijin's Tec~or~ fiber,
molecules are rigid, the only way to separate
however, is produced by wet spmrung fol-
them in tension is to break all of the hydrogen
lowed by drawing (Hongu and Phillips, 1990).
bonds at once. This requires a large force and

o
~c O~
oII~~ N -0- H
N
1 0

H
:

.
1
1

H
:
•
~c-o-~
0
:
II~~
N-
1
0-
0
H

N
1
~

0-" -0-1
-N 0 : H
~ H
~-~C~ 0

o ~ ~N~ -0- 11

H ~~c~~-o-
H

Fig. 10.2 Schematic shOWing hydrogen bonding between PPTA molecules in the crystallite.
 Aramid fibers 205

is the reason para-aramid fibers are exception- mechanical, thermal, physical and other proper-
ally strong in axial tension. However, since the ties. This anisotropy may produce design
bonds can be broken easily one at a time, the limitations, but can also be used to advantage.
fibers are quite susceptible to damage by
bending, buckling or transverse loading.
Physical and thermal properties
In meta-aramids, on the other hand, a
crooked chain results. Since even in pure ten- Table 10.1 compares the physical and thermal
sion the chain-extending bonds can flex and properties of some representative aramid
rotate, meta-aramids are much less rigid than fibers.
para-aramids and not as strong. However, Due to their highly aromatic and ordered
because the chains are more flexible, meta- structure, aramids have very high thermal
aramids are easier to manufacture than resistance for organic materials. They do not
para-aramids and are less expensive. melt prior to decomposition, in spite of the fact
that they are technically classified as thermo-
plastics. This is because melting of the
10.2.4 PROPERTIES
crystalline phase, like rupturing the fiber in
Aramid fibers offer some significant advan- tension, would require that all of the hydrogen
tages over other fibers, but also have their bonds between two molecules be severed at
drawbacks and limitations. Both advantages once. Nonetheless, because of decomposition,
and limitations will be described more fully in their temperature resistance is not equal to
the sections on properties and in the sections that of inorganic fibers. Thermogravimetric
on design considerations and applications. analysis of Kevlar fibers shows that weight
Both DuPont's Kevlar family of fibers and loss begins at above 350°C (660°F) in air (Penn
Akzo's Twaron fibers are based upon PPTA and Larsen, 1979; Yang, 1992), with complete
(Fig. 10.1(a». Teijin's Technora fiber and the decomposition occurring at between 427 and
para-aramid marketed by Hoechst AG in 482°C (800 and 900°F) (DuPont, 1992a).
Europe, on the other hand, are a para-aramid Exposure to elevated temperature will
copolymer, co-poly-p-phenylene/3,4'-oxy- degrade the properties of aramid fibers. Figure
diphenylene terephthalamide (Fig. 10.1(b». It is 10.3 shows the strength retention of Kevlar 29
likely that Kaiser VIAM's SVM fibers are poly- and Technora fibers as a function of time and
p-phenylene-benzirnidazole-terephthalamide temperature. This change in properties occurs
(PBIA), (Fig. 10.1(c» rather than PPTA as a result of slow oxidation. For this reason,
(Gerzeski, 1989). Kaiser VIAM's Armos fiber the long-term use temperature of para-aramid
may be PBIA or PPTA. Both DuPont's Nomex fibers is typically limited to about ISO-175°C
and Teijin's Teijinconex fibers are based upon (300-350°F).
MPIA (Fig. 1O.1(d». These chemical and struc- In the transverse direction para-aramids are
tural differences produce different properties like most other materials in that they expand
for the fibers. In addition, differences in spin- with increasing temperature. However, in the
ning conditions and, most importantly, longitudinal direction the fibers actually con-
post-spinning heat treatments are used to alter tract somewhat as temperature increases. The
properties further. For example, by changing negative thermal expansion coefficient of
processing conditions, Kevlar fibers can be para-aramids can be used to advantage to
produced with elastic moduli ranging from 63 design composites with tailored or zero ther-
to 143 GPa (9 to 21 Msi) and elongations at mal expansion coefficient.
break from 1.5 to 4.4%. Aramids are flame resistant but can be
Because of the anisotropy of their microstruc- ignited. While pulp or dust of Kevlar may con-
ture, para-aramid fibers have very anisotropic tinue to smolder once ignited, fabrics do not
206 Organic fibers

Table 10.1 Physical and thermal properties of representative aramid fibers

Fiber Kevlar 49 TwaronHM Technora Nomex Teijinconex

Type para-aramid para-aramid para-aramid meta-aramid meta-aramid
copolymer
Reference for data DuPont Akzo Teijin DuPont Teijin
1992a 1990,1991 1989,1993 1981,1993g 1991
Density 1.44 1.45 1.39 1.38 1.38
g cm-3 (lb in-3) (0.0520) (0.0524) (0.0502) (0.0499) (0.0499)
Melting temperature -538°C" >500 oe >371°e
(1000"F) (>932"F) (>700°F)
Decomposition 427-482°e saooe 500°C 371°C 400-430 oe
temperature in air (800-900°F) (930°F) (930°F) (700°F) (750-805°F)
Long-term use 149-177°e
temperature in air (300-350°F)
Longitudinal linear -4.9 -3.5 -6.2 +15 +20
thermal expansion (-2.7) (-1.95) (-3.4) (+8.3) (+11)
coefficienti'
10~/oe
(1O~/ OF)
Transverse linear +66"
thermal expansion (+37)
coefficienti':
/oe (/ OF)
Specific heati' 1.42 1.42 1.09 1.21 1.05
kJ Ikg K (BTU lIb OF) (0.34) (0.339) (0.26) (0.29) (0.25)
Longitudinal thermal
conductivityb 4.11< 4.0 0.13 0.13
1O-2 W/mK (2.38) (2.3) (22) (22)
(10-2 BTU /h ft OF)
Transverse thermal
conductivityb 4.82< 5.0
10-2 W/mK (2.79) (2.9)
(10-2 BTU /h ft OF)

Equilibrium 3.5% 3.5% 2.0% 4.5% 5.0-5.5%

moisture contenti'
Typical filament 12 or 15 12 12 max: 15-17 -10 to 15 X 45
diameter (0.48 or 0.59) (0.48) (0.48) (0.6-0.7) (-0.4 to 0.6 X 1.1)
~m (10-3 in)

Typical filament round round round oval to oval to

shape dogbone dogbone

a Data from Yang, 1992.

b Varies with temperature; room temperature values are given.
C Data from Chiao and Chiao, 1982.
 Aramid fibers 207

100

.~ 80
C
0
:;::
c
.! 60
•
-
II::
J:

-
~
c: 40
!
en
- Technorae
20

0
0.1 10 100 1000
Tim•• h

Fig. 10.3 Strength retention of Kevlar 29 and Technora fibers following elevated temperature exposure
(DuPont, 1992a; Teijin, 1989).

continue to bum when the flame source is fibers in the mid-1980s, aramid fiber composites
removed (DuPont, 1992a). The lower thermal had the highest specific strengths of all compos-
conductivity of aramids compared to inor- ite materials. Although composites from newer
ganic fibers can improve the fire resistance of fibers have taken over that position, aramids
their composites, since aramids do not readily still offer outstanding combinations of proper-
conduct heat into the more volatile matrix. ties, such as high specific strength, toughness,
creep resistance and moderate cost, for specific
applications.
Mechanical properties
Table 10.2 compares the mechanical proper-
Composite materials are most commonly used ties in axial tension of several commercially
because of their superior strength and/or stiff- available aramid fibers.
ness at a given weight as compared to Aramid fibers have some definite limita-
conventional structural materials. Figure lOA tions. They are weak in bending and show
compares the specific strengths and specific obvious damage if subjected to kinking or
stiffnesses of various reinforcing fibers. (The buckling. As a result, they are also weak in com-
strengths and stiffness in Fig. lOA are expressed pression (where microbuckling is inevitable)
in units of grams per denier (gpd). This is a tex- and in transverse tension (where bond-by-bond
tile term often used for organic fibers which breakage of hydrogen bonds is likely). In addi-
measures specific strength and/ or stiffness. This tion, even though the para-aramid chain is
term is further explained in the appendix to this quite polar in nature, almost all of the polar
chapter.) As can be seen, aramid fibers perform groups are fully involved in hydrogen bonding
very well. In fact, until the emergence of high to other aramid molecules. As a result, para-
strength intermediate modulus carbon fibers aramid fibers do not form strong bonds with
and the commercialization of polyethylene other materials such as composite matrices,
208 Organic fibers

Table 10.2 Axial tensile mechanical properties of representative aramid fibers

Fiber Reference Specific Initial tensile Tensile strength, Elongation at

gravity modulus, GPa (Msi) MPa (ksi) break, %
Bare" Epoxy- Bare" Epoxy- Bare"
impregnated" impregnated"
Kevlar DuPont, 1993h 1.44 71.8 2920 3.6
Type 956, (10.4) (424)
1500 denier
Kevlar 29 DuPont, 1992a, 1.44 70.5 83.0 2920 3600 3.6
Type 964, DuPont, 1993g (10.2) (12.0) (424) (525)
1500 denier
Kevlar49 DuPont, 1992a, 1.44 112.4 124.0 3000 3600 2.4
Type 965, DuPont, 1993g (16.3) (18.0) (435) (525)
1140 denier
Kevlar 68 DuPont, 1993g 1.44 99.13 3050 2.9
Type 989B, (l4.4) (442)
1420 denier
Kevlar 119 DuPont, 1990 54.6 3050 4.4
1500 denier (7.9) (442)
Kevlar 129 DuPont, 1993i 1.44 96.0 3380 3.3
denier (13.9) (490)
unspecified
Kevlar 149 DuPont, 1993g 1.47 142.7 2340 1.5
Type 965A, (20.7) (339)
1140 denier
Kevlar HT DuPont, 1993g 1.44 99.1 3370 3.3
Type964C, (14.4) (489)
1000 denier
Kevlar KM2 DuPont, 1992d 63.4 3280 4.0
850 denier (9.2) (476)
Twaron Akzo, 1991 1.44 70 2800 3500 3.6
(10.2) (406) (508)
Twaron Perkins, 1993 1.44 88 3230 3.3
Type 2000, (12.8) (468)
930 denier
'microfilament'
TwaronHM Akzo, 1991 1.45 103 2800 3500 2.5
(14.9) (406) (508)
Armos Kaiser VIAM, 147 4592 3.2
58.8 tex 1993ac (21.4) (666)
SVM Kaiser VIAM, 1.43 123 2917 3720
58.8(300) 1993g" (17.8) (423) (540)
X17-1000 Gerzeski,1989
Continued on next page
 Aramid fibers 209

Table 10.2 Continued

Fiber Reference Specific Initial tensile Tensile strength, Elongation at

gravity modulus, CPa (Msi) MPa (ksi) break, %
Bare" Epoxy- Bare" Epoxy- Bare"
impregnatedb impregnatedb
Technora Teijin, 1989 1.39 73 3440 4.6
(10.6) (498)
Nomex DuPont, 1993g 1.38 11.6 596 28.0
Type 430, (1.68) (86.6)
1200 denier
Teijinconex Teijin, 1991 1.38 7.9-9.7 610-670 35-45
(1.1-1.4) (88-97)
Teijinconex Teijin, 1991 1.38 11.6-12.2 730--850 20--30
HI (1.7-1.8) (110--120)

a Data for DuPont fibers taken from conditioned yams tested according to AS1M Standard D885.
Modulus data for Akzo fibers from testing according to AS1M Standard D885M. Test technique
unspecified for Akzo fiber strengths and elongations and for all data from Kaiser VIAM and Teijin fibers.
b Data for DuPont fibers taken from epoxy-impregnated strands tested according to ASTM Standard
D2343. Data for Akzo fibers from testing according to impregnated strand test method DIN 65356, part
2. Test technique unspecified for Kaiser VIAM fibers.
< Preliminary data.

"C
0.. • PBO
bO 40 • PBO HM

..c:
.....
Armas
•
Ql) Carbon TI000G • • Spectra 1000
I=: Dyneema SK60 • Carbon 1MB
Q)

.....s... 30 Technora •• Tekmilon

Ul •
Vectran HS •Spectra 900
Q) • Kevlar 49
......
..... Twaron •• SVM • Carbon T-300
••
rIl
I=: ·Twaron HM
Q) 20 S-Glus
E-< • Kevlar 149 • Carbon T-50
.....
C)
..... • • PE (H.C.) • Boron
..... E-Glas8
C)
Q)
10
•
Carbon P-I00
0..
Ul
• Steel

0
0 SOO 1000 1500 2000 2500 3000 3500 4000 4500

Specific Tensile Modulus, gpd

Fig. 10.4 Specific tensile strength against specific tensile stiffness for various reinforcing fibers.
210 Organic fibers

further aggravating the poor transverse, bending and/ or prepared using other fabrication
and compressive properties of the fiber itself. processes, the general trend is valid: aramid
The basic chemical structure differences fiber composites have poor off-axis properties.
between the aramid fibers produce many of In axial tension, both aramid fibers and
the mechanical property differences seen in their composites are linear to failure. In spite
Table 10.2. The ether (-0-) linkages in the of this fact, the same microstructural charac-
backbone of the Technora copolymer fiber pro- teristics which lead to the weakness of aramid
duce a lower modulus than that of Kevlar and fibers in buckling also make them very tough.
Twaron PPTA-based fibers. On the other hand, During failure, the widespread bending, buck-
the additional cyclic ring in the SVM PBlA- ling and other internal damage to the fibers
based fibers produces a higher basic modulus. absorbs a great deal of energy. Similarly, the
However, heat treatment and other fabrication strength of aramid fibers is not very strain rate
steps can also alter mechanical properties sig- sensitive: an increase in strain rate of more
nificantly, as is seen in the property differences than four orders of magnitude only decreases
between the various Kevlar fibers. the tensile strength by about 15%. (Abbott et
The mechanical properties of aramid com- al., 1975) This property alone provides design
posites are illustrated in the data of Table 10.3. advantages over all inorganic and many other
For this filament-wound composite the longi- organic fibers.
tudinal compressive strength was about The mechanical properties of aramid fibers
one-eighth that in longitudinal tension, the in- decrease with increasing temperature. Figure
plane shear strength was one-seventy-fifth and 10.5 shows the fiber elastic modulus as a func-
the transverse tensile strength over two hun- tion of temperature for several organic fibers.
dred times smaller. While the relative values of At 177°C (350°F) the modulus of para-aramid
properties may change for composites made fibers is about 80% of that at room tempera-
from other aramid fibers and/ or other matrices ture. Figure 10.6 compares the fiber tensile

Table 10.3 Mechanical properties of a filament-wound composite of 60 vol % aramid fiber in a room-tem-
perature curable epoxy matrix (Clements and Moore, 1977)
Fiber: DuPont's Kevlar 49, Type 968, 1420 denier
Matrix: 100 parts Dow Chemical DER 332 (diglycidyl ether ofbisphenol-A epoxy) and 4S parts Jefferson
Chemical Jeffamine T-403 polyether tria mine
Cure: 1 day at room temperature, postcure 16 hat 8S0C (18S°F)
Elastic constants:
Longitudinal Young's modulus En' GPa (Msi) 81.8 ± 1.5" (11.9 ± 0.22)
Transverse Young's modulus E22, GPa (Msi) 5.10 ± 0.10 (0.74 ± 0.014)
Shear modulus G12, GPa (Msi) 1.82 ± 0.09 (0.26 ± 0.013)
Major Poisson's ratio '\)12 0.310 ± 0.035
Minor Poisson's ration 1)21 0.0193 ± 0.0014
Ultimates: Tension Compression In-plane shear
Longitudinal strength, MPa (ksi) 1850 ± 50 (268 ± 7.3) 235 ± 3 (34.1 ± 0.4)
Longitudinal ultimate strain, % 2.23 ± 0.06 0.48 ± 0.3
Transverse strength, MPa (ksi) 7.9 ± 1.1 (1.15 ± 0.15) 53 ± 3 (7.7 ± 0.4)
Transverse ultimate strain, % 0.161 ± 0.023 1.41 ± 0.12
Shear stress at 0.2% offset, MPa (ksi) 24.2 ± 2.4 (3.51 ± 0.35)
Shear strain at 0.2% offset, % 1.55 ± 0.16
a Limits are 95% confidence limits. Each value is the result of five or more tests.
 Aramid fibers 211

Temperature, of
100 200 300 400 500 600

15
100

75

50

5
25

Temperature, DC
Fig.tO.s Modulus as a function of temperature for several para-aramid fibers (DuPont, 1992a; Teijin, 1989).
Temperature of

100 200 300 400 500

3500 500

3000
400
2500
0
a. ·ii
300

-
:2 2000 ~

£ ~
0, 01
c
c 1500 II

~ Polyester
200 ~

1000
100
500

0 0
a 50 100 150 200 250 300
Temperature. °C

Fig. 10.6 Tensile strength as a function of temperature for two para-aramid fibers and for two polymer
fibers and steel (DuPont, 1993h; Teijin, 1989).
212 Organic fibers

strength as a function of temperature for sev- wet transverse tensile and in-plane shear
eral organic fibers. For Kevlar fiber the strengths were only about half of the 52% r.h.
strength at 177°C (350°F) is about 80% of that values. The data in boiling water illustrate that
at room temperature, while for Technora the the drops in strength due to the presence of
strength is about 70% of the room temperature moisture alone were almost as severe as those
value. On the other hand, at cryogenic tem- due to the combined presence of moisture and
peratures modulus increases slightly and elevated temperature. This relative loss in
strength is not degraded. properties is less for the Technora para-aramid
The presence of moisture also reduces the co-polymer fiber. Care must be exercised when
mechanical properties of aramid fibers and using aramid composites in high moisture
their composites. The effect upon longitudinal applications.
tensile properties is relatively small, but the Both para-aramid co-polymers and homo-
loss is pronounced for off-axis properties. polymers exhibit very little creep. In general,
Table 10.4 illustrates this loss for Kevlar 49 creep strain increases with increasing tempera-
fiber in a room-temperature curable epoxy. ture, increasing stress and decreasing fiber
The longitudinal tensile strength in water at modulus. Like all high performance fibers,
room temperature was 88% of that for com- under long term loading, para- aramids are
posites equilibrated at room temperature and subject to stress rupture, i.e. failure of the fiber
52% relative humidity (r.h.). The wet longitu- under sustained loading with little or no
dinal compressive strength, on the other hand, accompanying creep. Figure 10.7 compares the
was only 75% of the 52% r.h. value, while the stress rupture performance of Kevlar 49 to that

Table 10.4 The effect of environments on the mechanical properties of a filament-wound composite of 50
vol % of an aramid fiber in a room-temperature curable epoxy matrix (Wu, 1980)
Fiber: DuPont's Kevlar 49,4560 denier
Matrix: 100 parts Dow Chemical DER 332 (diglycidyl ether ofbisphenol-A epoxy) and 45 parts Jefferson
Chemical Jeffamine T-403 polyether triamine
Cure: Infrared heating, postcure 2 hat lOO°C (212°F)
Strength, MPa (ksi)
23°e, dry 23°C, 52% r.h. 23°e, water 100°C, water
Longitudinal 1370 ±62' 1340 ± 112 1190 ± 62 1150 ± 124
tension (199 ±9) (194 ± 16) (173 ± 9) (167 ± 18)
Longitudinal 188 ± 12 169 ± 20 126 ± 22 107 ± 21
compression (27.3 ± 1.7) (24.5 ± 2.9) (18.3 ± 3.2) (15.5 ± 3.0)
Transverse 7.6 ± 1.6 74 ± 1.2 3.9 ± 0.7 3.6 ± 0.2
tension (1.10 ± 0.23) (1.07 ± 0.17) (0.57 ± 0.10) (0.52 ± 0.03)
Transverse 31.3 ± 3.2 29 ± 4.0 22.5 ±3.2 22.1 ± 23.6
compression (4.54 ± 0.46) (4.21 ± 0.58) (3.26 ± 0.46) (3.20 ± 3.42)
In-plane shear 27 ± 3.0 26.5 ± 1.6 13.8 ± 2.2 13.6 ± 2.5
(3.92 ± 0.44) (3.84 ± 0.23) (2.00 ± 0.32) (1.97 ± 0.36)
Hygrothermal Properties
Equilibrium moisture 4.1 7.8 8.9
concentration, %
• Limits are 95% confidence limits. Each strength is the average of five tests.
 Aramid fibers 213

•
~100 r----,----~----_r----~----~----r_--~

I
.~

!! 90 Kevlare 49
:s
~ 80 50' of
specimens

"
failed
2
••
• 70

~
• 60
8-g1ass

•
1.... 50
...DI
: 40 '-----....J'--------I.-------..1.____--L____..l-____L -__-.J

1 10
Li.fetime. h

Fig. 10.7 Stress-rupture behavior of epoxy-impregnated Kevlar 49 fibers compared to that of epoxy-
impregnated S-glass fibers (Chiao, Chiao and Sherry, 1976).

of S-glass. Para-aramids perform well under (DuPont, 1986). Para-aramids also can be
these conditions, but the phenomenon of stress expected to perform better than carbon fibers in
rupture must be considered in any design fatigue (Teijin, 1989; Yang, 1992). Technora
where long term loading is anticipated. para-aramid co-polymer is found to have even
Strength retention cannot be used to estimate better fatigue resistance than the para-aramid
the remaining life of aramid fibers or compos- homopolymer fibers (Teijin, 1989).
ites under long term load (Chiao, Sherry and
Chiao, 1976), so estimates of long term behavior
Chemical and environmental properties
must be derived from actual data, or acceler-
ated testing methods (Chiao and Chiao, 1982). PPTA fibers are quite stable chemically; their
Para-aramid fibers and their composites per- resistance to neutral chemicals is usually very
form very well in fatigue. For aramids, high. They are, however, subject to attack by
tension-tension fatigue generally is not of sig- acids and bases, especially by strong acids.
nificant concern in applications where an Because the spin process used for Teijin's
adequate static safety factor has been used Technora para-aramid co-polymer produces a
(Yang, 1992). Aramid composites have been very pure polymer, the chemical and environ-
found to be superior to glass fiber composites in mental resistance of Technora is superior to
both tensile-tensile and flexural fatigue load- that of the PPTA fibers. Table 10.5 reports the
ing. For the same lifetime (cycles to failure), resistance of Kevlar and Technora fibers to
Kevlar 49/ epoxy composites can operate at a various chemicals. Technora has better acid
significantly larger percentage of their static and alkali resistance than PPTA and its steam
strength than can glass-reinforced composites resistance is also superior.
214 Organic fibers

Table 10.5 Stability of para-aramid fibers in various chemicals

Concentration, Temperature, Time, Effect on breaking strength"

Chemical % °C (OF) hr None Slight Moderate Appreciable Degraded
Acids
Acetic 40 21 (70) 1000 K1>
40 95-99 (203-210) 100 T> K
Formic 90 21 (70) 100 K
90 95-99 (203-210) 100 T K
Hydrochloric 20 20 (68) 100 T
10 71 (160) 10 K
Nitric 10 20-21 (68-70) 100 T K
Phosphoric 10 20-21 (68-70) 100 K,T
10 21 (70) 1000 K
10 99 (21O) 100 K
Sulfuric 10 99 (210) 10 K
20 95 (203) 100 T
40 95 (203) 100 T
Alkalis
Ammonium hydroxide 28 21 (70) 1000 K
Sodium hydroxide 10 21 (70) 1000 K
10 95-99 (203-210) 100 T K
Portland cement saturated 95 (203) 100 T
saturated 180 (356) 15 T
Organic solvents
Acetone 100 boil 100 K
Benzene 100 20 (68) 784 T
100 21 (70) 1000 K
Carbon tetrachloride 100 boil 100 K
Ethylene chloride 100 20 (68) 1000 T
Ethylene glycol/water 50/50 99 (210) 1000 K
Ethylene glycol 100 95 (203) 300 T
Gasoline 100 20 (68) 784 T
Gasoline-leaded 100 21 (70) 1000 K
Methyl alcohol 100 21 (70) 1000 K,T
N-Methyl pyrrolidone 100 95 (203) 100 T
Other
Sodium chloride 3 21 (70) 1000 K
10 99 (210) 100 K
10 121 (250) 100 K
Sea water 100 95 (203) 1000 T
Sea water (New Jersey) 100 1yr K
Steam 100 120 (248) 400 T
100 150 (302) 48 K
100 150 (302) 100 T
100 200 (392) 100 T
Water, tap 100 99 (210) 100 K
• None, 0-10% strength loss; slight, 11-20% strength loss; moderate, 21-40% strength loss; appreciable,
41-80% strength loss; degraded, 81-100% strength loss.
b K is for Kevlar aramid fiber (DuPont 1989, 1993h); T is for Technora aramid fiber (Teijin, 1989).
 Aramid fibers 215

Para-aramids are strong ultraviolet (UV) in the para-aramid homopolymers and thus
absorbers. Upon exposure, the yellow or gold improved electrical properties. Technora fiber
fibers turn first orange and then brown, due to has a resistivity of 5 x 1014 il/ cm (Teijin, 1989).
degradation. The degradation occurs only in The dielectric constant of PPTA is 3.85 (Allied,
the presence of oxygen and is not enhanced by 1989).
either moisture or atmospheric contaminants The refractive index of Kevlar 49 fiber is 2.0
(DuPont, 1992a). Extended exposure may parallel to the fiber axis and 1.6 perpendicular
cause a loss of mechanical properties. Bare (DuPont, 1986). Aramid fibers are opaque and
1667 dtex (1500 denier) Kevlar 29 was found to are yellow to gold in color.
have 71 % strength retention after 1 month of
outdoor exposure in Wilmington, DE and 43%
10.2.5 TREATMENTS
after 4 months (Yang, 1992). In both processing
and applications, para-aramids must be pro- Unlike inorganic fibers, few surface treatments
tected from UV exposure, such as by painting are used on aramid fibers to promote matrix
or coating. However, since para-aramids are adhesion. One reason is the futility of increas-
self-screening, UV protection may also be ing the matrix bonding to the surface of a fiber
effected simply by dense packing of the fiber which readily fails by defibrillation. Most dra-
itself, with or without a matrix. Thus, bare matic improvements in fiber/matrix bonding
12.7 mm (0.5 in) 3-strand Kevlar 49 rope was give only modest improvements in off-axis
found to have 90% strength retention after 6 strengths since they simply move the locus of
months outdoors in Florida and 69% strength failure from the surface to the interior of the fil-
retention after 24 months (DuPont, 1986). ament. In other cases, longitudinal tensile
Unlike inorganic fibers, aramid fibers absorb strengths are adversely affected by otherwise
water. For some aramid fibers the equilibrium successful surface treatments. Not all attempts
moisture content (see appendix on page 241 for at designing surface treatments have been
definition) is quite high (5% for SVM, 7% for unsuccessful, but for the most part the surface
Kevlar, Kevlar 29 and Twaron), moderate for treatment used on commercial fibers is mini-
others (3.5% for Kevlar 49 and Twaron HM) mal compared to that used for inorganic fibers.
and reasonably low for some (2% for Armos Finishes - lubricants which aid in subse-
and Technora and about 1% for Kevlar 149) quent processing steps - are applied to aramid
(Akzo, 1991; Kaiser VIAM, 1993a; Teijin, 1989; fibers for some applications. Available finishes
Yang, 1992). The equilibrium moisture content are designed for such purposes as lubrication
is directly proportional to the relative humidity, during weaving operations, improving abra-
rising for Kevlar 49 to 6.2% at 96% r.h. (DuPont, sion resistance for cable applications or better
1992a). Absorbed moisture has only a small performance in rubber goods. If the fiber is to
effect upon the tensile properties of the fibers, be used in a high performance composite,
but a significant effect upon the transverse ten- however, the user will usually wish to avoid or
sile, compressive, shear and flexural properties remove any finish before impregnating. the
of the composite. The gain of moisture is com- fiber with a matrix.
pletely reversible and once removed produces Commercial aramid fibers may also be
no permanent property changes. twisted. Twist may be quite useful in some
applications and a small amount of twist will
increase the strength of bare yam or cord.
Electrical and optical properties
[This optimum twist for Kevlar fibers occurs at
Aramid fibers are electrical insulators. The a twist multiplier of 1.1. At about this value,
process used to make the Technora fiber, how- the strength of bare yam is the highest and the
ever, leaves it with fewer ionic impurities than modulus is only slightly decreased from the
216 Organic fibers

untwisted level (DuPont, 1992b).] Twist will staple, pulp and in various fabrics.
make the fiber easier to handle, make subse- While the meta-aramid fibers are not usu-
quent weaving or braiding operations easier ally used as fiber reinforcements in
and will improve the abrasion resistance of the composites, they are used extensively as rein-
fiber. It is also required for rope and cable forcements for honeycomb sandwich core
applications. However, once the fiber is used materials. The use of such materials along
in composite matrix fiber, twist is not desir- with composite face sheet panels has greatly
able. This is because twist interferes with full extended the overall usage of composite mate-
impregnation of the fiber with resin and with rials, particularly in the aerospace industry.
stress transfer between adjacent fiber bundles. Information about the availability and
It also increases stress concentrations, particu- package sizes of the fibers shown in Table 10.6,
larly at higher twist levels. For this reason, about other products and about special formu-
most of the aramid fiber manufacturers supply lations can be obtained from Table 10.7. At the
most or all of their fibers untwisted or with time of this publication, Kaiser VIAM's Armos
minimal twist. and SVM fibers are just being imported from
Russia. For this reason, information on the
fibers and their availability is limited in this
10.2.6 FORMS AND AVAILABILITY
chapter, but should be readily available later
Table 10.6 lists most of the commercial types of from the contact given in Table 10.7.
para-aramid fibers. Some of these fibers are
readily available in a variety of fiber deniers,
Pricing
package sizes, finishes and so forth, while oth-
ers are available only in limited quantities for Para-aramid fibers are currently priced from
specific applications. Due to constant changes about $20 per pound for the larger denier fibers
in market conditions and other factors, the to about $60 per pound for most of the small
user is advised to check with the fiber manu- denier, higher modulus fibers. (However, some
facturer concerning current availabilities. The of the very fine denier specialty fibers from
mechanical properties of fibers with different some manufacturers cost hundreds of dollars
deniers and/or finishes and other treatments per pound.) Prices can vary significantly for
will vary somewhat from each other and from similar fibers of different deniers or from dif-
the nominal values given in Table 10.2. ferent manufacturers and thus price quotes
In addition to the yams, tows and rovings should always be obtained before any decision
listed in Table 10.6, DuPont's Kevlar fibers are is made upon use of a specific fiber.
also available as staple (short fibers), floc (pre-
cision cut fibers of very short lengths) pulp
10.2.7 DESIGN CONSIDERATIONS
(very short and high fibrillated fibers) and in
specialty compounded forms (DuPont, 1992a, In the 1970s and early 1980s aramids began to
1993h). A variety of fabrics are also produced. replace carbon and glass fibers in many appli-
In addition, DuPont produces a colored fiber, cations. However, the development of high
Kevlar 100, in sage green, yellow, black and strength intermediate modulus carbon fibers in
royal blue (Yang, 1992). Both Teijin's Technora the mid-1980s and the commercialization of
and Akzo's Twaron fibers are available as sta- tough, high strength polyethylene fibers
ple and chopped fiber (very short lengths) and reversed this trend. Today aramid fibers are
in a variety offabrics (Teijin, 1989; Akzo, 1991). used mainly in applications where they offer a
Technora is also marketed in black as well as unique combination of properties, such as high
natural color. Kaiser VIAM's Armos and SVM specific strength combined with toughness and
fibers are also expected to be offered as tape, creep resistance.
 Aramid fibers 217

Table 10.6 Availability of commercial para-aramid fibers a

Product Count Filament Comments/typical applications

(reference)
number/ diameter
dtex (den) yarn 11m (1(13 in)
Kevlarb (DuPont, 1993g; Yang, 1992)
Type 950 1110 (1000) 666 12 (0.48) Finish: tire reinforcement
1670 (1500) 1000 12 (0.48)
2500 (2250) 1000 15 (0.59)
3330 (3000) 1333 15 (0.59)
Type 956 800 (720) 490 12 (0.48) Mechanical rubber goods: hoses, belts, etc.
1110 (1000) 666 12 (0.48)
1670 (1500) 1000 12 (0.48)
2500 (2250) 1000 15 (0.59)
3330 (3000) 1333 15 (0.59)
Kevlar 29 (DuPont, 1992b, 1993a, 1993b, 1993f, 1993g)
Type 960 1670 (1500) 1000 12 (0.48) Cordage finish: high lubricity for improved
3330 (3000) 1333 15 (0.59) abrasion resistance; ropes and cables
17000(15 000) 10000Rc 12 (0.48)
Type 961 1110 (1000) 666 12 (0.48) Textile finish; ropes and cables
1670 (1500) 1000 12 (0.48)
3330 (3000) 1333 15 (0.59)
5000 (4500) 3000 12 (0.48)
17000(15000) 10000R 12 (0.48)
Type 962 1670 (1500) 1000 12 (0.48) No finish; ropes and cables
3330 (3000) 1333 15 (0.59)
Type 963 3330 (3000) 1333 15 (0.59) Textile finish; non-apparel ballistic armor
Type 964 215 (200) 134 12 (0.48) Textile finish; ballistics and apparel,
430 (400) 267 12 (0.48) ignition cables
1110 (1000) 666 12 (0.48)
1670 (1500) 1000 12 (0.48)
Kevlar 49 (DuPont, 1992b, 1993c-g)
Type 965 61 (55) 25 15 (0.59) Textile finish; woven reinforcement in
215 (195) 134 12 (0.48) aerospace composites, ballistic armor, and
420 (380) 267 12 (0.48) printed circuit boards
1270 (1140) 768 12 (0.48)
1580 (1420) 1000 12 (0.48)
2400 (2160) 1000 15 (0.59)
Type 968 215 (195) 134 12 (0.48) No finish; marine composites, fiber optic
420 (380) 267 12 (0.48) cable reinforcement, ropes, filament-wound
1270 (1140) 768 12 (0.48) composites
1580 (1420) 1000 12 (0.48)
2400 (2160) 1000 15 (0.59)
3160 (2840) 1333 15 (0.59)
4800 (4320) 2000R 15 (0.59)
5070 (4560) 3200R 12 (0.48)
7900 (7100) 5000R 12 (0.48)
Continued on next page
218 Organic fibers

Table 10.6 Continued

Product Count Filament Comments/typical applications

(reference)
number/ diameter
dtex (den) yarn Ilm (1(r3 in)
Type 978 1580 (1420) 1000 12 (0.48) Cordage finish: high lubricity for improved
2400 (2160) 1000 15 (0.59) abrasion resistance; ropes and cables
5070 (4560) 3200R 12 (0.48)
Type 989 1580 (1420) 1000 12 (0.48) Textile finish; fiber optic cable
2400 (2160) 1000 15 (0.59) reinforcement
3160 (2840) 1333 15 (0.59)
4800 (4320) 2000R 15 (0.59)
6300 (5680) 2666R 15 (0.59)
7900 (7100) 5000R 12 (0.48)
9500 (8520) 4000R 15 (0.59)
Kevlar 68 (DuPont, 1992b, 1992c, 1993g)
Type 956B 215 (195) 90 12 (0.48) High performance mechanical rubber goods
1580 (1420) 1000 12 (0.48)
Type 989B 420 (380) 267 12 (0.48) Textile finish; fiber optic cable
1580 (1420) 1000 12 (0.48) reinforcement
2400 (2160) 1000 15 (0.59)
3160 (2840) 1333 15 (0.59)
4800 (4320) 2000 15 (0.59)
7900 (7100) 5000R 12 (0.48)
Keviar 119 (DuPont, 1990, 1993h)
Type 956E 1670 (1500) 1000 12 (0.48) Power transmission belts, high-performance
tires, high fatigue applications
Kevlar 129 (DuPont, 1992c, 1993h, 1993i)
Type 956C 1110 (1000) 666 12 (0.48) Mechanical rubber goods
Type 964C 830 (750) 500 12 (0.48) Personal body armor
930 (840) 600" 12 (0.48)
1110 (1000) 666 12 (0.48)
1580 (1420) 1000 12 (0.48)
Keviar 149 (DuPont, 1992c, 1993g)
Type965A 420 (380) 267 12 (0.48) Woven reinforcement in aerospace composites,
1270 (1140) 768 12 (0.48) hard ballistic armor, printed circuit
1580 (1420) 1000 12 (0.48) boards
Type 968A 1270 (1140) 768 12 (0.48) No finish; marine composites, fiber optic
1580 (1420) 1000 12 (0.48) cable reinforcements, ropes, filament-wound
4730 (4260) 3000R 12 (0.48) composites
7890 (7100) 5000R 12 (0.48)
Kevlar HT (DuPont, 1993g)
Type964C 1110 (1000) 666 12 (0.48) Advanced ballistic protection
Kevlar KM2 (DuPont, 1992d)
945 (850) 560 12 (0.48) Ballistic protection: helmets, composite armor

Continued on next page

 Aramid fibers 219

Table 10.6 Continued

Product Count Filament Comments/typical applications

(reference)
number/ diameter
dtex (den) yarn J.Lm (lo-3in)
Twaron d (Akzo, 1990, 1991; DeCos, 1993)
Type 1000 420 (380) 250 12 (0.48) Standard finish; multipurpose
840 (760) 500 12 (0.48)
1100 (990) 750 10.5 (0.41)
1260 (1130) 750 12 (0.48)
1680 (1510) 1000 12 (0.48)
2520 (2270) 1500R 12 (0.48)
3360 (3020) 2000R 12 (0.48)
Type 1001 420e (380) 250 12 (0.48) Adhesive-activated finish; tires,
840 (760) 500 12 (0.48) mechanical rubber goods, composites
110Qe (990) 750 10.5 (0.41)
1260 (1130) 750 12 (0.48)
1680 (1510) 1000 12 (0.48)
3360 (3020) 2000R 12 (0.48)
Type 1010 1680 (1510) 1000 12 (0.48) Very low finish level; composites
3360 (3020) 2000R 12 (0.48)
Type 1020 1680 (1510) 1000 12 (0.48) Special finish for increased abrasion
resistance; cables, ropes, nets
Type 1030 17000(15 300) 5000R 12 (0.48) PI'FE + silicone oil impregnated; braided packings
Type 1031 14 000"(12 600) 5000R 12 (0.48) PI'FE + silicone oil impregnated; braided packings
Type 1040 420 (380) 250 12 (0.48) Tangled yarn; multipurpose
840 (760) 500 12 (0.48)
1100" (990) 750 10.5 (0.41)
1260 (1130) 750 12 (0.48)
1680 (1510) 1000 12 (0.48)
Type 1041 1260" (1130) 750 12 (0.48) Adhesive-activated finish; fabrics
1680" (1510) 1000 12 (0.48)
Type 2000 930 (840) 1000 6.6 (0.26) Standard finish; high tenacity for
1110 (1000) 1000 8 (0.31) ballistic applications. 930 dtex
1680 (1510) 1000 12 (0.48) fiber is 'microfilament'.
Twaron HM (Akzo, 1991)
Type 1055 1210 (1090) 750 11.5 (0.45) Standard finish; multipurpose
1610 (1450) 1000 11.5 (0.45)
2420 (2180) 1500R 11.5 (0.45)
3220 (2900) 2000R 11.5 (0.45)
4830 (4350) 3000R 11.5 (0.45)
6440 (5800) 4000R 11.5 (0.45)
8050 (7245) 5000R 11.5 (0.45)
Type 1056 1210 (1090) 750 11.5 (0.45) Very low finish level; composites
1610 (1450) 1000 11.5 (0.45)
2420 (2180) 1500R 11.5 (0.45)
6440 (5800) 4000R 11.5 (0.45)
8050 (7245) 5000R 11.5 (0.45) Continued on next page
220 Organic fibers

Table 10.6 Continued

Product Count Filament Comments/typical applications

(reference)
number/ diameter
dtex (den) yarn J,lm (10-3 in)

Twaron IM (Akzo, 1991)

Type 1111 420" (380) 250 12 (0.48) Easily removed finish; fiber optic cable
1260 (1130) 750 12 (0.48) reinforcements, ballistics, composites
1680 (1510) 1000 12 (0.48)
2520 (2270) 1500R 12 (0.48)
Armos (Kaiser VIAM, 1993a)
588 (530) Twisted 48 tim; multipurpose
SVM(Kaiser VIAM, 1993b-j)
63 (57) Type A11ubricating finish
143 (130) Lubricating finish on 'acidic' fiber
294 (265) Lubricating finish on 'neutral' fiber
294 (265) Lubricating finish on 'acidic' fiber
294 (265) Type A11ubricating finish on 'acidic' fiber
588 (530) Two different heat treatments available
588 (530) Type Al lubricating finish on 'acidic' fiber
Technora f (Teijin, 1989; Mahn 1993)
T-200 1110 (1000) 666 12 (0.48) Rubber reinforcement
1670 (1500) 1000 12 (0.48)
T-202 440 (400) 1667 12 (0.48) Rubber reinforcement, pre-activated type
1670 (1500) 1000 12 (0.48)
T-220 1110 (1000) 666 12 (0.48) Rope, cable, and cord
1670 (1500) 1000 12 (0.48)
T-221 1110 (1000) 666 12 (0.48) Rope, cable, and cord
1670 (1500) 1000 12 (0.48)
T-230 1670 (1500) 1000 12 (0.48) Fiber-reinforced plastics, rope
T-240 60 (55) 36 12 (0.48) Woven and knitted fabrics, fiber-reinforced
110 (100) 67 12 (0.48) plastics
220 (200) 133 12 (0.48)
440 (400) 267 12 (0.48)
1110 (1000) 666 12 (0.48)
T-241 1670 (1500) 1000 12 (0.48) Woven and knitted fabrics, fiber-reinforced
8330 (7500) 5000R 12 (0.48) plastics
T-360 60s (55) 36 12 (0.48) 'Spunnized' yam (made up of long but not
220 (200) 67 12 (0.48) continuous filaments) for protective
440 (400) 133 12 (0.48) clothing and other fabric applications
1110 (1000) 267 12 (0.48)
1670 (1500) 1000 12 (0.48)
T-370 220 (200) 133 12 (0.48) High tenacity 'spunnized' yam for
440 (400) 267 12 (0.48) reinforcement of rubber, etc.
For footnotes see next page
 Aramid fibers 221

Table 10.7 Sources of information on commercial aramid fibers

Product Information source

Armos and SVM Kaiser VIAM; 880 Doolittle Drive, San Leandro, CA 94577, USA
fibers
Kevlar fibers DuPont Fibers; P.o. Box 80705, Wilmington, DE 19880-0705, USA, (800)
4-KEVLAR
Technora fibers Teijin Limited, 11, l-chome, Minamihonmachi, Chuo-ku, Osaka 541, Japan
Teijin America Inc; 10 East 50th Street, New York, NY 10022, USA
Twaron fibers Akzo, Aramide Maatschappij v.o.f., P.O. Box 9300, 6800 SB Arnhem,
Westervoortsedijk 73, The Netherlands
Akzo Fibers Inc., 80l-F Blacklawn Rd., Conyers, GA 30207, USA

The outstanding toughness of aramids is Aramids are strong UV absorbers and dete-
often the reason they are used over cheaper, riorate when exposed to ultraviolet light.
stiffer or even stronger fibers. Unlike glass and Protective coatings or the self-screening ability
carbon composites, aramid composites loaded in of the fiber should be used to avoid deterio-
compression, flexure or shear fail in a non-brittle ration.
manner, with significant work being required to Aramid fibers are opaque and thus the pen-
fail the composite. Their fatigue resistance is also etration of resin into the fiber bundles cannot
excellent. If other concerns such as cost or stiff- be determined visually for a aramid compos-
ness preclude the use of aramid composites, ite as it can for those made with glass fibers.
aramids are often used as a hybrid with another In fabric applications the weave used is
fiber to improve the toughness of the composite. important to the resulting properties. The
The poor off-axis and compressive proper- same is true for sandwich construction. In
ties of aramid fibers must be considered in any these cases, the fiber, fabric, or honeycomb
design. However, because of their high strength supplier can provide design assistance.
in axial tension and their toughness, aramid The choice of resin system for use with
fibers are often outstanding in applications aramid fibers is an important one. Epoxy resins
such as pressure vessels where the loading is give better translation of fiber properties than
almost totally in longitudinal tension. do polyesters, producing better shear strength
Aramid fibers absorb moisture. Where and flexural properties, but lower impact resis-
either the physical swelling of the fiber or the tance. Vmyl ester resins give both good shear
amount of moisture absorbed is of significant strength and impact resistance. Thermoplastic
concern, one of the lower absorption aramids, matrices are also used, particularly in chopped
such as Kevlar 149, Armos, or Technora should fiber composites, because of their improved
be considered. impact resistance over thermosets. However,

Footnotes for Table 10.6

• All availabilities are subject to market conditions and should be verified with the manufacturer.
b All Kevlar fibers are supplied untwisted.
, The 'R' indicates that this fiber is a 'roving,' meaning in this case that it is composed of more than one 'end' of yam.
d Twaron fibers are normally supplied untwisted. In some circumstances twist may be supplied on special request.
e Under development.
f Technora fibers are supplied with twist as requested. Finishes are supplied as requested or as is appropriate to the
application. For special applications, Technora fibers can be supplied in larger than 12 ~ filament diameters.
g These fibers, and others 'spun' yams (composed of discontinuous filaments) are normally measured by '(English)
cotton count' (ECC) rather than dtex or denier, where ECC = 5315/ denier.
222 Organic fibers

for thermoplastics the penetration of the resin Many of these are not as structural composites.
into the fiber bundle and the quality of the For example, aramids are used in many rope
fiber-matrix bond is almost always of concern. and cable applications. In mooring ropes to
Because aramids are very tough fibers, they secure oil tankers and to anchor off-shore oil
are somewhat difficult to cut and their com- platforms, the lighter weight compared to
posites can be difficult to machine. Special steel makes the aramid ropes much easier to
shears and other tools are available for cutting handle. In addition, they do not corrode, are
aramids and many successful machining tech- easier to maintain and have an extension
niques have been developed. The fiber under load which is far superior to both steel
manufacturers are an excellent source of infor- and other organic fibers.
mation in this area. Aramids are widely used to reinforce
As with all high performance fibers, aramids mechanical rubber goods. The largest volume
should be handled with care before and during of such usage is in pneumatic tires, where
processing. Rough handling will damage any aramids are lighter than steel and offer higher
high performance fiber. In addition, because of strength and modulus than other organic
their sensitivity to ultraviolet light, aramids fibers. Significant usage is also seen in belts
should be protected from such exposure. The and hoses. The excellent fatigue and creep
fibers also should not be exposed to excessive resistance of aramids are important factors in
moisture prior to processing. If the fiber is to be their usage in these applications. Corrosion
twisted, braided, or woven, it is preferable to resistance and electrical resistivity may also be
condition the fiber for one to two days at room important. Aramids are also used in athletic
temperature and intermediate moisture content shoes and in rubberized sheet materials as
prior to processing (DuPont, 1993h). However, used in aircraft evacuation slides and life rafts.
if the fiber is to be resin-impregnated and In some cases, non-composite applications
processed directly into a composite, so long as have led to composite uses. For example,
fiber handling is careful, superior properties aramids have long been used in soft body
may be attained by drying the fiber prior to armor, where the fibers absorb and disperse
processing. This is because of improved bond- bullet impact energy to other fibers in the fab-
ing of resin to the filament surfaces. ric weave. This application has now seen a
Aramid fibers present minimal safety or derivative usage in rigid composite ballistic
environmental concerns. In lifelong animal armor, composite helmets and composite spall
inhalation studies with Kevlar fibers, no liners. In these applications the toughness, RF
health effects were observed at any workplace transparency and fire and corrosion resistance
levels. Nonetheless, as with any textile fiber, of aramid fibers were significant factors in
inhalation of fibrous particles should be their selection.
avoided. Extensive animal and human skin In spite of significantly higher fiber costs
patch tests with Kevlar fibers have shown no than glass, aramids are used in canoes, kayaks,
sensitivity and little irritation, and rat feeding racing shells and small boats where maximiz-
studies have shown oral toxicity to be very ing strength and minimizing weight are
low. Combustion by-products are similar to important. Aramids offer weight savings for
wool. Aramid yarns are also essentially inert superior speed and better handling and/or
in the environment (DuPont, 1993h). improved range and fuel economy. Toughness
and overall durability and vibrational damp-
ing are also superior with aramids. The
10.2.8 APPLICATIONS
superior properties of aramids allow boats to
Aramid fibers are used in numerous applica- be built at an overall cost only 10-15% higher
tions, some of which are listed in Table 10.6. than with glass fibers and with superior perfor-
 Extended chain polyethylene fibers 223

mance (DuPont, 1983). These same properties properties but they also have limitations that
have led to the use of aramids in skis. must be considered in design.
Their high strength-to-weight ratio com- Commercially available high strength, high
bined with outstanding toughness has led to modulus polyethylene fibers include Spectra®
numerous applications of aramids in aero- fibers from Allied-Signal Corporation,
space. In both civilian and military aircraft, the Dyneema® SK60 from Dyneema Vof, Tekmilon®
toughness of aramids - and resulting resis- from Mitsui Petrochemicals and a new, as yet
tance to damage from impacts ranging from unnamed, fiber from Hoechst Celanese.
bird strikes to shrapnel - insures their contin-
ued usage. Engine nacelles and the tail cone on
10.3.2 MANUFACTURE
the McDonnell Douglas DC-9-80 are made
from Kevlar composites and approximately The traditional method of producing fibers
10% of the empty airframe weight of De from polyethylene is to spin them from a poly-
Havilland Aircraft's DASH-8 turboprop com- mer melt. This technique yields fibers
muter aircraft is Kevlar composite. Aramid composed of folded-chain crystalline regions
composites are also widely used in rotorcraft with non-crystalline regions interspersed. With
and other vertical lift aircraft. extraordinary means, the modulus of the
absolute best of such fibers can be brought to
about 80 GPa (11.5 Msi). It was long recognized,
10.2.9 CONCLUSIONS
however, that if polyethylene could somehow
Although composites of other fibers have now be produced with extended chain crystallinity,
supplanted aramid composites as having the a very high modulus fiber would result. [The
highest specific strengths, aramids still offer theoretical modulus for polyethylene is
combinations of properties not available with 320 GPa (46 Msi) (Adams and Eby, 1987).]
any other fiber. For example, aramids offer Following earlier work by Pennings, in the
high specific strength, toughness and creep late 1970s Smith and Lemstra of DSM (The
resistance, combined with moderate cost. Netherlands) developed a process with com-
However, the applications of aramid compos- mercial potential which yielded a highly
ites continue to be limited by their poor oriented extended-chain polyethylene fiber
compressive and off-axis properties and in (Hongu and Phillips, 1990). At the same time,
some applications, their tendency to absorb both Toyobo Inc. of Japan and Allied Chemical
water. Nonetheless, aramids will continue to Company in the USA were working on a sim-
be the fiber of choice where properties such as ilar approach. DSM, however, was the first to
outstanding impact resistance combined with patent the process and both Toyobo and Allied
creep resistance are critical. judged it impossible to circumvent the basic
patent filed by DSM. Thus, both companies
entered into technical association with DSM to
10.3 EXTENDED CHAIN POLYETHYLENE
produce polyethylene fibers. Toyobo Inc.
FIBERS
linked with DSM to form the joint venture -
Dyneema Vof - to produce and market the
10.3.1 OVERVIEW
new fiber. In the USA, Allied-Signal is licensed
High performance polyethylene fibers, with from DSM/Stamicarbon to produce and mar-
outstanding strength-to-weight and stiffness- ket a similar fiber.
to-weight performance, show promise in The process which is used to produce most
various specialized applications. While such commercial high strength, high modulus poly-
fibers are not as widely known as aramid and ethylene fibers is called gel spinning, the name
carbon fibers, they possess many superior derived from the gel-like appearance of the
224 Organic fibers

as-spun and quenched fibers. Ultra-high mol-

ecular weight linear polyethylene is dissolved
in a volatile solvent to form a dilute isotropic
solution that is then spun through a spinneret
and quenched in cold water to form a gel pre-
cursor fiber. Following solvent extraction, this
fiber is then hot-drawn to a very high draw
ratio ("" 30), yielding a very highly oriented,
highly crystalline, lightweight fiber
(Dyneema, 1987; Jaffe, 1989; Ward and
McIntyre, 1986; Yang, 1992).
Another approach to producing a high
strength polyethylene fiber is melt extrusion
followed by multiple stage drawing of a much
lower molecular weight polyethylene. The
modulus of an experimental fiber of this type,
220 GPa (32 Msi), is the highest ever achieved
for polyethylene (Adams and Eby, 1987). The
new polyethylene fiber from Hoechst Celanese
is the only commercial version of such a fiber.
This fiber has only about 50% of the strength
and 75% of the modulus of gel spun fibers. In
this case, the expense of dealing with a volatile
n
and potentially toxic solvent is avoided, low- (a) (b)
ering the overall price of the fiber significantly.
Fig. 10.8 Schematic illustrating the difference
between (a) conventional polyethylene fibers and
10.3.3 STRUCTURE (b) gel-spun extended chain fibers.

Figure 10.8 illustrates the difference between

conventional polyethylene fibers and gel-spun hydrogen bonds nor strong covalent bonds
or melt-extruded and drawn extended-chain between them. They are, in fact, held together
fibers. Figure 1O.8(a) is a schematic of a con- by weak dispersion-type van der Waals bonds
ventional melt-drawn polyethylene fiber. The which have a distinct effect upon properties.
fiber consists of folded chain crystallites,
mostly oriented in the draw direction, which
10.3.4 PROPERTIES
are joined to one another by tie molecules and
have between them interspersed non-crys- Polyethylene fibers offer a unique combina-
talline material. Figure 10.8(b) is a schematic of tion of properties: low specific gravity, high
a gel-spun and hot-drawn extended-chain specific modulus, high specific strength, high
polyethylene fiber. Such fibers show minimal energy to break, high abrasion resistance,
chain folding, high crystallinity and a very excellent chemical resistance, good ultraviolet
high degree of axial orientation (>95%). resistance and low moisture absorption. They
Since these fibers are based on polyethyl- have outstanding anti-ballistic and vibrational
ene, they have a density of only two-thirds damping characteristics, as well as a low
that of aramid fibers and about one-half that of dielectric constant. However there are trade-
carbon fibers. However, the polyethylene crys- offs involved in the use of polyethylene fibers.
tallites have neither relatively strong They are limited to fairly low use temperatures,
 Extended chain polyethylene fibers 225

they produce composites with poor off-axis treated Spectra 900 fiber in an epoxy matrix
and compressive properties and have poor was found to be -9 x 10--{j;oe (-5 x 10--{j ;OF) in
creep resistance. the axial direction and 100 x 1O--{j jOe (56 x
As with aramid fibers, the anisotropy of 1O--{j ;OF) in the transverse direction. The axial
their microstructure gives polyethylene fibers thermal expansion coefficient of a similar com-
anisotropic mechanical, thermal and physical posite of Spectra 1000 fiber was -10 x 10--{j;oe
properties which can be used to advantage in (-5.6 x 10--{j ;OF) and the transverse coefficient
some applications. was 105 x 10--{j;oe (58 x 10--{j;OF) (Allied, 1989).
Polyethylene fibers are the only high perfor-
mance fibers with a specific gravity of less than
Physical and thermal properties
1 and thus are the only fibers that float. Their
Polyethylene fibers have a relatively low melt- density is about two-thirds that of aramid
ing point [147°e (297°F)] and thus a low use fibers and about half that of carbon fibers.
temperature. In general, polyethylene fibers Polyethylene fibers will bum slowly if ignited,
are limited to use below 1000 e (212°F). They decomposing into carbon dioxide and water.
will, however, tolerate brief exposure (30 min The filament diameters of commercial poly-
or less) at temperatures near the melting point ethylene fibers are relatively large, typically
without major property loss (Dyneema, 1987; 23-38 !lm (0.91-1.50 x 10-3 in), although the
Weedon and Tam, 1986). diameter of Mitsui's Tekmilon monofilament
As would be expected from the lower melt- fibers can be as large as 121 !lm (4.76 x 10-3 in).
ing temperature, the properties of polyethylene The filament cross-section is typically irregu-
fibers are much more sensitive to temperature lar and somewhat elliptical.
than are aramids. Like aramid fibers,
polyethylene fibers contract with temperature
Mechanical properties
in the axial direction, while expanding in the
transverse direction. The thermal expansion Gel-spun polyethylene fibers offer some
coefficient of a composite of 60 vol% plasma- tremendous advantages over other fibers. As

Table 10.8 Axial tensile mechanical properties of representative high performance polyethylene fibers

Fiber Reference Fiber type Specific Tensile Tensile Elongation

gravity modulus, strength, at break, %
CPa (Msi) MPa (ksi)
Dyneema Dyneema, 1987 gel-spun 0.97 87 2620
SK60 (12.7) (380)
Hoechst Hoechst melt- 0.96 55 1300 4
Celanese fiber Celanese, 1993 extruded (8.0) (189)
Spectra 900 Allied,1993 gel-spun 0.97 86-103 2080-2400 3.6-3.7
(12.5-14.9) (300-350)
Spectra 1000 Allied,1993 gel-spun 0.97 128-171 2740-3000 2.8-3.1
(18.6-24.8) (397-435)
Tekmilon Mitsui,1989 gel-spun 0.96 59-98 1470-3430 4--6
monofilament (8.6-14.2) (213-498)
Tekmilon Mitsui, 1989 gel-spun 0.96 88.3 2450 3
multifilament (12.8) (356)
226 Organic fibers

can be seen in Fig. 10.4, these fibers offer very very poorly to a matrix. Although gas plasma
high specific stiffnesses and specific strengths, surface treatment can improve the interfacial
equivalent or superior to all of the aramid bond strength significantly, polyethylene fiber
fibers and to most of the carbon fibers. This composites will still have poor off-axis proper-
superior performance is offered at a lower ties.
price than that of competitive fibers. Table 10.9 gives mechanical properties for
Table 10.8 compares the mechanical proper- Spectra fiber composites, including those
ties of representative commercially available made from plasma-treated fibers.
polyethylene fibers. In spite their weak transverse strength, but
Like aramid fibers and for similar reasons, because of the non-stick nature of polyethylene
polyethylene fibers have poor compressive and thus its low coefficient of friction, polyeth-
and off-axis properties. Since the fiber is held ylene fibers perform much better than aramids
together internally by only very weak van der in abrasion resistance and polyethylene fabrics
Waals bonds, the transverse strength of the are much less easily damaged than are those of
fiber is even worse than that for the aramids. aramid fibers. The abrasion resistance of poly-
In addition, the inertness of the polyethylene ethylene fibers can be up to ten times that of
fiber means that the untreated fiber bonds aramids (Dyneema, 1987) and can be improved

Table 10.9 Mechanical properties of Spectra polyethylene fiber composites' (Allied, 1989)

Matrix: Bisphenol A based epoxy

Spectra 900 Spectra 900 PT' Spectra 1000 Spectra 100 PT
Axial Tensile Properties:
Volume percent fiber 58 50 53 54
Modulus, GPa (Msi) 27 ±1 24±1 50 50 ±3
(4.0 ± 0.1) (3.5 ± 0.1) (7.3) (7.2 ±0.5)
Strength, MPa (ksi) 552 ±90 676 ± 103 1034 ±228 889 ±55
(80 ± 13) (98 ± 15) (150 ± 33) (129 ± 8)
Elongation, % 3.6 ±0.2 2.1 ±0.4
Axial Compressive Properties:
Volume percent fiber 70 70 55 65
Modulus, GPa (Msi) 32±5 40±5 19 ±6 54±3
(4.7±0.7) (5.8 ± 0.7) (2.7 ± 0.9) (7.8 ± 0.4)
Strength, MPa (ksi) 52±2 59 ± 1 72±3 69 ±1
(7.5 ± 0.3) (8.6 ± 0.2) (10.5 ± 0.4) (10.0 ± 0.2)
Elongation, % 3.8 ±0.5 3.8 ±0.2
Flexural Properties:
Volume percent fiber 58 54 54 53
Modulus, GPa (Msi) 22±1 30±1 23±1 38±3
(3.2 ± 0.2) (4.3 ± 0.2) (3.3 ± 0.2) (5.5 ± 0.5)
Strength, MPa (ksi) 145±7 200±7 159 ±7 214 ±7
(21 ± 1) (29 ± 1) (23 ± 1) (31 ± 1)
Short Beam Shear Properties:
Volume percent fiber 58 54 54 53
Strength, MPa (ksi) 8.3±0.7 28.3 ±0.7 9.0 ±0.7 21 ±3
(1.2 ± 0.1) (4.1 ± 0.1) (1.3 ± 0.1) (3.1 ± 0.4)
• Numbers of specimens tested and criteria for limits not specified.
b PT indicates a fiber with gas plasma surface treatment.
 Extended chain polyethylene fibers 227

even further by the use of lubricants. exposure is 20-30%. The loss in both modulus
Because of their high strength, polyethylene and strength are reduced if annealing is per-
fibers exhibit very high energy to break. On a formed under tensile loading.
per-weight basis, the impact energy absorp- Unlike aramid fibers and their composites,
tion of polyethylene composites is superior to polyethylene fibers and composites show very
that of all other fiber composites. little or no loss of properties, axial or off-axis,
Polyethylene fibers are more affected by when exposed to moisture.
temperature than are higher melting point Creep resistance of extended-chain polyeth-
fibers. The loss in modulus as function of tem- ylene is of concern. Because of its low melting
perature is shown in Pig. 10.9 for Tekmilon temperature, the resistance of the fiber to creep,
multifilament fiber and Spectra fibers. Pig. 10.10 even at room temperature, is less than ideal.
shows the loss in strength as a function of tem- This is significant, since the creep of carbon,
perature for Tekmilon multifilament, Spectra glass and aramid fibers is minimal. Spectra
900 and Spectra 1000 fibers. Because of their 1000 is a stabilized' version of the fiber, which
I

very high specific strength at room tempera- shows better creep resistance than the Spectra
ture, however, polyethylene fibers still 900 fiber. Figure 10.11 shows the creep
outperform most other fibers to about 100°C response of the two Spectra fibers at room, ele-
(212°P). vated and low temperatures. At low load levels
Room temperature strength retention of at room temperature and/ or at low tempera-
polyethylene fibers following annealing at tures the creep encountered is not severe,
temperatures of up to 125°C (260 0 P) is excel- especially for the Spectra 1000 fiber, but at
lent, while modulus loss following such higher loads or temperatures the creep is much

Temperature. OF
50 100 150 200 250 300

15
100

0
11.
75 'iii
c.:l 10 :I

...
..:!
iii
:l
:l :;
"0
0
50 "0
0
:I :I

5
2S

OL....L..L....L..L-l...L-l...L-l...L-l....I.-.L-.l.-.L-.I-I-..........L..I.............L...L..................... O
o 25 50 75 100 125 150
Temperature. ·C

Fig. 10.9 Modulus as a function of temperature for Spectra 900, Spectra 1000, and Tekmilon multifilament
polyethylene fibers (Prevorsek, 1989; Mitsui, 1989).
228 Organic fibers

Temperature, Dr
50 100 150 200 250 300
3500 500

3000
400
2500

c
a.. 300 ~

-.c
2000
-
:::IE
.c

-
I:J)
I:J)
c c
I! 1500 I!
en 200 t;

1000
100
500

0 0
0 50
25 75 100 125 150
Temperature, DC
Fig. 10.10 Strength as a function of temperature for Spectra 900, Spectra 1000, and Tekm.ilon multifilament
polyethylene ftbers (Allied, 1991e; Mitsui, 1989).

Spectra 900
10 Spectra 1000

RT. 30" Load

--- --
- ... ----
5

o ~~~~~~~~~~~~~~~~~

o 10 20 30 40 50 60 70 80 90 100
time, h
(a)

Fig. 10.11 Creep of Spectra extended chain polyethylene fibers (a) at room temperature and 10% of static
ultimate and at room temperature and 30% of static ultimate.
 Extended chain polyethylene fibers 229

Spectra 900
Spectra 1000

5 °c, 20~ Load

------..- --- ---- ---

75 100 125 150 175

time, h
(b)

100
70 °C, 275 MPa

75
N
Spectra 900

-
c
'0
L
Spectra 1000
I /)

c..
50

u
"f

25

----_ ... ---- ------

-_ ... -... -- --_ ... --
Ol..L.i.L..U....I..L.l..L.l...I...Ww....I....L...L.J...L-l..JU-L.I..LL....L.L.L...I...I-............J...&...L..L.W
o 10 20 30 40 50 60 70 80

time, h
(e)

Fig. 10.11 (Continued) Creep of Spectra extended chain polyethylene fibers (b) at SoC (41°F) and 20% of
static ultimate, and (c) at 70°C (160°F) and 275 MPa (40 ksi), which is 18% of static ultimate for Spectra 900
and 11% of static ultimate for Spectra 1000 (Allied, 1991a, 1991b, 1991c, 1991d).
230 Organic fibers

more significant. The creep of polyethylene ment Gaffe, 1989). It shows superior chemical
fibers does not preclude their use in applica- resistance to PPTA in hydrochloric, nitric and
tions such as sailcloth or structural sulfuric acid (Dyneema, 1987). Table to.10 com-
reinforcement, but does require that the creep pares the chemical resistance of Spectra
demands of an application be carefully evalu- polyethylene fiber to that of aramid fiber.
ated. Because of their relatively poor creep Polyethylene fibers also show good resis-
resistance, polyethylene fibers are often tance to UV exposure. After 100 hours UV
hybridized with other, more creep resistant exposure in a fadeometer, Dyneema SK60
fibers in applications where prolonged loading retained 70% of its original strength and after
is anticipated. 1500 hours, retained 25% strength. This latter
The fatigue resistance of polyethylene fibers exposure is equivalent to about 2 years of out-
is excellent. In one test of loading and unload- door exposure (Dyneema, 1987).
ing of ropes, polyethylene fiber ropes Polyethylene fibers are hydrophobic and
withstood approximately eight times the cycles thus absorb very little moisture. The moisture
before break as aramid fiber ropes (Weedon and regain of polyethylene fibers is less than 1%.
Tam, 1986). This indicates a superiority in ten- Their weatherability is excellent: after 600
sile fatigue even to aramid fibers, which are hours exposure in a Weatherometer, Tekmilon
known for their excellent fatigue resistance. fiber retained 80% of its strength and 90% of its
modulus. Following similar exposure, an
aramid fiber had only 40% strength retention
Chemical and environmental properties
(Mitsui, 1989). Because of their excellent chem-
Polyethylene is inert. It is stable in almost all ical and moisture resistance, articles made
organic solvents and in a variety of other chem- with polyethylene fibers can be cleaned in
icals. It is also biologically inert. It is the best of soap and water.
all high modulus fibers in an alkaline environ-

Table 10.10 Comparison of strength retention after chemical immersion for polyethylene and aramid
fibers (Allied, 1989)

Strength retention, %
Spectra Aramid
6 months 2 years 6 months 2 years
Chemical (4380 h) (17500 h) (4380 h) (17500 h)
Sea water 100 100 100 98
Hydraulic fluid 100 100 100 87
Kerosene 100 100 100 97
Gasoline 100 100 93 a

Toluene 100 100 72

Glacial acetic acid 100 100 82
1M hydrochloric acid 100 100 40
5M sodium hydroxide 100 100 42
Ammonium hydroxide (29%) 100 100 70 a

Perchloroethylene 100 100 75 a

10% detergent solution 100 100 91

Chlorine bleach 91 73 0 0
a Too weak to test
 Extended chain polyethylene fibers 231

Electrical and optical properties bonding of the fiber to matrices. Through

1992, Allied-Signal marketed plasma-treated
Polyethylene fibers are electrically non-con-
Spectra fibers. However, polyethylene surface
ductive. The dielectric constant of Spectra
treatments are available after fiber purchase
fiber is 2.2, with a loss tangent of 2 x 1Q-4
from specialty companies. The fiber manufac-
(Allied, 1993). This compares to dielectric con-
turers can suggest sources of these services.
stants for aramid fibers of 3.85, quartz fibers of
The inertness and high abrasion resistance
3.78 and E-glass fiber of 6.31 (Allied, 1989).
of polyethylene fibers means that little or no
They are white in color and are transparent to
finish is required. This is fortuitous since adhe-
X-rays, radar and sonar.
sion to polyethylene is so difficult that finishes
generally will not stay on the fibers. All Spectra
fibers are supplied with 'low percent process
10.3.5 TREATMENTS
finishes' whose purpose is simply to aid in
Because of their chemical inertness, polyethyl- holding the filaments together in the bundle.
ene fibers bond poorly to matrices, with Like aramid fibers, polyethylene fibers may
consequent negative effects upon the mechan- be twisted, particularly for applications such as
ical properties of their composites. Surface marine cables. For bare fiber, twist initially
treatment by acid etch, plasma etch or corona increases the strength, although it decreases
discharge can significantly improve the the modulus of the fiber. For Spectra fiber, the

Table 10.11 Availability of commercial polyethylene fibers

Product Count Filament Comments

(reference)
number/ diameter
dtex (den) yarn ~m (10"3 in)
Dyneema SK60 444 (400)
(Dyneema, 1987) 888 (800)
1780 (1600)
Hoechst Celanese fiber 100 (90)b
(Adams, 1993) 200 (180)
Spectra 900 722 (650) 60 38 (1.50) All Spectra fibers are supplied
(Allied,1990,1993) 1333 (1200) 120 38 (1.50) untwisted and with only sufficient
5333 (4800) 480 38 (1.50) finish to hold the fiber bundle
together
Spectra 1000 239 (215) 60 23 (0.91)
(Allied, 1990, 1993) 417 (375) 60 30 (1.18)
722 (650) 120 28 (1.10)
1444 (1300) 240 28 (1.10)
Tekmilon (Mitsui, 1989)
Monofilament 5.6 (5) 1 27 (1.06)
22.2 (20) 1 54 (2.13)
111 (100) 1 121 (4.76)
Multifilament 555 (500) 50 ..38 (..1.5)
1110 (1000) 100 ..38 (",1.5)
• All availabilities are subject to market conditions and should be verified with the manufacturer.
b This fiber is newly commercially available. It will be supplied in multiples of 100 dtex.
232 Organic fibers

strength of the bare fiber is optimized at a twist Pricing

multiplier of 3, but at this value there is also
Spectra fibers are currently priced from about
about a 20% loss in modulus. As with aramids,
$15/lb for larger denier fibers to about $45/lb
however, twist is not desirable in fibers used in
for the small denier fibers. Dyneema fibers are
composites. Polyethylene fiber manufacturers
not marketed in the USA. Because of the strong
provide most of their fibers untwisted.
yen, at the time of writing Tekmilon fibers are
significantly more expensive than Spectra
10.3.6 FORMSANDAVAILABILIlY fibers in the USA. The new melt-extruded
polyethylene fiber from Hoechst Celanese is
Table 10.11 lists the availability of most of the
designed to offer good properties at a lower
c~mmercial high performance polyethylene
cost than the gel spun fibers. It costs 15-50%
fibers. In most cases, each denier of fiber is
less than the comparable Spectra fibers.
available in only one package size and weight.
The mechanical properties of different deniers
may vary somewhat from the nominal values 10.3.7 DESIGN CONSIDERATIONS
given in Table 10.9.
Like aramid fibers, polyethylene fibers are
In addition to the fiber forms listed in Table
mainly used in applications where they offer
10.11, Dyneema SK60 is available in various
a unique combination of properties. This
fabrics. Spectra fibers are also available as
prop~~ty combination includes outstanding
chopped fiber in lengths from 6 to 20 mm (0.25
speCIfIc strengths and stiffnesses, high
to 0.8 in) and in fabrics of various weaves.
toughness, outstanding abrasion resistance
Tekmilon is available as monofilament, multi-
and very low density.
filament and tape.
In any design, the relatively low melting
Information about the availability of the
~oint and low use temperature of polyethylene
fibers shown in Table 10.11; information about
fib~rs as well as the relatively poor creep
other products and about special formulations
reSIstance must be considered. The creep resis-
can be obtained from the sources given in
tance is improved in a fiber such as
Table 10.12.

Table 10.12 Sources of information on commercial polyethylene fibers

Product Information source

Dyneema fibers Dyneema Vof, Dr. Nolenslaan 119A, PO Box 599,6130 AN Sittard, The
Netherlands
Dyneema Japan Ltd., 2-8, Dojima Harna 2-chome, Kita-ku, Osaka 530,
Japan

Hoechst Celanese's Hoechst Celanese Corporation, PO Box 32414, Charlotte, NC 28232-2414, USA
high performance
polyethylene fiber
Spectra fibers Allied Fibers, Allied-Signal Inc., High Performance Fibers Technical
C~ter, PO Box 31, Petersburg, VA 23804, USA
Tekmilon fibers Mitsui Petrochemical Industries Ltd., Advanced Materials and Products
Department, Kasumigaseki Bldg., 2-5, Kasumigaseki 3-chome, Chiyoda-
ku, Tokyo 100, Japan
Mitsui Petrochemicals (America), Ltd., 1000 Louisiana, Suite 5690,
Houston,
TX 77002, USA
 Extended chain polyethylene fibers 233

Allied-Signal's Spectra 1000, which has signifi- Polyethylene fibers can be damaged by
cantly improved creep resistance compared to rough handling and should be handled with
Spectra 900. Polyethylene fibers are often care before and during processing. They pre-
hybridized with other more creep-resistant sent minimal safety or environmental concerns
fibers in applications where prolonged loading and most are biocompatible, offering another
is anticipated. Without hybridization polyeth- potential design advantage.
ylene fibers must be limited to applications
where long term, high load level, or elevated
10.3.8 APPLICATIONS
temperature loading is not anticipated, or
where creep is otherwise not of concern. Most of the current applications of polyethyl-
As with aramids, the poor off-axis and com- ene fibers are not in structural composites.
pressive properties of polyethylene fiber One of the main uses is in ropes and cables,
composites may also be concern. This is not a particularly in marine and off-shore applica-
problem, of course, in applications where the tions. The fibers are used because of their high
loading is mainly in axial tension. strength, outstanding abrasion resistance (up
The X-ray and radar transparency of the to ten times that of aramids), low density
fiber can present significant design advan- (since they float), good UV stability, resistance
tages. The UV resistance of the fiber is very to seawater and high durability. The fibers are
good and thus does not present a design prob- also used as marine sewing threads.
lem in most applications. In another major application, UV- and
Because of its chemical inertness, polyethyl- water-resistance are again important. Both
ene fibers are almost impossible to dye, Spectra and Dyneema fabrics are used, typi-
although color can be added during the fiber cally with a film coating, in sails. Unlike
spinning process. As with aramids, polyethyl- aramids, polyethylene sail can be folded and
ene fibers are optically opaque, so resin repacked numerous times without damage.
penetration within a composite cannot be This latter quality also led to the selection of
determined visually. Spectra fabric for the anchor balloon of the
Polyethylene fibers can be used with a vari- Hilton Earthwinds round-the-world balloon
ety of resins, iI)cluding polyurethanes, flight project and to their usage in lightweight,
epoxies, vinyl esters, polyesters and thermo- durable backpacks.
plastics, so long as the composite can be Polyethylene fabrics are used as filter
processed below 120°C (2S0 0 P). Polyesters are cloths, where the excellent chemical resistance
economical, vinylester resin systems provide is a tremendous advantage. Spectra fabric has
outstanding impact properties and epoxies been used in oil containment and recovery
give better translation of structural properties. systems following the Persian Gulf war. The
The preferred thermosetting matrix cure tem- fabric is treated so that water passes through it
perature is 93-104°C (200-2200P) (Allied, but oils and other floating pollution do not.
1990). The fiber manufacturers can be very Because of their excellent biocompatibility,
helpful in choosing an appropriate resin sys- polyethylene fibers are used as sutures and as
tem for an application. artificial ligaments. They are also used in sur-
As with aramids, polyethylene fiber com- gical gloves, because of their excellent cut
posites are relatively hard to machine and resistance, biocompatibility and low absorp-
producing a smooth final machined surface tion of fluids. While the poor temperature
requires special techniques. Machining tech- resistance rules out the use of polyethylene
niques developed for aramid composites can fibers in thermal protection, they are used in
be used successfully, as can hot knife or hot industrial protective clothing and in ballistic
wire cutting (Allied, 1989). protection and impact shields, with or without
234 Organic fibers

a matrix. Allied makes a special non-woven CONCLUSIONS

polyethylene fabric called Spectra Shield®,
which has alternating unidirectional layers For many applications extended-chain poly-
held together by a polymer matrix and gives ethylene fibers are superior to all other fibers,
outstanding performance in such applications. particularly when properties such as tough-
As composites, polyethylene fibers and fab- ness, dielectric constant, and/or hydrolytic
rics are used in boat hulls, water skis, stability are of concern. However, because of
sailboards, canoes and kayaks. They have their relatively low melting temperature, poly-
been explored for sporting goods ranging ethylene fibers must be limited to moderate
from archery bows to ski poles. In all these temperatures and to applications where the
examples the excellent impact resistance they creep response is acceptable.
impart to their composites is another signifi- For sail-cloth, marine rope, pressure vessels
cant advantage. In applications such as skis and other applications where the service tem-
and tennis rackets, as well as speaker cones, perature is not a governing factor,
the excellent vibrational damping capability is polyethylene fibers can be expected to make
also an advantage. However, in spite of this serious inroads into or even dominate their
superior performance, the much higher cost of respective industries. However, they cannot
polyethylene compared to glass has limited supplant aramid, carbon, or glass fibers in
the usage in boat hulls and sporting goods to applications where elevated service tempera-
high end applications such as racing competi- ture or creep resistance are critical.
tions where performance and/or safety are
more important than cost. In some cases, how- 10.4 OTHER ORGANIC FIBERS
ever, the addition of polyethylene fiber
actually lowers the overall cost of the product. 10.4.1 AROMATIC POLYESTER FIBERS
One such application is in wrapping ice
hockey sticks with Spectra fiber to improve Aromatic polyester fibers are prepared by
their durability. In this case, the additional cost spinning from a liquid crystalline melt fol-
to wrap the stick is more than offset by the lowed by heat treatment to form high strength,
longer life achieved. high modulus fibers. While many such fibers
Polyethylene fibers are also used to rein- have been synthesized since the late 1970s, the
force rubbers and elastomers. In many of only fiber commercially available in the USA
these applications, their excellent vibrational today is Vectran® fiber from Hoechst Celanese
damping characteristics are important. Inc. The general structural formula of this fiber
However, the temperature sensitivity limits is shown in Fig. 10.12.
the usage in tires to off-road vehicles Vectran was developed in the 1970s in
(Dyneema, 1987). Because of their X-ray, response to tire customers who wanted equiva-
radar and sonar transparency and low dielec- lent performance to aramid fibers but at a lower
tric constant, they have significant potential
in applications such as radomes, sonar domes
and X-ray tables.
As mentioned before, polyethylene can be
hybridized with other fibers to provide signif-
icant improvements in impact resistance.
Current hybrid applications range from bike
frames to impact shields.
Fig. 10.12 General structural formula of Vectran
aromatic polyester fiber. .
 Other organic fibers 235

cost. Hoechst Celanese was successful in pro- over a broad temperature range and excellent
viding a fiber with the desired performance, chemical resistance.
but unfortunately the resulting cost was even The density of Vectran HS is 1.41 g cm-3
higher than the aramids. Shortly thereafter, (0.0509 lbs in-3) (Hoechst Celanese, 1990). It
Hoechst Celanese stopped marketing Vectran melts at 330°C (636°P). Like aramid fibers,
fiber and marketed Vectran resin instead. Vectran HS has a negative axial coefficient of
Vectran resin quickly became the material of thermal expansion. Prom 20°C (68°P) to 145°C
choice in the electronics and computer indus- (293°P) its longitudinal linear thermal expan-
tries for small, very close tolerance connectors, sion coefficient is -4.8 x 10~ rc (-2.7 x 1O~
plugs and other components. Based upon this rp). The coefficient increases to -14.6 x 10~
success, in 1989 Hoechst Celanese reintroduced JOC (-8.1 x 1O~ rp) from 145°C (293°P) to
Vectran as a fiber product. Since the fiber is 200°C (392°P) and to -26.7 x 10~ rc (-14.8 x
more expensive, by 1.5 to 3 times, than aramids, 1~ rp) from 200°C (392°P) to 290°C (554°P)
the marketing focus is on areas where aramids (Beers and Ramirez, 1990). It has good temper-
do not meet the performance requirements ature resistance, although not as good as
(Adams and Parrow, 1993a). aramid fibers since it melts at high tempera-
Vectran is a polyester-polyarylate fiber. ture. Its shrinkage in hot air at 177°C (350 0 P) or
Unlike the aramids, Vectran melts at high tem- in boiling water is less than 0.5%.
perature. It is melt spun on conventional Vectran HS fiber is outstanding in its
polyester spinning equipment and the as-spun mechanical properties. Its axial mechanical
fibers are then heat treated in a sequence of properties are summarized in Table 10.13.
steps (Adams and Parrow, 1993b). It is the only Vectran displays no creep when tested for 2760
commercially available melt-spun liquid crys- h at 50% of its ultimate tensile strength. This
talline polymer fiber (Hoechst Celanese, 1990). behavior is significantly better than both
aramid and polyethylene fibers. Vectran also
has excellent vibrational damping characteris-
Properties
tics, better than aramids. Vectran HS has
Vectran HS offers a unique combination of superior abrasion resistance to Kevlar 29,
properties: high strength, no creep, low mois- although not as good as polyethylene fibers. In
ture absorption, negative coefficient of flexural fatigue, Vectran HS braid exhibited a
thermal expansion, good property retention 10% reduction in strength after one million

Table 10.13 Axial tensile mechanical properties of representative non-aramid, non-polyethylene organic
fibers

Fiber Reference Specific Tensile Tensile Elongation

gravity modulus strength, at break, %
CPa (Msi) MPa (ksi)
VectranHS Hoechst Celanese, 1.41 64.8 2840 3.3
1990 (9.4) (412)
Adams and Farrow, 1.41 62-86 2500-3100 2.2-2.5
1993a (9.0-12.5) (363-450)
PBO(Dow) Burk, 1993 1.56 152 5650 3.5
(22) (820)
PBO, high Burk,1993 1.56 276 5520 1.5
modulus (Dow) (40) (800)
236 Organic fibers

cycles (and maintained this strength level to glass/Vectran M products, the glass fibers are
five million cycles), while a Kevlar 29 braid producer-sized.) (Hoechst Celanese, 1990).
showed a 30% strength reduction under the More information on Vectran can be
same conditions (Beers and Ramirez, 1990). obtained from the source listed in Table 10.14.
Vectran absorbs very little water, having a
moisture regain of less than 0.1%. It is
Applications
hydrolytically stable. It has excellent chemical
resistance being resistant to organic solvents, As with aramid and polyethylene fibers, ropes
to acids at less than 90% concentration and to and cables are an important usage. Marine
bases at less than 30% concentration (Hoechst cables, fish nets, towing ropes, cargo tie
Celanese, 1990). downs, slings, sails, bicycle brake cables and
Vectran's dielectric constant is 3.3 at 1 kHz optical fiber reinforcement all have been made
(Hoechst Celanese, 1990). from Vectran HS. Olympic target archers use
Like aramid and polyethylene fibers, bow strings from Vectran HS having a propri-
Vectran is difficult to cut and its composites are etary abrasion resistant finish. The result is
difficult to machine. Typical aramid composite increased arrow speed with no creep of the
machining techniques can be used successfully. string. At the last America's Cup yacht races,
Vectran HS was used in at least six yachts,
either in sails or in marine cables. Vectran sails
Forms, availability and treatments
stretch far less than either aramid or polyeth-
Vectran is produced in the USA by Hoechst ylene and they have four to six times the life of
Celanese Corporation and in Japan by the aramid sails (Adams and Farrow, 1993a).
Kuraray Corporation under license from Safety materials and protective garments
Hoechst Celanese. It is available as Vectran have also been made of Vectran in industries
HS, a high strength reinforcement fiber, ranging from meat packing to metal working.
Vectran M, a high performance matrix fiber, In these applications, Vectran is superior to
and as engineered, commingled combinations aramids, which have poor resistance to bleach,
of Vectran HS with Vectran M and S-2 Glass and to polyethylene fibers, which are sensitive
fiber with Vectran M. to the high temperatures used in drying laun-
Vectran HS fiber is available in dtex (deniers) dered garments.
of 222 (200), 833 (750), 1000 (900) and 1667 Vectran composites have been used in aero-
(1500). These fibers are composed of 40, ISO, 180 space applications and in recreation and
and 300 filaments respectively, with the fila- leisure applications such as canoes, golf clubs,
ments being 23 !lID (0.91 x 10-3 in) in diameter. baseball bats, hockey sticks, tennis rackets,
Vectran HS fibers are offered with or without a bicycles, skis, ping pong paddles, paragliders
standard textile finish to assist in processing and stereo speaker cones. Vectran properties of
and/or to provide (dramatically improved) importance in these applications include the
abrasion resistance. Commingled Vectran HS or low moisture absorption, the excellent damp-
M fibers have no finish. (In the commingled 5-2 ing characteristics, high stiffness, lack of creep

Table 10.14 Sources of information on non-aramid, non-polyethylene organic fibers

Product Information source

Vectran fibers Hoechst Celanese Corporation, P.O. Box 32414, Charlotte, NC 28232-2414, USA
PBI fibers Hoechst Celanese Corporation, P.O. Box 32414, Charlotte, NC 28232-2414, USA
PBO fibers The Dow Chemical Company, Midland, MI 48674, USA
 Other organic fibers 237

and good flexural fatigue properties. Vectran

Nt: ~C-o-
is used where the cost of the fiber is secondary o 0
to its performance (Adams and Farrow,
1993a). -C~N N
Hybrid tennis rackets have been made by
Prince Manufacturing Company and Dunlop.
Vectran HS is combined with carbon fiber to
Fig. 10.14 Structural formula of poly-p-phenylene
give greater speed and power with vibration benzobisoxazole (PBO) (Yang, 1992).
characteristics as good as wood. Jennifer
Capriatti played with a Vectran racquet at
Wimbledon in 1992. in 1983. It has excellent chemical and solvent
Other actual or potential uses include resistance and does not burn. It is, however,
antenna guy wires, chemical resistant pack- more expensive than the aramids and has an
ings and gaskets, heat and creep resistant intrinsically high moisture absorption. It is
belting, medical and surgical equipment, pres- used mainly in woven form in fireblocking lay-
sure vessels, printed circuit board substrates ers, including aircraft seat cushions and
and aerial tow ropes. fire-fighting overgear. It was also used in
chemical warfare suits in Operation Desert
Storm. In order to reduce cost, PBI is also used
10.4.2 AROMATIC HETEROCYCLIC POLYMER
in blends with aramids for thermal protective
FIBERS
apparel. PBI has potential applications as a
Two aromatic heterocyclic polymer fibers are fiber reinforcement in composites, but cur-
currently available or in development in the rently its only composites application is as a
United States. These are PBI fiber from matrix resin or as a matrix-precursor for car-
Hoechst Celanese and PBO fiber from Dow bon-carbon composites (Yang, 1992; Conrad,
Chemical Company. 1993). More information on PBI can be
obtained from the source given in Table 10.14.
PBI fiber
PBO fiber
PBI fiber is produced from a high performance
polybenzimidazole. Chemically it is poly-2,2'- PBO fiber is a polybenzoxazole, specifically
m-phenylene-5,5'-benzimidazole, with the poly-p-phenylene benzobisoxazole, with the
structural formula shown in Fig. 10.13. The structural formula shown in Fig. 10.14. PBO
fiber was commercialized by Hoechst Celanese fiber resulted from a US Air Force program
aimed at developing high strength fibers for
advanced composites. In the late 1980s, Dow
Chemical purchased worldwide rights to the
polymer. Dow has now constructed pilot plant
facilities for monomer, polymer and fiber and
the fiber is available for evaluation in pre-pro-
duction quantities (Burk, 1993).
As with aramids, PBO fibers are spun from
a liquid crystalline solution using dry-jet wet
spinning. This is, however, a more difficult
process than for aramids. The fiber is then heat
Fig. 10.13 Structural formula of poly-2,2'-m-pheny- stretched to improve its orientation and prop-
lene-5,5'- benzimidazole (PBI) (Yang, 1992). erties (Wolfe, 1990).
238 Organic fibers

PBO is one of the most thermally and tions include composites loaded in tension,
thermo-oxidatively stable organic polymers such as pressure vessels, missile cases and ten-
known. No weight loss was observed for PBO sile beams. PBO fiber composites may also be
held at 316°C (600 0 P) (Wolfe, 1990) and weight used in non-load-bearing applications where
loss of only 0.06%/h was observed at 370°C high temperature exposure or harsh chemical
(700°F) (Burk, 1993). Its decomposition tem- environments are anticipated, such as rocket
perature is 600 0 e (1110°F) (Burk, 1993). insulation systems and brake and transmis-
Exposed to flame, PBO chars, but does not sion systems. The high strength could also
support combustion. (Wolfe, 1990) Dow's PBO lower the weight of composites used in space-
fiber has a longitudinal coefficient of thermal craft and in recreation and sporting goods.
expansion of -6 x 10-6 ;oe (-3.3 x 10-6 lOP) PBO also has significant potential application
(Burk, 1993). to ballistics, where, as a fabric or composite, it
PBO fiber has a significantly higher tensile performs equally well at half the weight of an
strength and modulus than any other known aramid. PBO composites could provide out-
organic fiber. PBO fibers have been produced standing containment systems for high speed
with tensile moduli of as high as 470 GPa rotors and turbines where high temperature
(68 Msi). Dow's current pre-production fibers exposure is of concern. The fibers also have
do not achieve these high levels, but do potential for bomb containment systems, for
nonetheless have excellent axial mechanical fire resistant and cut resistant apparel and fire
properties, as shown in Table 10.13. However, blocks, as well as ropes and cables (Burk,
like all other high performance organic fibers, 1993).
PBO fibers are quite weak in compression, More information on PBO can be obtained
with a fiber compressive strength comparable from the source given in Table 10.14.
to that of aramids (Burk, 1993). They also bond
poorly to epoxy matrices, so their off-axis
10.5 CONCLUSIONS
properties are also poor (Wolfe, 1990). For
these reasons, as with other organic fiber com- While high performance organic fibers are not
posites, PBO composites are limited to competitive with inorganic fibers in all of their
applications where structural loading is properties, they offer certain properties and
mainly in axial tension. combinations of properties that are unavail-
Moisture regain for Dow's PBO is 2.0% for able with inorganic fibers. All suffer from
the standard fiber and less than 0.5% for the certain limitations, such as poor off-axis and
high modulus version. The moisture resis- compression properties and I or temperature
tance is significantly better than aramids limitations. However, if these limitations are
(Burk, 1993). PBO is highly resistant to hydrol- properly considered, high performance
ysis, acid chemical attack, bases, solvents, organic fibers can make possible designs that
electron bombardment and laser radiation. Its can be achieved in no other way.
UV stability is outstanding (Wolfe, 1990).
Dow's PBO fiber has a lower and more stable
REFERENCES
dielectric constant than that of aramids, 3.0 at
100 kHz (Burk, 1993). Abbott, N.J., Donovan, J.G., Schoppee, M.M. and
The price for commercial PBO will be vol- Skelton, J. 1975. Some mechanical properties of
ume dependent, but will be higher than that Kevlar and other heat resistant, nonflammable
fibers, yarns, and fabrics. Technical Report
for aramids. PBO fiber will be used where AFML-TR-74-65, Part III. Wright Paterson Air
aramids and other fibers do not meet the per- Force Base: Air Force Materials Laboratory.
formance needs, particularly for strength, Adams, P.M. 1993. Private communication. 10-28-93.
modulus and flammability. Potential applica- Charlotte, NC: Hoechst Celanese Corporation.
 References 239

Adams, P.M. and Farrow, G. 1993a. Advanced fiber materials. Composites 7: 107-109.
materials for specialty applications from fully Chiao, T.T., Chiao, c.c. and Sherry, RJ. 1976.
aromatic polyesters. Presented at Textile Lifetimes of fiber composites under sustained
Research Institute, 63rd Annual Conference, 5-6 tensile loading. UCRL- 78367. Livermore, CA:
May 1993, Princeton, NJ. Lawrence Livermore Laboratory.
Adams, P.M. and Farrow, G. 1993b. Processing, Chiao, c.c. and Chiao, T.T. 1982. Aramid fibers and
properties and applications of fibers from fully composites. In Handbook of Composites, ed.
aromatic polyesters. Unpublished paper. George Lubin, pp. 272-317. New York: Van
Charlotte, NC: Hoechst Celonese Corp. Nostrand Reinhold.
Adams, P.M. Farrow, G. and Beers, D. 1995. Clements, L.L. and Moore, RL. 1977. Composite
Advanced fiber applications: properties and Properties of an Aramid Fiber in a Room-
applications of fibers from fully aromatic poly- Temperature-Curable Epoxy Matrix. SAMPE
mers. TAPPI J., 78(11), 169-174. Quarterly 9: 6-12.
Adams, W. Wade and Eby, RK. 1987. High-perfor- Conrad, D. 1993. Private communication. 7-19-93.
mance polymer fibers. MRS Bulletin 12 (12): Charlotte, NC: Hoechst Celanese.
22-26. DeCos, L. 1993. Private communication. 9-27-93.
Akzo. 1990. The aramid fiber for high-performance Conyers, GA: Akzo Fibers, Inc.
composites Twaron. Arnhem, The Netherlands: DuPont. 1981. Properties of Nomex aramid filament
Akzo Fibers and Polymers Division. yams. Bulletin NX-17. December 1981.
Akzo. 1991. Twaron product information yams, Wilmington, DE: E.I. DuPont de Nemours & Co.
fibers and pulp. Arnhem, The Netherlands: Inc.
Akzo Fibers Division, Aramid Fibers. DuPont. 1983. Kevlar aramid, the fiber of choice in
Allied. 1989 (received). Spectra high performance boat hull reinforcement. E-46814. 10/83.
fibers for reinforced composites. Undated. Wilinington, DE: E.I. DuPont de Nemours & Co.
Petersburg, VA: Allied Fibers, Allied-Signal, Inc. Inc.
Allied. 1990. Spectra high performance fibers. DuPont. 1986. Data manual for Kevlar 49 aramid.
Petersburg, VA: Allied Fibers, Allied-Signal, Inc. May 1986. Wilmington, DE: E.I. DuPont de
Allied. 1991a. Creep at 10% load (room tempera- Nemours & Co., Inc.
ture). LB006. 6/17/91. Petersburg, VA: Allied DuPont. 1989 (received). Kevlar aramid, the uncom-
Fibers, Allied-Signal, Inc. mon material for uncommon solutions.
Allied. 1991b. Creep at 30% load (room tempera- H-05500-1. Undated. Wilmington, DE: DuPont,
ture). LB003. 6/91. Petersburg, VA: Allied Fibers Department.
Fibers, Allied-Signal, Inc. DuPont. 1990. Presenting Kevlar 119 aramid fiber,
Allied. 1991c. High temperature creep. LBOOS. 6/91. for longer service life in demanding applica-
Petersburg, VA: Allied Fibers, Allied-Signal, Inc. tions, 5/90. Wilmington, DE: DuPont Fibers
Allied. 1991d. Spectra creep 20% load 5°C. LB004. Department.
6/91. Petersburg, VA: Allied Fibers, Allied- DuPont. 1992a. Kevlar aramid fiber technical guide.
Signal, Inc. 12/92. Wilmington, DE: DuPont Fibers, Kevlar
Allied. 1991e. Extended chain polyethylene tensile Products.
properties at temperature. LB007. 7/91. DuPont. 1992b. Kevlar aramid, properties and uses
Petersburg, VA: Allied Fibers, Allied-Signal, Inc. of Kevlar 29 aramid, Kevlar 49 aramid, Kevlar
Allied. 1993 (received). Spectra high performance 68 aramid in fiber optic and electromechanical
fibers, product specifications. Undated. cables. Information bulletin K- 506C, revised
Petersburg, VA: Allied Fibers, Allied-Signal, Inc. November 1992. H-37390. Wilmington, DE:
Beers, D.E. and Ramirez, J.E. 1990. Vectran high- DuPont Fibers Department.
performance fibre. J. Textile Institute 81 (4): DuPont. 1992c. Internal price list, Kevlar yam,
561-574. 11/23/92. Wilmington, DE: DuPont.
Blades, H. 1973. US Patent. 3 767 7S6. DuPont. 1992d (received). Kevlar aramid KM2, pre-
Blades, H. 1974. US Patent. 3817941. liminary information bulletin, H-35645.
Burk, W.R 1993. Private communication. 7-16-93. Undated. Wilmington, DE: DuPont Fibers,
Midland, MI: The Dow Chemical Company. Kevlar Products.
Chiao, c.c., Sherry, RJ. and Chiao, T.T. 1976. DuPont. 1993a. Prices for Kevlar 29 yarns used in
Strength retention and life of fiber composite textile processing. Price list effective 1/4/93.
240 Organic fibers

Wilmington, DE: DuPont. Kaiser VIAM. 1993a. Certificate, synthetic aramid

DuPont. 1993b. Prices for Kevlar 29 yarns used in filament, 58.8 tex. San Leandro, CA: Kaiser
ropes and cables. Price list effective 1/4/93. VIAM.
WIlmington, DE: DuPont. Kaiser VIAM. 1993b. Certificate, fiber SYM 6.3(40)-
DuPont. 1993c. Prices for Kevlar 49 yarns used in AI. San Leandro, CA: Kaiser VIAM.
textile processing. Price list effective 1/4/93. Kaiser VIAM. 1993c. Certificate, fiber SYM 14.3-Al
WI1mington, DE: DuPont. 'acidic.' San Leandro, CA: Kaiser VIAM.
DuPont. 1993d. Prices for Kevlar 49 yarns used in Kaiser VIAM. 1993d. Certificate, fiber SVM
fiber optics. Price list effective 1/4/93. 29.4(200)-A1 'N' (neutral). San Leandro, CA:
WI1mington, DE: DuPont. Kaiser VIAM.
DuPont. 1993e. Prices for Kevlar 49 yarns used in Kaiser VIAM. 1993e. Certificate, fiber SYM-K 29.4
ropes and cables. Price list effective 1/4/93. 'acidic.' San Leandro, CA: Kaiser VIAM.
WIlmington, DE: DuPont. Kaiser VIAM. 1993f. Certificate, fiber SYM 29.4-A1
DuPont. 1993f. Rope and cordage products. H- 'acidic.' San Leandro, CA: Kaiser VIAM.
37399. 2/93. Wilmington, DE: DuPont Fibers Kaiser VIAM. 1993g. Certificate, fiber TOW SYM
Department. DuPont. 58.8(300)X17- 1000; VTV heat treatment. San
DuPont. 1993g. Properties of DuPont industrial fil- Leandro, CA: Kaiser VIAM.
ament yarns, DuPont nylon, Cordura nylon, Kaiser VIAM. 1993h. Certificate, fiber TOW SYM
Dacron polyester, Nomex aramid, Teflon fluoro- 58.8(300)X17- 1000; TOSN heat treatment. San
carbon, Kevlar aramid. Technical Information, Leandro, CA: Kaiser VIAM.
Multifiber Bulletin X-273, April 1993. A-90240. Kaiser VIAM. 1993i. Certificate, fiber SYM 58.8
Wilmington, DE: DuPont, Fibers Department. 'acidic.' San Leandro, CA: Kaiser VIAM.
DuPont. Kaiser VIAM. 1993j. Certificate, fiber SYM 6.3(40)-
DuPont. 1993h. Properties and processing of AI. San Leandro, CA: Kaiser VIAM.
DuPont Kevlar Aramid Yam for Mechanical Kwolek, S.L. 1971. U.S. Patent 3 600360.
Rubber Goods. Technical Information, Kevlar Kwolek, S.L. 1972. US Patent 3 671 542.
Bulletin K-10, June 1993. Wilmington, DE: Kwolek, S.L. 1974. US Patent 3 819 587.
DuPont Fibers Department. Mahn, Harry M. 1993. Director R&D. Private com-
DuPont. 1993i (received). The second generation of munication. 9-27-93. New York: Teijin America,
ballistic protection, new Kevlar 129. H-13653. Inc.
Undated. Wilmington, DE: DuPont. Mitsui. 1989. High-performance fiber material
Dyneema. 1987. Dyneema SK60, high Tekmilon. 7/89. Tokoyo, Japan: Mitsui
strength/high modulus fiber, properties and Petrochemical Industries, Ltd.
applications. Sittard, The Netherlands: Penn, Lynn and Larsen, Fred. 1979. Physiochemical
Dyneema Vof. properties of Kevlar 49 fiber. J. Appl. Polym. Sci.
Gerzeski, Roger H. 1989. Vniivlon/polyamidobenz- 23: 59-73.
imidazole - USSR's aramid fiber forming Perkins, G. 1993. Private communication. Conyers,
polymer. In Reference Book for Composites GA: Akzo Fibers, Inc.
Technology, Vol. 1, ed. Stuart M. Lee, pp. Prevorsek, D.C. 1989. Ultrahigh modulus/strength
271-325. Lancaster, PA: Technomic Publishing polyethylene fibers and composites. In Reference
Company. Book for Composites Technology, Vol. 1, ed. Stuart
Hoechst Celanese. 1990. Vectran liquid crystalline M. Lee, pp. 167-174. Lancaster, PA: Technomic
polymer fiber. Charlotte, NC: Hoechst Celanese Publishing Company.
Corporation. Teijin. 1989. High tenacity aramid fibre Technora.
Hoechst Celanese. 1993 (received). Typical proper- Technical Information TIE-05/89.11. Osaka,
ties of high modulus polyethylene. Charlotte, Japan: Teijin Limited.
NC: Hoechst Celanese Corporation. Teijin. 1991. Heat resistant aramid fiber
Hongu, Tatsuya and Phillips, Glyn O. 1990. New TeijinconeX®. Technical Information CN02/91.2.
Fibers,. London: Ellis Horwood Ltd. Osaka, Japan: Teijin Limited.
Jaffe, M. 1989. High-modulus high-strength organic Teijin. 1993 (received). Technora - a para aramid
fibers. In Concise Encyclopedia of Composite copolymer fiber. Bulletin No. L 1.0. Undated.
Materials, ed. Anthony Kelly, pp. 129-134. Osaka, Japan: Teijin Limited.
Oxford: Pergamon Press. Ward, I.M. and McIntyre. 1986. High-modulus
 Appendix 241

fibers. In Encyclopedia of Materials Science and as the weight in grams of 1000 meters of the
Engineering, Vol. 3, ed. Michael B. Bever, pp. material. A related term is decitex (dtex), 0.1
2139-40. Cambridge, MA: MIT Press. tex, which is often used in order to be com-
Weedon, G.c. and Tam, T.Y. 1986. New PE fibers for
high-end composites. Modern Plastics.
parable to the US quantity 'denier.'
63(3):64-68. Equilibrium moisture content: moisture
Wolfe, J.E 1990. Polybenzothiazoles and polyben- absorbed by a fiber after it has been dried at
zoxazoles. In Encyclopedia of Polymer Science and 50De (122°P) for 2 h and then equilibrated at
Technology, Vol. 11, ed. J.L. Kroschwite, pp. 20 c e (68°P) and 55% relative humidity
601-635. New York: John Wiley. Strength retention: percent of room tempera-
Wu, E.M. 1980. Strength degradation of aramid- ture strength retained following exposure
fiber/epoxy composites. AMMRC-TR-80-19.
to the conditions indicated
Watertown, MA: Army Materials and
Mechanics Research Center. Tenacity: the ultimate failure strength of a fiber
Yang, H.H. 1992. Kevlar Aramid Fiber. New York: per unit original area per unit weight. The
Wiley-Interscience. most commonly used units are 'grams per
Yang, H.H. 1989. Aromatic High-Strength Fibers. New denier' (gpd) and 'newtons per tex'
York: Wiley-Interscience. (N/tex).

10.6 APPENDIX Conversion factors

DEFINITIONS AND CONVERSION FACTORS Fiber size:

1 tex = 9 denier = 10-6 kg/m
Definitions Twist:
1 tpi (turns per inch) = 39.37 tim (turns per
Denier: Term in common usage in the fiber meter)
industry in the USA to describe the fineness Twist multiplier = [tim (dtex)1I2]/3000 = [tpi
of a fiber or fiber bundle. The denier is (denier)1I2/73]
defined as the weight in grams of 9000 Modulus, stress, strength, and tenacity:
meters of the material. This is also known as 1 kgf/mm2 =9.806550 MPa
the 'count'. Its inverse measure is the 1 ksi =6.894757 MPa
'yield', expressed in yards per pound or 1 cN/tex =0.01 N/tex =10 Pt MPa =1.45 Pt ksi
meters per kilogram. where Pt is the specific gravity of the fiber
Tex: Term in common usage in the fiber indus- 1 gpd =8.826 cN/tex = 88.26 Pt MPa = 12.8 Pt
try outside the USA to describe the fineness ksi
of a fiber or fiber bundle. The tex is defined
PARTICULATE FILLERS 11
Harry S. Katz

11.1 INTRODUCTION 2 wt.% in the composite material. The only

exception is the inclusion of a brief discussion
The preceding chapters of this handbook dis-
~ussed fiber reinforcements. Fibers play an
of coupling agents, which serve an important
unportant role as the outstanding means for ~cti~n in a~hieving optimum physical prop-
erties In particulate filler composites.
obtaining great increases in the strength and
modulus of the matrix material. The focus of There are many good reasons for using par-
ticulate fillers in plastic, metal or ceramic
attention on fiber reinforcements has usually
blinded potential end users of composite matrices, in addition to the obvious usual
materials to the many beneficial uses of non- reduction in cost of the final product. In the
case of plastics, the addition of fillers provides
fibrous, or particulate, fillers. However,
a reduction of shrinkage during the cure of a
industrial experts are now becoming more
aware of these benefits and there will thermoset polymer system or the injection
undoubtedly be increased particulate filler molding of a thermoplastic resin. This reduced
usage in future polymers, ceramics and met- shrinkage results in important benefits such as
als. Particulate fillers can provide improved avoidance of the warpage or cracking that
materials, as compared with the unfilled may occur, especially in the case of large
matrix and can also be synergistic with a fiber molded parts. Also, the thermal conductivities
reinforcement to further improve the system of mineral fillers are usually in the order of ten
performance. times greater than the thermal conductivity of
polymers. Therefore, the filled polymer has a
A fiber may be described as a particle with
much greater thermal conductivity than the
a length-to-diameter ratio of greater than 10 to
1. A particulate filler may be described as a
unfilled resin. This provides an important
~on fibrous solid that fits the definition given
advantage in processes such as injection mold-
In the Compilation of ASTM Standard
ing, where the cycle time is often determined
Definitions, 'a solid compounding material, by the time to cool the part in the mold. The
faster cooling rate of the filled plastic will pro-
usually in finely divided form, which may be
vide cost savings due to the faster cycle time.
added in relatively high proportions to a poly-
Another advantage of the higher thermal con-
mer for technical or economic reasons.' This
ductivity is the faster dissipation of localized
chapter will not include additives such as
hot spots, which could cause thermal decom-
an~oxidants or internal mold release agents,
position of a polymer or failure of a sensitive
which serve important compounding func-
tions, but are used at low levels, usually below electronic component adjacent to the plastic.
There are many types of particulate fillersl.
They may be classified as mineral, natural or
synthetic, inorganic or organic. The mineral
Handbook of Composites. Edited by S.T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 fillers include calcium carbonate, clay,
 Common particulate fillers 243

feldspar, nepheline syenite, talc, alumina tri- 11.2.1 CALCIUM CARBONATE

hydrate, natural silicas and mica.
Calcium carbonate is one of the lowest cost
The importance of particulate fillers is read-
particulate fillers. Price varies depending on
ily apparent:
the grade and quantity purchased. During
'In 1993, about 4.9 billion pounds of fillers and 1994, the listed price for good quality standard
extenders, which included minerals, synthetic uncoated grades ranged from about $80 to
and organic based materials, were used by the $200 per ton. Special grades, such as precipi-
US plastics industry. The estimated value of tated and coated can cost over 25 cents per
these materials was $1.2 billion. Minerals rep- pound. Calcium carbonate is non-toxic, non-
resent the majority of these products. The irritating, odorless and is white in color with a
annual growth rate of synthetic fillers was esti- low refractive index. It is nonabrasive and is
mated at 7%.'2 the standard material for Mohs' 3 rating on the
Mohs' hardness scale. There is an abundant
Manufactured fillers include products such
supply of this mineral filler and it is commer-
as glass beads, metal powders and synthetic
cially available in a wide range of particle
silicas.
sizes. It exhibits a low oil absorption rating
There are many particulate organic fillers
and is readily dispersed at high loadings in
such as wood flour, carbon black, various
liquid systems. As is the case with other par-
starches, ground rice hulls, peanut shell and
ticulate fillers, its use reduces shrinkage
reclaimed rubber.
during molding and curing. Because of this
Low density particulate fillers are available
characteristic, it is a usual filler in low shrink
and these include perlite and hollow micros-
reinforced polyester sheet molding com-
pheres of glass, ceramic, or organic polymers
pounds. Even with high loadings, it provides
such as phimolic and polyvinylidene chloride.
relatively low stiffness to the composite. It is
Specific particulate fillers can give a com-
stable up to 800°C.
posite material special characteristics such as
electrical conductivity, biodegradability, ther-
mochromic, photo chromic, low surface 11.2.2 FELDSPAR AND NEPHELINE
friction or magnetic properties. Feldspar and nepheline are similar minerals
and serve the same purposes. These are anhy-
drous, alkali, aluminosilicates having excellent
11.2 COMMON PARTICULATE FILLERS
chemical resistance and are offered in coarse to
There are many commercial particulate fillers medium particle sizes. Because the index of
and the end user must take into consideration refraction of this mineral is equivalent to
many facts before deciding whether to use a polyvinyl chloride and other polymers, its use
particulate filler and which material will be best can result in retention of transparency or
for the application. The following paragraphs translucency in most polymers. Thus, less
present some information as a starting point for color pigment may be required to obtain a
possible selection of one of these materials for a given intensity. These mineral particulates
specific end use. Most of these descriptions are usually provide ease of wetting and are read-
condensations of the more complete details that ily dispersed and de-aired. They cause little to
can be found in The Handbook of Fillers for no effect on promoters, accelerators, or addi-
Plastics1• It should be noted that the descrip- tives that are used in liquid polymer systems.
tions are mainly aimed at the use of these Because of the low oil absorption values,
materials in a polymer matrix, but some of high loadings of these minerals can be added
these particulate fillers may have similar uses in to liquid polymer systems without running
metal or ceramic matrix composites. into the problem of excessively high viscosity.
244 Particulate fillers

Feldspar and nepheline syenite are not con- amorphous phase, of silica has very high sur-
sidered to be a health hazard or toxic and these face area. When added to liquid systems, this
minerals have been recommended as suitable type of product can provide a thixotropic
for use in composites that contact food. effect and reinforcing properties in cured rub-
Feldspar is obtained from white granite by bers or polymers. Flocculated varieties of the
quarrying the rock and crushing it. Nepheline colloidal form give flatting characteristics and
is the mineral derived from syenite rock and is composite properties different from the partic-
referred to as nepheline syenite. High-grade ulate colloidal phase. The synthetic forms are
deposits consisting of millions of tons are so fine in particle size that they offer little in the
mined and processed northeast of Toronto, way of hardness in the finished composites.
Canada. Natural silica products, as in the case of the
low quartz phase, come from quartzite and
tripolitic forms, which are generally termed
11.2.3 KAOLIN (CLAY)
natural, microcrystalline quartz. These are
Kaolin is the common term for the mineral usually less expensive than the synthetic, col-
kaolinite, a hydrous aluminosilicate mineral, loidal forms. They offer excellent electrical
one of a number of minerals called clays. It is insulation, good mechanical properties and
available in two primary grades: the natural lower surface area, hence better rheological
hydrous form and a calcined, anhydrous form. flow. Drawbacks are increased wear of pro-
The larger average particle sizes consist of cessing equipment and the lack of reinforcing
mixtures of blocky and platy particles. The properties.
finer fractions are almost entirely thin plates. Fused silica is manufactured by subjecting
Clay is a rock term applying to soft, earthy high purity quartzite to temperatures in excess
ores which are plastic when mixed with water. of 2200°C (4000°F). Finely divided fused silica
The hydrous forms are non-abrasive, chemi- has a very low coefficient of thermal expan-
cal resisting and have relatively high surface sion. This is desirable in composites subjected
area, which results in high viscosities in liquid to wide variations in temperatures.
polymer systems. They disperse readily in most Silica surfaces can be modified to offer
plastics, especially with the aid of dispersants organofunctionality, hydrophobicity and non-
or surfactants. In reinforced and thermoset functionality. Commercial activity in surface
plastics, they control flow properties to provide modification is limited now, but the future will
more uniform composite properties. surely see more functional particulate silica
Calcined kaolin is considerably harder and materials with the interface tailored to specific
provides improved electrical properties in ther- end uses.
mosets and thermoplastics. Many surface-
treated types of each gr;:tde are available to pro-
11.2.5 ANTIMONY OXIDE
vide outstanding water resistance, electrical
properties and reinforcement, in a variety of Antimony oxide is an opaque white mineral
systems. powder which has been used as a pigment
because of its good color and high hiding
11.2.4 SILICA power. It is no longer competitive with more
economical white pigments and its high cost
There are over 20 distinct phases of silica and and density makes it unattractive as an ordinary
each is chemically Si02 • A number of them are filler or extender. The major use now is as a
useful as fillers in plastics. These different flame retarder.
phases offer advantages and disadvantages, Synergism between antimony oxide and
for example a synthetic colloidal form, or halogens accounts for the widespread use of
 Common particulate fillers 245

antimony oxide in plastics. Antimony oxide is mal conductivity to a polymer composite. In

added at 1-15 wt.% and affects physical prop- each application, the grade and manufacturer
erties in the same way as other fine particle must be carefully selected in order to obtain
size minerals. The halogen compound has a optimum performance.
greater effect on plastic properties because There are a wide variety of commercial car-
larger quantities are added. The halogen com- bon black products. Differences include
pound is a plasticizer, which lowers modulus particle size and particle size distribution, spe-
and strength. An additive system must be cific surface area, surface structure, amount of
selected for the fabrication method to be used agglomeration, moisture content and minor
and for flammability and physical require- contaminants, such as metals, sulfur, oxygen
ments of the specific application. and hydrogen. These differences originate
from the fact that various hydrocarbon raw
materials and production methods are used to
11.2.6 ALUMINA TRIHYDRATE
produce different grades of products, which
Alumina trihydrate (Gibbsite) is a refined min- include furnace, thermal, acetylene, lampblack
eral which offers a bonus of flame retardance and channel blacks. Carbon blacks exhibit pH
and smoke suppression in most plastics appli- of 2-9 when slurried in water.
cations. It is available in different grades, which Many polymers are susceptible to pho-
contain various amounts of minor impurities todegradation upon exposure to sunlight or
and range in particle size from coarse materials, UV radiation. An effective method for reduc-
almost entirely +325 mesh, to very fine prod- ing or eliminating this effect is by adding
ucts less than 1 ~m in mean particle diameter. about 3 wt.% of an appropriate carbon black to
Alumina trihydrate is non-toxic and rela- the polymer.
tively inert chemically. In electrical There are a number of specific grades of car-
applications, it provides the benefit of arc and bon black, especially acetylene blacks or
track resistance. Since its surface has a low furnace-produced carbons, that are very effec-
sorptive capacity, trihydrate has low resin tive in providing high electrical conductivity
demand in liquid polyester, epoxy and acrylic in a polymer composites. High electrical con-
systems. Because it undergoes thermal decom- ductivity can also be obtained by use of the
position, it is not suitable for processing at many different grades of graphite flakes and
temperatures above about 200°C (428°F). The powders that are commercially available.
endothermic evolution of water at the decom-
position temperature is the basis for the fire
11.2.8 MICA
retardant property of this particulate filler.
Mica is a generic term for a family of potas-
sium alumininosilicate minerals, which
11.2.7 CARBON BLACK AND GRAPHITE
include muscovite, phlogopite, biotite and flu-
Particulate carbon is available commercially in orphlogopite. The exact composition of these
many forms and from many suppliers. The minerals can vary from mine to mine. In all
filler known as carbon black is usually cases, the mineral is readily ground and split
obtained by the pyrolysis of hydrocarbon into thin flexible sheets or flakes. Mica flake
gases or oils. Its largest end use is for filling fillers have had a high growth rate as a plastics
and reinforcement of rubber polymers in filler because they can provide a much greater
products such as automotive tires. However, it increase of matrix modulus than most other
also finds widespread applications as a filler mineral fillers. This improvement is further
and coloring agent in plastics, inks, paints and optimized when the mica surface is treated
paper. It can also provide electrical and ther- with appropriate coupling agents. 3
246 Particulate fillers

11.2.9 TALC ture initiation of failure of a polymer matrix

composite.
The talc of major interest to plastics users is a
There are many types of solid microspheri-
finely ground product consisting of thin
cal fillers, including glass, ceramics and
platelets and is white in color. Due to the platy
plastics. Commercial 'A' glass spheres are
nature of this form of talc, it is considered to be
available in a range of diameters, usually from
a reinforcing filler in many plastic applica-
30-750 11m. Incorporating these products into
tions. Its low cost qualifies it as an extender
a resin matrix may result in poor bonding
while its platiness qualifies it as a reinforcing
between the resin and the sphere. For opti~
filler. White, platy, fine, ground talcs are avail-
mum composite properties, a surface
able from a number of different suppliers.
treatment that is specifically tailored for the
Composites with platy talc exhibit higher
resin should be applied to the sphere. Various
stiffness and creep resistance at ambient and
surface treated products are commercially
elevated temperatures, when compared with
available.
equivalent composites filled with a particulate
There are many types of hollow micro-
filler such as calcium carbonate. For example:
spherical fillers. These include glass, ceramics,
polypropylene filled with a 40% loading of
polvinylidene chloride and phenolic plastic.
talc, the flexural modulus or stiffness triples
These fillers can be used to prepare polymer
from about 1380 MPa (200 000 psi) for the
composites with low densities and relatively
unfilled polypropylene to about 4140 MPa
high compressive strengths.
(600000 psi) for the talc-filled equivalent. A
40% loading of calcium carbonate only
increases the modulus to about 2760 MPa 11.3 METHODS OF PRODUCTION
(400 000 psi).
There are many different methods for produc-
High loadings of talc are usually accompa-
tion of particulate fillers. As indicated above,
nied by a reduction of the impact strength of
the mineral fillers are simply produced by min-
the polymer matrix composite. However, this
ing, grinding and then removing undesirable
can be minimized through proper selection of
components/ contaminants and separation
the particle size and size distribution, surface
into various size fractions.
treatments on the talc and resin formulation.
In many cases, the end user has a choice
between the natural mineral filler and a syn-
thetic product that can provide superior
11.2.10 MICROSPHERES
performance for specific applications. Calcium
There are two main reasons why microsphere carbonate is the most widely used filler or exten-
fillers are an excellent choice for many appli- der pigment in the plastics industry. For many
cations. The spherical shape provides the applications, the filler is manufactured by
lowest possible surface to volume ratio, which merely crushing the mineral, followed by pul-
limits the area of resin adsorption. This per- verizing and air classification into the desired
mits a relatively high loading of the spheres in particle size. However, a precipitated calcium
a fluid matrix without the resultant high vis- carbonate is usually specified where purity, uni-
cosity that would occur if the same volume of form small particle size and superior whiteness
equivalent size non-spherical particulate were is required. The two types of calcium carbonate
added. Second, the spherical shape provides a have their advantages and disadvantages.
uniform stress distribution on its surface, in Different manufacturing methods will pro-
contrast to the sharp edges of most other par- duce different size distributions of the
ticulates. Sharp edges create a localized high particulate fillers and size distribution is an
stress concentration, which can lead to prema- important factor to evaluate in the choice of a
 Methods of production 247

filler. Manufacturers of particulate fillers pre- Precipitated CaC03 is produced by heating

sent this type of data as graphs (Fig. 11.1). CaC03 to form CO2 and CaO. The latter prod-
Each of the four calcium carbonate fillers, uct is reacted with water to form calcium
Atomite, duramite, Snowflake P.E. and hydroxide, which is then recombined with the
Supermite, has quite different particle size dis- CO2 to produce CaC03 • Median particle sizes
tributions even though they are all produced are usually 0.2-2.0 /lm. This material is avail-
by a wet grinding process. The mean particle able in high purity grades for USP, FDA and
size of these products is 3, 11,5.5 and 1 in the food contact applications. A disadvantage is
order cited above and the oil absorption val- that the cost is higher than the ground mineral
ues, by the rub out method, are 15, 8, 9 and 19. products.
Note that the mean particle size of Snowflake Fumed silica is prepared by the hydrolysis
P.E. is one half that of Duramite, whereas the of silicon tetrachloride in a flame of hydrogen
oil absorption number is about the same. The and oxygen at a temperature of 1000°C or
relatively flat particle size distribution curve higher. While still molten, the primary small
of Snowflake P.E. gives it good packing char- particles of silica fuse into secondary particles
acteristics so that it provides high loading called aggregates. Particle size and surface
levels with low viscosity in liquid resins. area of the resultant fumed silica are con-
Therefore, it is a good choice for a wide variety trolled by varying the ratio of the reactants.
of polyester compounding methods, such as After passing through a coagulation and cool-
bulk molding compounds, sheet molding ing zone, the silica particles are separated from
compounds, transfer molding compounds, the HCl-containing combustion gases by
pultrusion and spray up applications. means of cyclones or filters. The residual HCl

90

80

70

60
?f.
~ SO~~----~----~'---~r---~--~--~~----~
~
(])

u::c 40

30

20

10

S040 20 10 S 3 2 O.S
Equivalent Spherical Diameter !J,m
Fig. 11.1 Size distributions of calcium carbonate fillers (a) Duramite, (b) Snowflake P.E., (c) Atomite,
(d) Supermite. (Courtesy of ECC International, Atlanta, GA.)
248 Particulate fillers

is removed by treating the silica with moist hot shape from the previous two materials. This
air. emphasizes that the choice of a particulate
Microsphere fillers are produced by a num- filler should not be made on the basis of the
ber of different methods. These include the fire generic type, but must be made after evalua-
polishing of ground particles, atomization of tion of different grades that are available from
molten materials and calcining of crushed par-
ticles. There are a number of patents on
methods for production of glass micros-
pheres4- 7• Fly ash microspheres are
commercially available. These are obtained by
beneficiation of the waste fly ash product that
is generated in large quantities by the many
utility companies that bum powdered coal.

11.4 TYPES OF PARTICULATE FILLERS

11.4.1 CHARACTERIZATION BY SHAPE

The shape of a particulate filler is an important Fig. 11.2 Montana talc. (Courtesy of Speciality
consideration in the selection of the optimum Minerals Inc.)
filler for a specific end use.

Microspheres
As described above, spheres have many
advantages over irregularly shaped particu-
lates when used as fillers in a polymer matrix.
Therefore, when particulate fillers are being
considered for use in a composite materiat the
selection process should include the evalua-
tion of microsphere fillers

Flakes Fig. 11.3 Italian talc. (Courtesy of Speciality

Minerals Inc.)
Flake or platelike (lamellar) particulates can
provide important benefits for composite
materials. In coatings, this shape of filler can
lead to a highly loaded composite that is a
good permeation barrier, since any penetrant
must go through a tortuous path within this
material. Polymer composites, with flake rein-
forcements, can exhibit planar isotropy and
higher modulus than composites with an
equivalent loading of an irregular shaped par-
ticulate filler. The crystal form of talc may be
platelike or flakes as in Montana talc (Fig. 11.2)
and Italian talc (Fig. 11.3). In contrast, the New Fig. 11.4 New York talc. (Courtesy of Speciality
York talc (Fig. 11.4) is quite different in particle Minerals Inc.)
 Types of particulate fillers 249

various sources and companies. There can be 11.4.2 CHARACTERIZATION BY FUNCTION

dramatic differences in compounding charac-
In addition to the categories or classifications
teristics and in critical properties, such as
discussed above, particulate fillers may be
composite strength or modulus, based o~ the
characterized by the function they serve.
specific grade of a particulate filler, how It ~as
been processed and whether or not there IS a
surface treatment. In addition to talc, other Fire retardants
flake or platy fillers include mica, some clays
An important type of particulate filler is the
and graphite.
solid fire retardants discussed above. The use
Mica flakes have been discussed above. A
of these fillers will continue to increase as a
high performance type of this material i~ the
means for reducing the loss of life and prop-
high aspect ratio grades. The aspect ratio of
erty due to fires in hotels and aircraft.
flakes is defined as the ratio of the flake aver-
age diameter to the thickness. The higher
aspect ratio flakes are more effective than low Lightweight fillers
aspect ratio flakes in the transfer of s~ess from
Particulate fillers can reduce the specific grav-
the relatively weak polymer matnx to the
ity of the composite material. Among the
stronger flake reinforcement. Optimizing .the
particulates for this category are perlite and
properties of mica filled polymer composItes
the many types of hollow microspherical
has involved a number of developments,
fillers discussed above. Perlite is an excellent
including surface treatment with coupling
low density filler, derived from a natural min-
agents, improvements in flake separations and
eral, which is mined, ground, expanded by a
orientations and reduction of flake damage
heat process and screened to various size frac-
during processing. The impact strength of
tions. It is a multicellular material, which has
these composites is often poor, but improve- the advantage that if some of the outer wall is
ments have been obtained by combinations of
broken, the remaining closed cell structure
mica flakes with high strength short fibers and prevents penetration by a liquid resin. Perlite
by use of polymer impact modifiers.
is a low cost filler, especially when evaluated
The predominant filler for manufacture of
in terms of cost per unit volume.
conductive polymer composites and inks is sil-
ver flake.
High density fillers
Irregular shaped particulates Particulate fillers may be used to increasing
the specific gravity of a polymer compo~ite.
Most commercial fillers may be considered as
Appropriate fillers in this category are banum
irregular in shape, even though the basic crys- sulfate, nickel and lead powders.
tal structure may be a square or block type
lattice. As mentioned above, mineral particu-
late fillers that are obtained by simple High hardness fillers
grinding and classifying into various size frac-
High hardness fillers such as alumina and
tions are essentially irregular shaped
crystalline silica may be used to increase
particulate fillers. These fillers, such as crys-
hardness and provide abrasion resistance for
talline silica or ground quartzite, have sharp
polymer matrix composites.
edges or points that result in high stresses
within a polymer composite and can lead to
initiation of premature failure of the structure.
250 Particulate fillers

High thermal conductivity fillers 11.4.3 CHARACTERIZATION BY SURFACE

PROPERTIES
Beryllium oxide powder has been used in
polymer matrices because it provides high The surface characteristics of a particulate
thermal conductivity and excellent electrical filler are important in determining the final
resistivity. However, this powder is highly properties of the composite. Surface treat-
toxic and this has severely limited its usage. ments are well established as a means for
Alternate choices for this end use have improving the processing characteristics dur-
included magnesium oxide and aluminum ing the addition of particulate fillers into
oxide fillers. As noted above, mineral fillers liquid resins and for improving the physical
have much higher thermal conductivity than properties and retention of the composite's
polymers, so most of these fillers will provide properties upon aging. Therefore, even though
a polymer matrix composite with much higher this chapter does not deal with additives, their
thermal conductivity than the polymer matrix. importance in achieving optimum physical
properties in particulate filler composites,
merits brief discussion.
Electrically conductive particulate fillers
In composites technology, the term cou-
Electrically conductive particulate fillers such pling agents has been used to deSignate
as silver flake, nickel powder and graphite chemicals that are used to treat the surface of
particulates can be used to provide polymer fillers and reinforcements in order to obtain
matrix coatings and composites with static optimum physical properties and for long-
discharge characteristics and electrical con- term retention of physical properties.
ductivity. Coupling agents are chemical molecules with
dual functionality, wherein one part of the
molecule will adhere to one surface, e.g. filler
Magnetic particulate fillers
or reinforcement, while another part of the
Ferrite or ceramic magnetic powders are used molecule provides a bond to the matrix mate-
as fillers in polymers to produce bonded mag- rial. Thus, a bonded bridge is formed between
nets. The loading of these powders ranges two different materials.
from 85-90 wt. %. Plastic magnets are available Coupling agents can provide benefits as
in either a flexible or rigid form. processing aids for polymer matrix compos-
ites. Dramatic reductions in viscosity can
occur in some highly filled liquid polymers
Low friction particulate fillers
when a relatively small quantity of an appro-
Particulate fillers such as Teflon® powder, priate coupling agent is added to the
molybdenum disulfide and graphite have formulation.
been added to plastics in order to obtain low
friction composites. Typical end uses have
Silanes
been oil-free bearings and molded plastics
inserts for sliding drawers or windows. Silanes are currently the predominant cou-
pling agents or surface treatments for
particulate fillers. Many commercial fillers are
available with various silane treatments and
each silane is designed for optimum use in a
specific type of polymer.s
 Representative properties of particulate fillers 251

Titanates and zirconates vary widely for the many different types and
grades of particulate fillers.
Titanates and zirconates produced by
Kenrich Petrochemicals Inc., Bayonne, New
Jersey, have received much attention in recent 11.5.2 OIL ABSORPTION NUMBER
years and show promise of affording some
The oil absorption number is an important
remarkable improvements in processing
characteristic of a particulate filler. An analo-
characteristics and final properties of many
gous situation exists in the paint and coatings
composites systems. These coupling agents
industry, where two ASTM specifications have
are analogous to the silanes. Each product has
been used to determine the oil absorption of
organic functionality and an inorganic back-
pigments. The objectives of these tests are 'to
bone, so that one end can interact with the
obtain information about the vehicle demand
matrix resin and the inorganic component will
of the pigment when it is used in a pigment
have an affinity for the filler or reinforcement
paste.' In ASTM D1483, linseed oil is added in
surface.
drops to the gently stirred pigment. The
amount of oil required to form a paste is used
Miscellaneous coupling agents to calculate an oil absorption value. In ASTM
D281, linseed oil is added dropwise to a small
There have been many polymer additives and
quantity of pigment, while rubbing vigorously
modified polymers that have been used to
with a spatula. The end point occurs upon
improve the properties of particulate filled
obtaining a 'stiff, putty-like paste, that does
polymer composites.9
not break or separate.' ASTM D281 refine-
Polypropylene has been grafted with
ments, which can involve the use of
acrylic acid and this modified polypropylene
appropriate liquids other than linseed oil,
has provided improved bonding to mica and
have been described for determination of oil
talc in polypropylene composites. Copolymers
absorption or packing fractions of fillers.lO In
of styrene and maleic anhydride have been
most applications, it is desirable to choose
shown to have improved mechanical proper-
fillers with low oil absorption values, since
ties in filled thermoplastics compared to a
these fillers will cause the minimum increase
polystyrene matrix. Dupont produces a line of
in viscosity of the matrix material.
fluorosurfactants, with the trade name Zonyl,
that have been recommended for use as a
polymer additive to improve wetting and 11.5.3 HARDNESS
bonding with fillers.
The particulate fillers that are commercially
available have a wide hardness range. They
11.5 REPRESENTATIVE PROPERTIES OF are usually rated by use of qualitative scale in
PARTICULATE FILLERS which the hardness of a mineral was deter-
mined by its ability to scratch or be scratched
11.5.1 PROPERTIES OF PARTICULATE FILLERS
by another mineral. The scale was set up by a
In considering the use of particulate fillers all mineralogist, Friedrich Mohs. In the Mohs' rat-
factors should be evaluated including chemi- ings, the standard hardness minerals are: talc, 1;
cal and physical properties, surface gypsum, 2; calcite, 3; fluorite, 4; apatite, 5;
characteristics, particle size and size distribu- feldspar, 6; quartz, 7; topaz, 8; alumina, 9 and
tion. Among physical properties are specific diamond, 10. The hardness can be important
gravity, bulk density, specific heat, coefficient in the choice of a particulate filler. In the case
of linear expansion, index of refraction and of a polymer matrix cast part that must be post
Mohs' hardness value. These properties can machined, it will usually be advantageous to
252 Particulate fillers

choose a soft filler such as talc or calcium car- much higher ratio of resin/ filler to fill the voids
bonate, since an alumina filler may result in before the system becomes sufficiently fluid for
severe abrasion of standard machining tools. most processing or molding procedures. Many
Also, in the extrusion or injection molding of applications will benefit from the use of a par-
polymer composites, a quartz or alumina filler ticulate filler and the proper packing of the
may cause severe abrasion of processing filler, filler blends or filler and fiber combina-
equipment or molds. tions will be useful to optimize the physical
properties of the composite.n

11.5.4 TOXICITY CONCERNS

11.6.2 SHRINKAGE OF PARTICULATE FILLED
Extremely fine particle size fillers can present PLASTICS
a health hazard and should be handled with
special precautions. During recent years, fillers An important advantage provided by the use
with a high content of fine crystalline silica of particulate fillers is the reduction in the
have been withdrawn from the market as a shrinkage of polymers, so that less warpage is
result of concern as a suspected carcinogen. obtained in large molded plastic parts.
Other particulates with well known health haz- However, designers of molds for unfilled plas-
ards are asbestos and beryllium oxide powders. tics must be careful to modify the design for
In general, most particulate fillers do not repre- use with particulate filled plastics. The lower
sent a significant hazard when handled by use shrinkage can result in parts that are out of tol-
of standard industrial safety procedures. erance, when mating structures are involved.
Also, there may be problems in efficient
removal of parts from a mold.
11.6 DESIGN CONSIDERATIONS IN THE USE
OF FILLERS
11.6.3 ORIENTATION OF FLAKE FILLERS
11.6.1 PACKING CONCEPTS
The design of molds for flake filled polymer
An understanding of packing concepts is composites must include the flow orientation
important for optimum physical properties in of the flakes. This can result in directional
most particulate filler composites. It is easy to strength problems unless injection gate shapes
visualize this concept in its simplest form, from and locations are properly designed. Another
the following considerations. When solid consideration for injection molding of these
spheres of any uniform size are placed in a con- polymer composites is that flow conditions
tainer, the spheres will occupy 62.5% of the often result in the formation of knit lines,
container volume and the interstitial void con- which may have much lower strength than the
tent will represent 27.5% of the volume. If much base material.
smaller spheres, with a size that permits them
to fit within the interstitial space of the larger
11.6.4 FILLER ADDITION REDUCES
spheres, are now placed into the container, it is
ELONGATION OF POLYMERS
obvious that the interstitial void volume will be
reduced. By appropriate choices of the sizes of The elongation at failure of a polymer will be
the large and small spheres, a void volume of greatly reduced when compounded with a
only 15% can easily be achieved with this particulate filler. This can lead to a significant
bimodal packing. The difference between 27.5% reduction of toughness or impact resistance.
voids and 15% voids can be extremely impor- This factor must be evaluated before making a
tant in compounding with liquid polymers or final choice in the use of particulate fillers and
resins, since the high void content requires a loading levels.
 References 253

11.7 TYPICAL END USES FOR PARTICULATE carbide fillers. Flexible and rigid plastic mag-
FILLERS nets contain about 90 wt. % loading of ferrite
powder. These magnets are used in many
The traditional source for particulate filled
plastics has been from plastics compounders. automotive and appliance applications, such
as door seals and refrigerator gaskets.
There are many custom compounders with
appropriate equipment such as Banbury mix- Mica-reinforced polypropylene has been
ers and twin screw extruders that are used to used in many automotive applications, such as
fan shrouds, seat backs, glove compartment
obtain a uniform blend of the filler and matrix.
moldings and inner fender linings.
Various filled polymer compounds, usually in
Among the many end products that are
the form of small pellets that have been
made by use of particulate filled composites
chopped from extruded rods, are available
are electronic components, toys, marine com-
from these compounders. In addition, large
resin producers have introduced filled grades ponents, grinding wheels, display items, tools,
housewares and cameras.
of their compounds. Among the many prod-
ucts that are commercially available are Nylon
and PBT resins filled with talc or mica from REFERENCES
GAFi Nylon filled with kaolin from Hoecht-
1. Katz, Harry,S., and Milewski, John v., Handbook
Celanese, Dupont (Minion), or Monsanto of Fillers for Plastics, New York: Van Nostrand
(Vydyne); and GE's Valox 700, which is a PBT Rheinhold (now Chapman & Hall), 1987.
resin filled with mica. GE's Valox HV7000 2. Fillers and Extenders for Plastics, Norwalk, CT:
series resins are PBT composites that are Business Communications Co., 1995.
highly loaded with mineral filler. Grade 7075 3. Canova, L.A., Effect of surface treatments in
has 68 wt.% filler. mica-filled polypropylene, Plastics Compounds,
1990, JulyI August, 38-43.
Most commercial particulate filled plastics 4. Davis, et al., US Patent 2 460 977, 1949.
have had a maximum filler loading of about 5. Searight, et al., US Patent 3138 444,1964.
40 wt. % because the physical properties have 6. Bland, et al., US Patent 3 150 947, 1964.
usually fallen dramatically when this level of 7. Schmidt, et al., US Patent 3 190 737, 1965.
filler has been exceeded. However, as the art 8. Plueddeman, E.P., Silane Coupling Agents, New
and science of compounding progresses, it York: Plenum Press, 1982.
may be anticipated that future composites will 9. Skeist, I., Handbook of Adhesives, New York: Van
Nostrand Reinhold, 1989.
effectively use much higher average loadings 10. Ferrigno, TH. in Handbook of Fillers for Plastics,
of mineral fillers. This trend is apparent from New York: Van Nostrand Reinhold (now
the high loading of minerals in GE's HV7000 Chapman & Hall), 1987, pp. 17, 19-20.
series mentioned above. Synthetic marble, 11. Milewski, J.V. and Katz, Harry 5., Handbook of
which is used to fabricate bathroom sinks, is Reinforcements for Plastics, New York: Van
usually formulated by use of polyester resins Nostrand Reinhold (now Chapman & Hall),
that have been highly loaded with calcium 1987, pp. 14-33.
SANDWICH CONSTRUCTION 12
Andrew C. Marshall

12.1 INTRODUCTION 1. a pair of thin, strong facings;

2. a thick, lightweight core to separate the fac-
This chapter covers a unique form of com-
ings and carry loads from one facing to the
posites known as 'structural sandwich
other; and
construction' .
A structural sandwich consists of three ele- 3. an attachment which is capable of transmit-
ting shear and axial loads to and from the
ments, as shown in Fig. 12.1:
core.
This chapter prOVides a general background
and a brief summary of the various materials
in common use; the design steps used to cal-
culate loads; some design details for solving
load point, edging and attachment problems;
and a few tables, charts and graphs containing
useful information for the designer. An
attempt is also made throughout the chapter
to provide suggestions and perspectives to
help a new user of sandwich structures tech-
nology to avoid some of the errors of his
predecessors.
Structural sandwich construction is one of
the first forms of composite structures to have
attained broad acceptance and usage. Virtually
all commercial airliners and helicopters and
Fig. 12.1 The elements of a sandwich structure are nearly all military air and space vehicles make
as follows: (a) two rigid, thin, high strength facings; extensive usage of sandwich construction. In
(b) one thick, low density core; and (c) an adhesive
attachment which forces the core and facings to act
recent years, most commercial space vehicles
as a continuous structure. The facings of a sandwich have also adopted this technology for many
panel act similarly to the flanges of an I-beam, components. The effectiveness of sandwich
resisting the bending loads and increasing the construction is shown in Fig. 12.2.
bending stiffness of the structure by spreading the In addition to air and space vehicles, the sys-
facings apart. However, unlike the I-beam's web, tem is commonly used in the manufacture of
the core gives continuous support to the flanges or cargo containers, relocatable shelters and air-
facings.
field surfacing, navy ship interiors, small boats
and yachts, duplicate die models and produc-
Handbook of Composites. Edited by S.T. Peters. Published tion parts in the automobile and recreational
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7
 Facing material 255

T
... _ ,- . -- '...... ~ !,: . l •

4t
T2t
-.L
Fig. 12.2 A striking example of how conversion to sandwich stiffens a structure without materially increas-
ing its weight. This example uses 1.6 mm (0.063 in) thick aluminum facings and 1/4-505237 kg/m3 (2.3
Ib/ft3) aluminum core.

vehicle industry, snow skis, display cases, resi- amounts of 'B' staged resins; high strength
dential construction materials, interior resins; tough, high peel adhesives requiring
partitions, doors, cabinets and a great many lower cure temperatures and pressures; as
other everyday items. well as the discovery of the resistance of sand-
Although the employment of sandwich wich to sonic fatigue.
design to produce lightweight or special pur-
pose load-carrying members is thought to
12.2 FACING MATERIAL
have originated as early as 1820, routine com-
mercial use of the idea did not occur until The primary function of the face sheets in a
about 110 years later. What started this sudden sandwich structure is to provide the required
acceptance was the successful commercial pro- bending and in-plane shear stiffness and to
duction of structural adhesives, starting in carry the edgewise and bending loads, as well
both UK and USA in the 1920s and 1930s. as the in-plane shear loading. In the aerospace
This early production began with the use of field, facings most commonly chosen are resin
casein glue and later urea-formaldehyde and impregnated fiberglass cloth or a laminate of
phenolics, with wood facings and cores. The unidirectional fibers (commonly called
search for better adhesives subsequently 'prepreg'), graphite prepreg, 2024 or 7075 alu-
resulted in the development of the rubber- minum alloy, titanium alloy, or any of several
phenolics and the vinyl-phenolics, which were stainless steel or refractory metal alloys. Even
suitable for use with metals. Commercial the most economical of these products repre-
adhesives such as 'Cycleweld,' (from Chrysler sents a substantial cost and customary practice
Motors), 'Plycosite,' (from US Plywood) and is to choose among them very carefully on a
'Redux' (from Bonded Structures, in Duxford, value engineering, or lowest lifetime cost,
UK) adhered well to both wood and metals basis.
and possessed rather high and predictable
strength.
12.2.1 SUITABILITY OF MATERIALS
The result was the beginning of a revolution
in bonding technology. Many further develop- When choosing facing materials (as well as the
ments followed in only a few years. They core, adhesive, or other materials) for an appli-
included improved cleaning methods for cation, it is wise to examine the less obvious
metal skins; low weight, high strength/ stiff- properties of the material, such as toughness
ness honeycomb core materials; 'B' staged or brittleness, mode of fracture, durability and
tape adhesives which could be stored for long weatherability, compatibility with rivets and
times; glass fabrics and collimated tapes bolts and other such attributes which may
preimpregnated with accurately measured directly affect the usability or success of the
256 Sandwich construction

end product, even though not directly applications such as boat hulls, large tanks and
involved in stress analysis or weight savings. airborne pallets and containers. This broaden-
An understanding of these requirements has ing usage is also prompted by its excellent
resulted in a switch from aluminum to fiber- compressive strength and modulus properties
glass skins and from fiberglass to aramid when compared to all but the aramid paper
(Nomex, from DuPont) cores for most aircraft honeycombs, which are much more expensive.
cabin interior panels. Complete information can be obtained from
the leading producer of these materials,
Baltek13, or Balsa Ecuador Lumber Company.
12.3 CORE MATERIALS

The primary function of a core in sandwich

structures is that of stabilizing the facings and Foam
carrying most of the shear loads through the
thickness. In order to perform this job effi- The use of foam as a structural core has been
ciently, the core must be as rigid and as light as and is now, extensive. Recent developments in
possible and must deliver uniformly pre- the technology of foam injection have sharply
dictable properties in the environment (such increased the use of these materials. The most
as high humidity) in which the finished part is novel of these is use of a cold-cavity die, in
to perform. which the foam is injection molded in a single
production step. A careful adjustment of the
mixing and curing reaction of the foam,
12.3.1 TYPES OF CORE MATERIALS together with the heat-sink effect of the mold
results in a part with facings which are simply
Wood
an un-foamed, higher density form of the
Several different materials are used exten- same polymer which constitutes the foamed
sively as sandwich cores. The oldest of these is core. The high productivity and modest cost of
wood, which continues to be used in many this scheme have resulted in many applica-
applications as a core for such common appli- tions in the automotive and industrial fields.
cations as doors, partitions and many other Another fast-growing form of the material is
'builder's supply' items. It is also used in the in cores for fiberglass snow skis and tennis
majority of snow skis, either flat-grain or end- rackets, in which an assembly of facings and
grain, although a few of the higher close-out details is placed in a closed cavity
performance skis employ honeycomb, foam, mold and foam injected to form both the core
or reinforced plastic cores. End-grain balsa has and the adhesive attachment to the pre-cured
broad acceptance in boat hulls up to lengths of glass fiber skins and various edge details. The
15.2 m (50 ft) or more and is still used for saving in labor over conventional assembly
replacement flooring for many older and a few methods has resulted in rapid acceptance of
new aircraft. the process and the construction of many new
The traditional advantage of the low cost of factories.
wood has been progressively eroded with the Foams can also provide special properties
passage of time and many users report diffi- such as insulation or radar transparency, when
culty in supply, even at prices higher than foam used with appropriate facing materials.
and sometimes approaching that of honey- The very low cost polystyrene foams are
comb. Even so, the ease of use and excellent used primarily in non-sandwich applications,
durability of the end product has led to sub- their role in structural parts for refrigerated
stantially increased usage, particularly of the vehicles and buildings having been largely
carefully selected grades of end-grain balsa, in taken over by the urethanes. The single major
 Core materials 257

exception to this statement lies in the extensive foams, even though this value is needed for
use of polystyrene foams as cores in several sandwich panel design. This property, even
thousand amateur-built composite aircraft. where listed, cannot be considered to be a reli-
This application was pioneered by Burt Rutan, able value. The actual value for an application
in his 'moldless construction', used in his at hand must be determined for the actual
series of high performance small aircraft and materials and conditions of use in order to be
the many similar designs offered by others in considered reliable. When a value for shear
subsequent years. strength is not available, it may be roughly
The polyvinyl chloride (PVC) foams, which estimated to be about 0.7 times the compres-
made an impact on the transport aircraft sive strength shown. Even the compressive
industry as flooring cores, have been largely strength cannot be considered to be reliable,
replaced by the more efficient high density however, as many differing methods of mea-
aramid honeycombs. suring this value are commonly used and each
The foam-in-place system of producing results in a substantially different value
sandwich structures has been used for more reported.
than 35 years, because of its simple concept.
However, users of this system have always
12.3.2 HONEYCOMB
had difficulty with the continuing problem of
producing uniform properties from one mix to Honeycomb types in common usage include
the next and in achieving uniformly high core products made from uncoated and resin-
and bond strengths to the metal or pre-cured impregnated kraft paper, various aluminum
glass fiber skins. The use of systematic incom- alloys, aramid paper and glass or carbon fiber
ing inspection, automatic mixmg and reinforced plastic in a number of cloth weaves
dispensing equipment and, in the case of criti- and resin systems. Honeycombs based on tita-
cal airframe parts, test coupons, produced nium, stainless steel and many others are used
integrally with the basic part, have all helped in lesser quantities. Most honeycomb cores are
to keep the problems under control. constructed by adhesively bonding strips of
It will be noted that Table 12.1 does not list thin material together, as shown in Fig. 12.3.
the shear strength of many of the various In the case of aramid paper honeycomb, the

~ T - -T_

HOBE Block
~'--~
HOBE Slice •

Expanded Panel _rflII!Y ~

Expansion Process of Honeycomb Manufacture

Corrugated Sheet Corrugated Block

. ~..-.
~ L _

Fig. 12.3 Most honeycomb is

Roll Corrugating Rolls produced by the expansion
w process. Actual cell shape
Corrugation Process 01
Honeycomb Manufacture
Corrugated produced by either method
Panel I
may vary greatly.
258 Sandwich construction

Table 12.1 Properties of several foam materials used as cores*

Type Compressive
Tensile strength Maximum
strength at 10% deflection service
Density (ASTM D1623) (ASTM D1621) temperature
IblJt3 kg/m 3 psi MPa psi MPa OF °C
ABS
(acrylonitrile
butadiene-styrene)
Injection molding
type pellets 40--56 641-897 2000--4000 13.8-27.6 2300-3700 15.8-25.5 176--180 80--82
Cellulois acetate
Boards and rods
(rigid, closed
cell foam) 6.0--8.0 96--128 170 1.2 125 0.86 350 177
Epoxies
Rigid closed cell, 5.0 80 51 0.35 90 0.62 350 177
precast blocks, 10.0 160 180 1.2 260 1.8 350 177
slabs, sheet 20.0 320 650 4.5 1080 7.4 350 177
Phenolics
Foam-in-phase ){-1~ 5-24 3-17 0.021-D.12 2-15 0.D14-D.1O
liquid resin 2-5 32-80 20--54 0.138-D.372 22-85 0.15-D.58 Continuous
7-10 112-160 80--130 0.552-D.896 158-300 1.09-2.07 service at 145
300
Polypropylene
Pellets 50 801 5500 37.9 7500 51.7 270 132
Polypropylene' 35.0 561 1600 11.03 2100 14.4
Polyurethaneb 1.3--3.0 21-48 15-96 0.1D-D.65 15-60 0.1D-D.41 180--250 82-121
4-8 64-128 90--290 0.62-1.99 70--275 0.48-1.90 200-250 93-131
9-12 144-192 230-450 1.58-3.10 290-550 1.99-3.79 250--275 121-135
13--18 208-288 475-700 3.28--4.83 650--1100 4.48-7.58 250-300 121-149
19-25 304-400 775-1300 5.34-8.96 1200--2000 8.27-13.8 250-300 121-149
Skinned molded
(rigid)
Skin 25-65 400--1041 100-2700 0.68-18.6 40--3000 0.28-20.7 150-250 66-121
Core 3-30 48--481 15-1500 15-1500 150-250 66-121
Polyvinyl chloride
Rigid closed cell 3 48 1000 6.90 95 0.65
and up and up
boards and billets 6 96 200 1.38
* Where shear strength and modulus properties are not shown, use a figure of 0.7 times the compressive strength shown
as a first approximation for design feasibility consideration. Always test actual material used for true value of shear
strength and modulus.
a High density, foam, molded, parts and shapes, with solid, integral skin.
b Rigid (closed cell) molded parts; boards, blocks, slabs; pipe covering; one-shot, two- and three-package systems for
foam-in-place; for spray, pour, or froth-pour techniques.
 Core materials 259

Table 12.1 Continued

Type Thermal
conductivity Shear Shear
strength modulus
BTU in
h-1Jr2 °F-l Wm- 1 K-l psi MFa psi MPa
ABS
(acrylonitrile
butadiene-styrene)
Injection molding
type pellets 0.58-2.1 0.08-0.30
Cellulose acetate
Boards and rods
(rigid, closed
cell foam) 0.31 0.04
Epoxies
Rigid closed cell, 0.26 0.04
precast blocks, 0.28 0.04
slabs, sheet 0.32 0.05
Phenolics
Foam-in-phase 0.21-0.28 0.03-Q.04
liquid resin 0.2O-Q.22 0.03-0.04
0.24-Q.28 0.03-Q.04
Polypropylene
Pellets 1.05 0.15
Polypropylene> 4.2 0.61
Polyurethaneb 0.11-0.21 0.2-0.4 20 0.14 226 1.56
0.15-0.29 0.02-0.04 90 0.62 1500 10.3
0.19-0.35 0.03-Q.05 180 1.24 4500 31.0
0.26-0.40 0.04-0.06
0.34-Q.52 0.05-0.07 450 3.1 15000 103.5
Skinned molded
(rigid)
Skin 0.12-0.80 0.02-0.12
Core 0.21-0.55 20-500 225-15000
Polyvinyl chloride
Rigid closed cell 2.0 at 70 65 0.45 1200 8.3
boards and billets 120 0.83 2200 15.2
* Where shear strength and modulus properties are not shown, use a figure of 0.7 times the compressive strength shown
as a first approximation for design feasibility consideration. Always test actual material used for true value of shear
strength and modulus.
>High density, foam, molded, parts and shapes, with solid, integral skin.
b Rigid (closed cell) molded parts; boards, blocks, slabs; pipe covering; one-shot, two- and three-package systems for
foam-in-place; for spray, pour, or froth-pour techniques.
260 Sandwich construction

inherent toughness and abuse resistance of the enced by the properties of the materials from
material makes cores of 16-48 kg/ m 3 (1-3 which they are manufactured. Some of these
lb / ft3) an excellent choice for aircraft cabin differences are obvious in the thennal conduc-
interior walls and ceilings, even with glass fab- tivity information shown in Fig. 12.4 and Fig.
ric-reinforced skins as low as 0.254 mm (O.OlD 12.5. However, several significant properties
in) in thickness. of honeycomb cores are peculiar to the materi-
Physical and mechanical properties of the als and should be separately noted.
honeycomb core materials are strongly influ-

r--"T""-....,.--.....-..--_-_--r_---. .16
Thermal Resistance - All.IT1inum Honeycomb
.028
0
u.
~1~.021
w
t---+---II--+---+--~ .12
'ii:~
,III
~ a:
« :E:
~ .014 t---t---I----i---+..,.".""-+-........,;;;!OOO""""---I---i .08
fB
a:
LLI
0
z
.04 ~
UI
m
a:
0
2.5 (1.0) 5.0(2.0) 7.6(3.0 10(4.0)
Core Thickness- em (in.)

TIIer.,..1 Reeistanc.Non Melallic Honeycomb

Fig. 12.4 Thermal conductivity .70 r--...,...---,~-_-...,.--,....-..,.:..--.._..--. 4
through sandwich panels can be 0
u.
isolated into the contribution of
each component: facings, core 3 'ii:i=
,ID
and adhesive. The resistances (or a:
:E:
reciprocal of conductivity) can 2
simply be added - including the LLI
0
Z
effect of boundary layer condi- .181--+--=i~~;j.ooII!!!:::-+--l--+--+--I ~
tions. The thermal properties of UI
il5
typical facing materials may be w
(b)
a:
found in many handbooks. 0
1.3(0.5) 2.5(1.0) 3.8(1.5) 5.0(2.0)
Thermal resistance values for Core Thickness- em (in.)
typical core to facing adhesives
are typically 0.03 for film adhe-
sives with a scrim cloth support Effect of Mean Temperature
and 0.01 for unsupported adhe- 2.0
sives. These graphs give the
resistance for aluminum and 1.5
non-metallic honeycomb at a
a
mean temperature of 23.9°C
~ 1.0
(75°F). Note that for non-metallic .~

honeycomb, it has been found ~

that the cell size is more critical 0.5
than core density. The reverse is
true with aluminum honeycomb. (c) 0
-129 -17.8 93 204
To correct for mean temperature, (·200) (0) (200) (400)
divide the resistance at 23.9°C
(75°P) by coefficient Q. Mean Temperature. 0(;, (OF)
 Core materials 261

1.2

1.1
, 140

1.0
~ ,
..,~~ ~ THESE CURVES ARE APPROXIMATE 120
.ItUJ""NIJII

\
~ .9
o ~"'l.ti~-.,.
'" ,""'--........"
.8
:~
\ - ---
--
.7
I>"'I>~~ 1'......
13 .8

---
K ~ ---
...............
w
~ .5
1' ........ TW
o
U .4
.3
'-- OR
-~-
1.0
----
o .5 1.5 2.0 2.5 3.0 3.5 4.0 IN •
(a) 0~--7---~2~~;~--~~--~~---7~--~f~~~~-.§r---t.1~CM

CORE THICKNESS
40
LEGENO:

Fig. 12.5 Measured core shear strength will vary -- 125 L8. KRAFT PAPER,
PHENOLIC RESIN,
depending upon the test method, core thickness, r- ;i'!INCH HEXAGON CELLS_

skin thickness and many other factors. The above

20
-- 3003-H19 ALUMINUM:
00024 -INCH FOIL,
curves may be used only for preliminary correction rlNCH HEXAGON CELLS

factors. Physical tests of the final design must be (b) 0 I

o I 2 3
used to confirm actual values obtained, as the CORE THICKNESS fiNCHES)
curves shown above are only approximate.

Density
All mechanical properties increase with higher the loads anticipated. Figure 12.7 shows typical
density, as shown in Fig. 12.6. differences in shear strength for the Land W
directions. In addition, some cell shapes allow
easy forming or curving at a small loss in
Cell shape
strength/weight ratio. This attribute can be of
All honeycombs are anisotropic and the result- great importance in manufacturing curved
ing directional properties should be adapted to parts of appreciable thickness.

MPa PSI
2000 HFT-G/
/
/
12 I /
/
/50561.

1600
/
/ / /5052 u.-;r:
/"
/

(/ ,/
/ /'
10 / //
/ /

1200 i /
/ //

i /~
/4"
/

V'
,//
I
/

r
/

~//-1
lalla
-= 6 100 /
/

.4
"X' /.
V
II at.r •• iltanV/
1ft 4 Pape '"/
fl"'~ "III Pitt noIiC;:::?~",,-:~" --
400 --

"-
,!,.Y~
--"
Kraft

~/";'
No.
2
.--" ,"Vi: FOG n
~
o f--= .-"-
~<"-=
2 3 4 5 6 7 • 9 10PCF
I ,
Fig. 12.6(a) Typical stabilized
I I I I ! !
o 20 40 60 10 100 120 140 160kglm'
Density compressive strengths.
262 Sandwich construction

_ PSI
100

A~ ~~AL VsOJ2
600
/
/
y.
1/'/
4 / /"
/
/

/
/
/
V /,,/

-
~ 400

f
.,./
/
~ ·SFi !-tl...~ !h0nolie
.....
--.......
/

! a••istal t Pap.,
/ ~ ~- ....,. ...
//
./
~

p .--' _..-" -~
,..:;<~~
200 ,.,~
~
.

........ ::.::.:::. .-.-

K"". .--::::::. PVC Foo~

°0 2 3 4 5 6 7 • 9 10 PCF
O~'~~2~'0~-'~~'~~~~'~~~~~~I~~~~I~~O~~I~~~u~~~
Fig. 12.6(b) Typical '1' shear
Density strengths.

..('....

~ 111111111
.-----r---,----.-----.~-_.

Lt7(7101 r---+---+---+--J~

--
A

i ..
8
(
=:
~(

! "'"(110) I---+-~",+-----,..s..o:+------l---I c o

12(. 14(", IICI) 121(IJ 110(10)

CORE DENSITY Kgrln3 (PCF) E F

Fig. 12.7 Plate shear test values may be signifi-

cantly different from test. results obtained from
testing beams, Values shown above are typical for
G H
5052 aluminum honeycomb.
Fig. 12.8 A few of the many cell configurations in
Cell shape variations common usage. B, G and H are only produced by
the corrugating method. F is a cell configuration
Cell shape variations may be either furnished nearly always used in the manufacture of welded
to specification by the core manufacturer, or, in metal honeycomb. C is flexible in one axis, while G
certain materials such as aluminum, shapes and H are flexible in both axes. A, C and 0 are
may be intentionally or inadvertently altered expanded from identical unexpanded slices, A
by the core user. It should be noted that the being normal expansion, C fully over-expanded
under- or over-expansion of the core changes and 0 50% expanded. B is a reinforced corrugated
core, with an extra layer of uncorrugated web mate-
its cell shape and density. The over-expanded rial placed between each layer of corrugated web
version of Fig. 12.8(c) changes directional material. Reinforcing layers may be added in dis-
properties such that the L direction becomes crete locations or patterns and may be of the same
slightly the weaker of the two major axes. or different web material or thickness.
 Core materials 263

Since the drop in the L direction strength can

amount to as much as 30%, such changes in
cell shape must not be allowed to occur by
error.

Cell size
Although cell size tends to be a secondary
variable for most mechanical properties of
core materials, it is primary in fixing the
strength level of the core-to-face attachment
(or, more accurately, in fixing the required
lower limit on core-to-panel adhesive weight)
and in determining stress levels at which
intracell buckling or face dimpling of facings
occurs.

Thickness
The shear and compressive properties noted
Fig. 12.9 Plate shear test for honeycomb shear
for a specific core type can only be realized strength and modulus 1.27 cm (0.50 in) thick steel
when test methods are carefully specified and plates are oven-cleaned and may be reused many
controlled and the correct thickness of core is times.
tested. Failure to allow for the effect of thick-
ness can affect observed values by a factor of 4 1

~
or more, as noted in Fig. 12.5. It should be f1'
I.

6J'--@-,-,-"'-1 III 11M) dl'

emphasized that the correction factor shown
may be considerably different, depending on
skin material and thickness, as well as the
exact test method used.

Specimen geometry and test method

Like thickness, these must be specified and
carefully controlled in order to obtain compa-
rability with test values obtained by others.
Shear strength values obtained using plate Fig. 12.10 Short beam shear test for honeycomb
shear test methods of Fig. 12.9 are quite nor- core. Note the ample bearing area provided at each
load and support point to preclude core crushing
mally up to 25% below those obtained when prior to core shear failure.
using the flexure method shown in Fig. 12.10.
Both methods are accepted and used and any
lack of understanding of the differences can
Paper honeycomb
lead to monumental, if nonsensical, problems.
It will be noted that the tables of mechanical Paper honeycomb is the first predecessor of all
properties for various honeycombs, Tables the types of honeycomb, having been pro-
12.4-12.12, specify the shear test method used duced for some 2000 years. Early forms were
in producing the data shown. not used as structural cores, but were
Table 12.2(a) Properties of 5052 alloy hexagonal aluminum honeycomb" IV
~
Compressive "i Plate shear
*' ~
;::t
~ Bare Stabilized *
..:: 'L' Direction 'w' Direction l=:I..
Honeycomb ~~ ~
designation, .5 -- Strength, Strength, Modulus, ii~ Strength, Modulus, Strength, Modulus,
E ~ i:.c
§:
cell - material - gauge ~{i psi psi ksi Uti) psi ksi psi ksi 8;::::
typical min typical min typical typical typical min typical typical min typical t:D
::r
;::
1/16 - 5052 - 0.0007 6.3 870p 910p 275p SlOp 90p 320p 40p n
.....
1/16 - 5052 - 0.001 9.0 1480x 1500x 420x 775x 105x 520x 53x c·
;::::
1/8 - 5052 - 0.0007 3.1 270 200 290 215 75 130 210 155 45.0 130 90 22.0
1/8 - 5052 - 0.001 4.5 520 375 545 405 150 260 340 285 70.0 220 168 31.0
1/8 - 5052 - 0.002 8.1 1400 1000 1470 1100 350 750 725 670 135 455 400 54.0
1/8 - 5052 - 0.003 12.0 2200p 2325p 900p llOOp 625p
5/32 - 5052 - 0.0007 2.6 200 150 215 160 55 90 165 120 37.0 100 70 19.0
5/32 - 5052 - 0.0025 8.4 1530 1070 1600 1180 370 800 760 690 140 475 420 56.0
3/16 - 5052 - 0.0007 2.0 130 90 135 100 34 60 120 80 27.0 70 46 14.3
3/16 - 5052 - 0.001 3.1 270 200 290 215 75 130 210 155 45.0 130 90 22.0
3/16 - 5052 - 0.002 5.7 770 560 810 600 220 390 460 410 90.0 300 244 38.5
3/16 - 5052 - 0.003 8.1 1400 1000 1470 1100 350 750 725 670 135 455 400 54.0
1/4 - 5052 - 0.0007 1.6 85 60 95 70 20 40 85 60 21.0 50 32 11.0
1/4 - 5052 - 0.002 4.3 480 350 505 370 140 230 320 265 66.0 210 155 29.8
1/4 - 5052 - 0.003 6.0 850 630 880 660 235 430 495 445 96.0 315 265 40.5
1/4 - 5052 - 0.004 7.9 1360 970 1420 1050 340 725 700 650 130 440 390 52.8
3/8 - 5052 - 0.0007 1.0 30 20 45 20 10 25 45 32 12.0 30 20 7.0
3/8 - 5052 - 0.001 1.6 85 60 95 70 20 40 85 60 21.0 50 32 11.0
3/8 - 5052 - 0.002 3.0 260 190 270 200 70 120 200 145 43.0 125 85 21.2
3/8 - 5052 - 0.005 6.5 970 700 1020 750 265 505 545 500 105 350 300 43.5
.. Corrugated 5052 and 5056 aluminum honeycomb is available in higher densities with crush strenghts up to 6000 psi. Test data obtained at 0.626 in thickness.
*.. Crush strength values shown are average or typical; actual values may vary because of density tolerances, etc.
p = preliminary properties;
x =predicted values
Note: contact core producer for complete information.
Table 12.2(b) Properties of 5052 alloy hexagonal aluminum honeycomb* (metric)

~ Compressive Plate shear

~ Bare Stabilized *'
..l::
* 'L'Direction 'W' Direction
Honeycomb <:l.? ..l::~
designation, .E '1J; Strength, Strength, Modulus, Strength, Modulus, Strength, Modulus,
E :o:! ~ E <:!
cell - material - gauge ~~ kPa kPa MPa 01;;~ kPa MPa kPa MPa
typical min typical min typical typical typical min typical typical min typical
1/16 - 5052 - 0.0007 101 5998p 6274p 1896p 3516p 621p
2206p 276p
1/16 - 5052 - 0.001 144 10204x 10 342x 2896x 5343x 724x
3585x 365x
1/8 - 5052 - 0.0007 50 1862 1379 1999 1482 517 896 1448
1069 310
896 621 152
1/8 - 5052 - 0.001 72 3585 2586 3758 2792 1034 1793 2344
1965 483
1517 1158 214
1/8 - 5052 - 0.002 130 9653 6895 10 135 7584 2413 5171 4999
4619 931
3137 2758 372
1/8 - 5052 - 0.003 192 15168p 16030p 6205p 7584p 4309p
5/32 - 5052 - 0.0007 42 1379 1034 1482 1103 379 621 1138
827 255 689 483 131
5/32 - 5052 - 0.0025 135 10549 7377 11 032 8136 2551 5516 4757
5240 965 3275 2896 386
3/16 - 5052 - 0.0007 32 896 621 931 689 234 414 827
552 186 483 317 99
3/16 - 5052 - 0.001 50 1862 1379 1999 1482 517 896 1448
1069 310 896 621 152
3/16 - 5052 - 0.002 91 5309 3861 5585 4137 1517 2689 3172
2827 621 2068 1682 265
3/16 - 5052 - 0.003 130 9653 6895 10135 7584 2413 5171 4999
4619 931 3137 2758 372
1/4 - 5052 - 0.0007 26 586 414 655 483 138 276 414
586 145 345 221 76
1/4 - 5052 - 0.002 69 3309 2413 3482 2551 965 1586 2206
1827 455 1448 1069 205
1/4 - 5052 - 0.003 96 5861 4344 6067 4551 1620 2965 3413
3068 662 2172 1827 279
1/4 - 5052 - 0.004 127 9377 6688 9791 7239 2344 4999 4826
4482 896 3034 2689 364
3/8 - 5052 - 0.0007 16 207 138 310 138 69 172 221
310 83 207 138 48
3/8 - 5052 - 0.001 26 586 414 655 483 138 276 586
414 145 345 221 76
3/8 - 5052 - 0.002 48 1783 1310 1862 1379 483 827 1379
1000 296 862 586 146
3/8 - 5052 - 0.005 104 6688 4826 7033 5171 1827 3482 3758
3447 724 2413 2068 300
* Corrugated 5052 and 5056 aluminum honeycomb is available in higher densities with crush strenghts up to 41000 KPa. Test data obtained at 16 rom thickness .
•• Crush strength values shown are average or typical; actual values may vary because of density tolerances, etc.
p = preliminary properties;
x = predicted values Q
Note: contact core producer for complete information. ~
~
~
@"
~.
til
N
0"-
U1
Table 12.3a Properties of several commonly used glass-reinforced plastic honeycombs" N
0'\
0'\

Compressive Plate shear (J'}

::::.
Bare Stabilized 'L'Direction 'W' Direction
Honeycomb ~
designation, Strength, Strength, Modulus, Strength, Modulus, Strength, Modulus, §:
material - cell - density psi psi lesi psi ksi psi ksi 8;:s
Hexagonal typical min typical min typical typical min typical typical min typical .....
'"
500 350 600 480 57 260 210 11.5 140 110 5.0
....
;:
HRP - 3/16 - 4.0 C'">
HRP - 3/16 - 5.5 800 600 940 750 95 425 370 19.5 220 190 8.5 .....
HRP - 3/16 - 8.0 1400 1100 1600 1280 164 660 600 34.0 400 370 15.0
o·
;:s
HRP - 3/16 -12.0 2280 1600 2300 1800 260p 940p 815 SSp 570 500 25p
HRP -1/4 - 3.5 350 260 500 400 46 230 170 9.0 120 100 3.5
HRP -1/4 - 6.5 1025 850 1180 900 120 450 25.0 260 11.0
HRP-3/8 - 2.2 150 105 200 145 13 105 75 5.0 60 45 2.0
HRP - 3/8 - 4.5 610 450 690 550 65 300 260 14.0 170 150 6.0
HRP - 3/8 - 6.0 900 750 1000 750 100 400 340 22.5 260 210 10.0
HRP - 3/8 - 8.0 1060 920 1200 150p 520 31p 320 13p
Ox-core
HRP lOX -1/4 - 4.5 520 350 625p 43p 210 8.0 250 15.2
HRP/OX-1/4-7.0 1150 1230p 84p 395p 14p 450p 20p
HRP lOX - 3/8 - 3.2 340 260 425p 32p 140 4.5p 150p 9p
HRP lOX - 3/8 - 5.5 700 580 S20p 60p 240 lOp 300p 17p
Flex-core
HRP /F35 - 2.5 180 240 25 125p 12.5p 70p 7.Op
HRP /F35 - 4.5 440 600 49 280 22.0 140 12.0
HRP IF50 - 3.5 300 425 300 37 195 140 20.0 100 75 10.0
HRP /F35 - 5.5 600p 880p 61p 390p 31.5p 205p 16.0p
,. Test data obtained at 0.500 in thickness.
p =preliminary properties.
Table 12.3b Properties of several commonly used glass-reinforced plastic honeycombs* (metric)

Compressive Plate shear

Bare Stabilized 'L' Direction 'W' Direction
Honeycomb
designation Strength, Strength, Modulus, Strength, Modulus, Strength, Modulus,
material - cell - density kPa kPa MPa kPa MPa kPa MPa
Hexagonal typical min typical min typical typical min typical typical min typical
HRP - 3/16 - 4.0 3447 2417 4137 3309 393 1793 1448 79 965 758 34
HRP - 3/16 - 5.5 5516 4137 6481 5171 655 2930 2551 134 1517 1310 59
HRP - 3/16 - 8.0 9653 7584 11 032 8825 1131 4551 4137 234 2758 2551 103
HRP - 3/16 - 12.0 15 720 11 032 15 858 12 411 1793p 6481p 5619p 379p 3930p 3447p 172p
HRP -1/4 - 3.5 2413 1793 3447 2758 317 1586 1172 62 827 689 24
HRP-1/4-6.5 7067 5861 8136 6205 827 3103 172 1793 76
HRP - 3/8 - 2.2 1034 724 1379 1000 90 724 517 34 414 310 14
HRP - 3/8 - 4.5 4206 3103 4757 3792 448 2068 1793 97 1172 1034 41
HRP - 3/8 - 6.0 6205 5171 6895 5171 689 2758 2344 155 1793 1448 69
HRP - 3/8 - 8.0 7308 6343 8274 lO34p 3585 214p 2206p 90p
Ox-core
HRP lOX -1/4 - 4.5 3585 2413 4309p 296p 1448 55 1724 105
HRP/OX -1/4-7.0 7929 8481p 579p 2723p 97p 3103p 138p
HRP lOX - 3/8 - 3.2 2344 1793 2930p 221p 965 31p 1034p 62p
HRP lOX - 3/8 - 5.5 4826 3999 5654p 414p 1655 69p 2068p 117p
Flex-core
HRP /F35 - 2.5 1241 1655 172 862p 86p 483p 48p
HRP /F35 - 4.5 3034 4137 338 1931 152 965 83
HRP /F35 - 3.5 2068 2930 2068 255 1344 965 138 689 517 69
HRP /F35 - 5.5 4137p 6067p 421p 2689p 217p 1413p 110p
• Test data obtained at 12.70 mm thickness.
p := preliminary properties,
()
<:)

~
~
;;,
~
....,
;:;'
rn
N
0\
'l
268 Sandwich construction

employed as decoration - and are still fre- Some of the above alloys are also available as
quently seen today as seasonal decorations in corrugated, corrugated and reinforced, over-
department stores in the form of expanded expanded and flexible cell configurations.
bells, spheres and so forth. Some have also been produced in a specially
Current materials used as sandwich cores tailored geometry to make all the cell axes lie
are different, in that much stronger kraft paper on a true radius of a cylinder, a sphere, or
is employed and 11-35% phenolic resin is fre- other unique configurations. These same alloy
quently used to improve mechanical foils can also be wound as a corrugated spiral
properties, as well as moisture and fungus to form a cylinder or tube for very light energy
resistance. Many variations are available in absorption applications.
cell sizes of to, 13 and 19 mm (%, ~ and %in) or The aluminum honeycomb cores remain
even larger sizes. The higher strength versions the most used, as well as the most versatile of
are only produced in the smaller cell size, with the various core materials obtainable and are
the to mm (% in) cell available as a water- often found to possess the most favorable per-
migration resistant grade meeting military formance/ cost ratio available. Expanded
specification MIL-H-2104Q. aluminum cores commercially available
Most applications are found in non-aircraft ranges from a low of about 32 kg/ m3 (2 lb / fr)
uses, where cost saving is the one primary to a high of 192 kg/m3 (12.0lb/ft3).
objective. Usage is growing rapidly in recre- Corrugated aluminum cores, however, start at
ational vehicles; for doors, walls and under 128 kg/m3 (8Ib/fr) and can be pur-
partitions; for factory produced kitchen cabi- chased up to 880 kg/m3 (55lb/ft3). At
nets; in packaged patio room additions for densities below 128 kg/m3 (8lb/fr) corru-
homes; in curtain wall panels; and in bearing gated core suffers a serious penalty in shear
walls for commercial building. properties when compared to expanded core.

Aluminum honeycomb Glass fiber-reinforced plastic honeycomb

This family of materials has been in produc- This family of materials is most commonly
tion and growing since about 1947. Aluminum used in electrically sensitive parts, such as
honeycomb now includes four alloys, at least radomes and antennae, or where a heat resis-
five cell shapes and many foil gauges to pro- tant resin and low thermal conductivity make
vide a range of densities. The alloys generally it a natural choice. It has also seen distin-
available are: guished service as a matrix for retaining
non-structural ablative materials, such as soft
• 3003-H19, the lowest strength of the group, silicone rubbers or syntactic rigid epoxy
usually used for non-aircraft applications; foams, which otherwise could not have been
• 5052-H39, the most often used aircraft used effectively as ablative heat shields on the
grade, available with a corrosion resistant Gemini and Apollo re-entry vehicles.
surface treatment. Mechanical properties Only high temperature phenolic and poly-
are listed in Table 12.2; imide cores are generally produced. They are
• 5056-H39, the strongest of the regular air- commonly available in cell sizes of 5, 6.3 and
craft grades, available with a corrosion 10 mm (%6, 74: and %in) with a 3 mm (% in) cell
resistant surface treatment; available in a bias weave glass reinforcement.
• 2024-T3 or T81, the most heat-resistant alloy Densities range from 32 to 192 kg/m3 (2 to 12
and slightly stronger in some properties lb / fr). Mechanical properties of several com-
than 5056-H39. Available with a corrosion mercially available glass fiber-reinforced cores
resistant surface treatment. are shown in Tables 12.3-12.6.
 Core materials 269

Table 12.4(a) Properties of glass-reinforced phenolic honeycomb (bias weave reinforcement)*

Compressive Plate shear

Bare Stabilized 'L'Direction 'w' Direction
Honeycomb
designation Strength, Strength, Modulus, Strength, Modulus, Strength, Modulus,
material - cell - density psi psi ksi psi ksi psi ksi
typical typical typical typical typical typical typical
HFT -1/8 - 3.0 300p 350p 22p 185p 17p 95p 7p
HFT -1/8 - 4.0 390p 575p 45p 300p 32p 150p 12p
HFT -1/8 - 5.5 525p 960p 67p 425p 42p 225p 17p
HFT -1/8 - 8.0 1450p 1625p lOOp 575p 48p 340p 25p
HFT - 3/16 -1.8 75p 120p 14p 105p 13p 50p 4p
HFT - 3/16 - 2.0 lOOp 170p 17p 115p 15p 60p 5p
HFT - 3/16 - 3.0 275p 375p 32p 200p 24p lOOp 9p
HFT - 3/16 - 4.0 435p 550p 45p 275p 30p 140p 12p
HFT lOX - 3/16 - 6.0 1000p 1100p 67p 290p 13p 335p 30p
* Test data obtained at 0.500 in thickness. Honeycomb is normally not tested for bare compressive strength.

Table 12.4(b) Properties of glass-reinforced phenolic honeycomb (bias weave reinforcement)* (metric)

Compressive Plate shear

Bare Stabilized 'L' Direction 'W' Direction
Honeycomb
designation Strength, Strength, Modulus, Strength, Modulus, Strength, Modulus,
material - cell - density kPa kPa MPa kPa MPa kPa MFa
typical typical typical typical typical typical typical
HFT -1/8 - 3.0 2068p 2413p 151p 1275p 117p 655p 48p
HFT -1/8 - 4.0 2688p 3964p 310p 2068p 220p 1034p 82p
HFT - 1/8 - 5.5 3619p 6618p 461p 2930p 289p 1551p 117p
HFT -1/8 - 8.0 9997p 11203p 689p 3964p 331p 2344 172p
HFT - 3/16 -1.8 517p 827p 97p 724p 89p 344p 27p
HFT - 3/16 - 2.0 689p 1172p 117p 792p 103p 413p 34p
HFT - 3/16 - 3.0 1896p 2585p 220p 1378p 165p 689p 62p
HFT - 3/16 - 4.0 2999p 3792p 310p 1896p 206p 965p 82p
HFT lOX - 3/16 - 6.0 6894p 7584p 461p 1999p 89p 2309p 206p
* Test data obtained at 12.70 mm thickness. Honeycomb is normally not tested for bare compressive strength.

Aramid paper honeycomb

what lower than aluminum, especially in
This is an especially tough and damage resis- modulus, but it possesses a unique ability to
tant product, based on a completely synthetic, survive overloads in local areas without per-
calendered 'Nomex' paper material produced manent damage. This translates into abuse
by DuPont. The core is expanded very much resistance when applied to very light interior
like aluminum or glass fabric honeycomb and aircraft panels or flooring and gives the mate-
then dip-coated with phenolic or other suit- rial a competitive edge even at the higher cost
able resin system. The mechanical properties it represents. The base material is relatively
of the material as a structural core are some- incombustible and the small amounts present
270 Sandwich construction

Table 12.5 HFT glass-reinforced phenolic honeycomb (Fibertruss bias weave)*

Compressive Plate shear

Bare Stabilized 'L' Direction 'W' Direction
Honeycomb
designation, Strength, Strength, Modulus, Strength, Modulus, Strength, Modulus,
material - cell - density psi psi ksi psi ksi psi ksi
typical typical typical typical typical typical typical
HFT - 1/8 - 3.0 250p 360p 21p 185p 16p 96p 6.4p
HFT - 1/8 - 4.0 460p 530p 45p 3 lOp 25p 150p 9.5p
HFT -1/8 - 5.5 850p 950p 65p 460p 34p 240p 13.5p
HFT -1/8 - 8.0 1600p 1750p 95p 600p 43p 340p 20.0p
HFT - 3/16 - 2.0 90p 140p 17p 1I8p 15p 55p 4.3p
HFT - 3/16 - 3.0 250p 320p 32p 170p 20p 90p 6.5p
HFT - 3/16 - 4.0 460p 530p 45p 310p 25p 150p 9.5p
HFT/OX - 3/16 - 6.0 1000p 1I00p 67p 290p 13p 335p 30.0p
• Test data obtained at 0.500 in thickness. p =preliminary properties.

in typical panels result in low volumes of space vehicles. In addition to this small usage,
smoke and gases given off in fire tests. Typical however, carbon fiber honeycomb is now used
applications make use of these properties very as the structural core for nacelle assemblies in
effectively. As a consequence, they have grown the Boeing Model 777 transport aircraft. The
to a commercial volume nearly as large as that constant pressure for lighter structures in such
of aluminum, for use in aircraft structures. designs has led to the use of carbon fiber fac-
Uses outside the aerospace industry are lim- ings, which have a potential corrosion
ited due to the high cost of the material, but problem when used with aluminum cores.
despite this it has seen some application in This concern for corrosion problems has sub-
boat hulls up to 10.2 m (40 ft) in length, as well sequently led to the adoption of a new class of
as in skis, racing shells and several other prod- carbon fiber honeycomb materials for this air-
ucts. craft and will possibly lead to further use in
Aramid core is normally produced in cell other future designs.
sizes of 3, 5, 6.5 and 10 mm (lil, 71'6, X and % in), Two types of carbon fiber cores are now
in densities of 24-192 kg/m3 (1.5-12Ib/ff3). being produced. One is for purely structural
Densities higher than 64 kg/m3 (4Ib/ff3) are applications, while the other has a require-
almost entirely used for aircraft flooring. ment for heat transfer through the thickness of
Mechanical properties of some of these core the panel. The former type uses only the usual
materials are shown in Table 12.6. pan based carbon fibers, while the latter
employs pitch based carbon fibers, which
duplicate the heat transfer properties of the
Carbon fiber honeycomb
aluminum core which it replaces. Although
Reinforced plastic honeycomb has for many neither of these materials is as yet in large vol-
years employed glass fabric reinforcement, but ume production, the economic impact is
only rarely employed other fibers. In the past substantiat since these honeycombs are
few years, however, both Kevlar and carbon markedly higher in price than the aluminum
fiber have become much more common as or Nomex cores they replace.
reinforcing fibers for honeycomb. Carbon Little data is yet available on these new
fibers only now are beginning to be used in cores, but it is likely they will see substantial
 Adhesive materials 271

use and public scrutiny in the next several be understood by the designer and fabricator
years. in order for the otherwise inevitable problems
to be avoided. Some factors which merit atten-
tion are discussed below.
Kevlar honeycomb
This honeycomb has been in use for a number
of years as a core for space vehicle antenna 12.4.1 PRODUCTS GIVEN OFF DURING CURE
reflectors. The purpose of the Kevlar honey- Some adhesive types, such as phenolic, give
comb is to allow transmission of radio signals off a vapor as a product of the curing reaction
through the panel, while at the same time the and the presence of these secondary materials
Kevlar facing acts as a partial reflecting antenna can lead to several problems:
for a different wavelength of a different signal.
Kevlar honeycomb, based on one of several • internal pressure, resulting in little or no
fabrics woven from Kevlar yam, is usually bond in some areas, or 'blisters';
produced in cell sizes of 6.3-9.5 rom (X-% in) . • core splitting, as the gas forces its way
Usual densities available range from 16 to through the core to a lower pressure area;
64 kg/m3 (l-4lb/ff). • core movement, sometimes several inches,
resulting in an unusable cured part;
• subsequent corrosion of core or skins by the
Kevlar paper honeycomb chemical action of the vapor or its residual
In addition to Kevlar honeycomb made from condensate.
woven fabric, DuPont has recently introduced
a new honeycomb, based on a Nomex-like 12.4.2 BONDING PRESSURE
paper, which is entirely composed of fibers
derived from Kevlar. This material has rather Adhesives such as the phenolics and some
surprising mechanical and physical proper- others actually require more than atmospheric
ties, with strengths well above both glass and pressure in order to prevent excessive poros-
Nomex honeycombs and dielectric properties ity. Certain forms may be suitable for solid
somewhat superior to Nomex. This material is cores like balsa, but cannot be used at all in
trade named 'Kortex' and is available in the open cores such as honeycomb or large cell
usual range of cell sizes and densities. foams. Also, most core materials will not alone
Because the material is somewhat more withstand compressive bonding loads exceed-
expensive than Nomex, no large scale replace- ing a few atmospheres and consequently
ment of Nomex honeycomb appears likely, cannot be used with any adhesive system
although many special purpose applications requiring higher pressures.
have been developed in both air and space
craft.
12.4.3 FILLET FORMING
In order to achieve a good attachment to an
12.4 ADHESIVE MATERIALS
open cell core, such as honeycomb, the adhe-
Adhesives, as they apply to sandwich struc- sive must have a unique combination of
tures, constitute a somewhat different family surface tension, surface wetting and controlled
of materials than those required to bond an flow during early stages of cure. Controlled
open cellular core to a stiff and continuous fac- flow prevents the adhesive from flowing
ing. Although these differences are less down the cell wall and leaving a low strength
important with some of the newer modified top skin attachment and an overweight bot-
epoxy materials, they remain basic and must tom skin attachment.
Table 12.6(a) Properties of Nomex paper honeycomb* N
tj
Compressive Plate shear C/'l
~
Honeycomb ;::l
Bare Stabilized 'L' Direction 'W' Direction ::::....
designation, 8
material - cell - density Strength, Strength, Modulus, Strength, Modulus, Strength, Modulus, ;:;.
~
(gauge) psi psi ksi psi ksi psi ksi 8;::l
Hexagonal typical min typical min typical typical min typical typical min typical til
..,.
.....
HRH 10 -1/8 -1.8 (1.5) 110 70 130 85 90 65 3.7 50 36 2.0 .:
,..,
HRH 10 -1/8 - 3.0 (2) 300 180 330 270 20 180 162 7.0 95 85 3.5 ....
o·
HRH 10 -1/8 - 4.0 (2) 500 330 560 470 28 245 225 9.2 140 110 4.7 ;::l

HRH 10 -1/8 - 5.0 (3) 900 600 925 660 325 235 175 120
HRH 10 -1/8 - 6.0 (3) 1075 800 1125 825 60 370 260 13.0 200 135 6.0
HRH 10 -1/8 - 9.0 (3) 1700 1400 1800 1600 90 520 370 17.0 270 240 9.0
HRH 10 - 5/32 - 5.0 (4) 800p 900p 360p 11.5p 180p 5.0p
HRH 10 - 3/16 - 2.0 (2) 150 90 170 105 11 110 72 4.2 55 40 2.2
HRH 10 - 3/16 - 4.0 (3) 500 320 560 470 28 245 215 7.8 140 110 4.7
HRH 10 - 3/16 - 6.0 (5) 650 580 700 650 390 330 14.5 185 150 6.0
HRH 10 -1/4-1.5 (2) 90 45 95 55 6 75 45 3.0 35 23 1.5
HRH 10 -1/4 - 3.1 (5) 275 180 285 240 170 135 7.0p 85p 60 3.0
HRH 10 -1/4 - 4.0 (5) 370 310 400 360 240 200 7.5 125 95 3.5
HRH 10 - 3/8 -1.5 (2) 90 45 95 55 6 75 45 3.0 35 23 1.5
HRH 10 - 3/8 - 3.0 (5) 285p 300p 17p 170 5.6p 95p 3.0p
OX-core
HRH 1O/0X - 3/16 -1.8 (2) 110 70 130 60 45 2.0 60 35 3.0
HRH 10/0X - 3/16 - 3.0 (2) 365 250 400 270 17 115 95 3.0 125 95 6.0
HRH 1O/0X -1/4 - 3.0 (2) 350 210 385 250 17 110 90 3.0 115 90 6.0
Flex-core
HRH 1O/F35 - 2.5 (3) 150 105 170 119 12p 70 49 4.0p 40 28 1.9p
HRH 10/F35 - 4.5 (5) 450p 490p 33p 270p 7.3p 150p 3.7p
HRH 10/F50 - 3.5 (3) 300 189 350 217 24 150 105 5.7p 80 56 2.8p
HRH 10/F50 - 5.0 (5) 550 625 525 37 330 300 8.0 190 160 4.1
• Test data obtained at 0.500 in thickness. Nomex is a registered trademark of DuPont.
p = preliminary properties
Table 12.6(b) Properties of Nomex paper honeycomb" (metric)

Compressive Plate shear

Honeycomb Bare Stabilized 'L' Direction 'W'Direction

designation, Strength, Strength, Modulus, Strength, Modulus, Strength, Modulus,
material - cell - density (gauge) kPa kPa MPa kPa MPa kPa MPa
Hexagonal typical min typical min typical typical min typical typical min typical
HRH 10 - 1/8 - 1.8 (1.5) 758 482 896 586 620 448 25 344 248 13
HRH 10 - 1/8 - 3.0 (2) 2068 1241 2275 1861 137 1241 1116 48 655 586 24
HRH 10 -1/8 - 4.0 (2) 34473 2275 3861 3240 193 1689 1551 63 965 758 32
HRH 10 -1/8 - 5.0 (3) 6205 4136 6377 4550 2240 1620 1206 827
HRH 10 -1/8 - 6.0 (3) 7411 5515 7756 5688 413 2551 1792 89 1378 930 41
HRH 10 - 1/8 - 9.0 (3) 11721 9652 12410 11 031 620 3585 2551 117 1861 1654 62
HRH 10 - 5/32 - 5.0 (4) 5515p 6205p 2482p 72 1241p 34
HRH 10 - 3/16 - 2.0 (2) 1034 620 1172 723 75 758 496 28 379 275 15
HRH 10 - 3/16 - 4.0 (3) 3447 2206 3861 3240 193 1689 1482 53 965 758 32
HRH 10 - 3/16 - 6.0 (5) 4481 3998 4826 4481 2688 2275 99 1275 1034 41
HRH 10 -1/4 -1.5 (2) 620 310 655 379 41 517 310 20 241 158 10
HRH 10 -1/4 - 3.1 (5) 1896 1241 1965 1654 1172 930 48 586 413 20
HRH 10 -1/4 - 4.0 (5) 2551 2137 2757 2482 1654 1378 51 861 655 24
HRH 10 - 3/8 - 1.5 (2) 620 310 655 379 41 517 310 20 241 158 10
HRH 10 - 3/8 - 3.0 (5) 1965p 2068p 117p 1172p 38p 655p 20p
OX-core
HRH 10/0X - 3/16 -1.8 (2) 758 482 896 413 310 13 413 241 20
HRH 10/0X - 3/16 - 3.0 (2) 2516 1723 2757 1861 11 792 655 20 861 655 42
HRH 10/0X -1/4 - 3.0 (2) 2413 1447 2654 1723 117 758 620 20 792 620 41
Flex-core
HRH 10/F35 - 2.5 (3) 1034 723 1172 820 82p 482 337 27p 275 193 13p ~
~
HRH 10/F35 - 4.5 (5) 3102p 3378p 227p 1861p 59p 1034p 25p ;::.-
HRH 10/F50 - 3.5 (3) 2068 1303 2413p 1496 16 1034 6929 39p 551 386 19p ~
HRH 10/F50 - 5.0 (5) 3792 4309 2275 ~.
3619 255 2068 55 1310 1103 28
~
* Test data obtained at 0.500 in thickness. Nomex is a registered trademark of DuPont. ~

p = preliminary properties ~
"'t
5·
S;

IV
2j
274 Sandwich construction

12.4.4 ADAPTABILITY

The requirements noted above must all be met

while also meeting all the requirements of a
skin-to-skin to skin-to-doubler attachment. In
the case of contoured parts, the adhesive must
also be a good 'gap-filler' without appreciable
strength penalty, since tolerance control of
details is much more difficult to achieve on
contoured than on flat panels and a greater
degree of latitude for misfit must usually be
allowed.

12.4.5 BOND LINE CONTROL

This is a need which exists because of misfitting

details and is approximately the opposite of
adaptability. It is the capability of the adhesive
to resist being squeezed out from between fay- Fig. 12.11 Climbing drum peel test for adequacy of
ing surfaces when excessive pressure is applied skin adhesion. The difference in diameter of the
to a local area of the part during cure. Many cylinders to which the straps are attached and the
adhesives are formulated to achieve good core cylinder to which the skin is attached causes the
filleting and are subsequently given controlled drum to rotate clockwise when tension is applied
flow by adding an open weave cloth or fibrous by the universal testing machine. This arrangement
web, cast within a thicker film of adhesive. This allows duplication of test results from one shop to
another.
'scrim cloth' then prevents the faying surfaces
from squeezing out all the adhesive, which virtue of being easily duplicated, as well as
would result in an area of low bond strength. possessing an obvious relationship to the
toughness whose value is sought. Values of
peel strength will vary considerably, depend-
12.4.6 TOUGHNESS
ingupon:
The word 'toughness' has many meanings in
the world of adhesives. Usually, it refers to the • toughness of the adhesive;
resistance shown by the adhesive to permit- • amount of adhesive used;
ting bond line cracks to grow under impact • density of the core;
loading. In the area of sandwich core-to-facing • cell size of the core;
bonds, it refers to the resistance shown by the • direction of the peel (with or across the rib-
adhesive toward loads which act to separate bon direction);
the facings from the core under either static or • adequacy of the surface preparation;
dynamic conditions. It has been found from • degradation of the adherend surface subse-
quent to bonding.
experience that greater toughness in the bond
line usually equates to greater durability and Because these variables can lead to widely dif-
thus to longer service life. fering peel strengths for the very same
Many types of tests have been devised to adhesive, all of them must be properly under-
measure toughness, but the most common one stood and controlled if the peel test is to be
used for sandwich structures is the climbing used and its value compared to other test
drum peel test (Fig. 12.11). This test has the results.
 Adhesive materials 275

The peel test is used to control quality 12.4.9 NITRILE RUBBER MODIFIED EPOXIES
throughout the sandwich industry. Values
These make up a broad group of more recent
obtained, provided the adhesive weight and
materials which provide much of the flow and
core material are in balance, will give indica-
toughness shown by the nylon--epoxies, along
tions of tooling or cure problems and of
with the durability and weather resistance of
adherend surface preparation problems. It is
the vinyl-phenolics. They are the most com-
particularly useful for this when an environ-
mon of the 'toughened' thermosetting
mental exposure involving both elevated
adhesives and are usually limited to about
temperature and high humidity is interposed
149°C (300°F) service temperature. Some of
between manufacture and test. It is also adapt-
these materials routinely achieve shear
able to use with nearly any skin material,
strengths of 34 500 kPa (5000 psi) and most
except that it becomes impractical with very
can be cured over a wide range of tempera-
thick or very stiff skins.
tures and pressures.
It can be readily seen that a number of points
of difference separate the sandwich adhesives
from other structural adhesives. Fortunately for 12.4.10 URETHANES
the sandwich user, many adhesives are avail-
able which satisfactorily meet both sets of Urethane based adhesives are used in many
requirements. The types available, along with commercial structures. Both moisture-cured
some salient features, are as follows. and two-part systems are available.

12.4.7 PHENOLICS BLENDED WITH VINYLS, 12.4.11 OTHER POLYIMIDES,

RUBBERS OR EPOXY THERMOPLASTICS AND HIGHLY
SPECIALIZED ADHESIVES
All of these families of adhesives give off at
least some water during cure and are therefore These are used in a number of applications
used only where their high strength, durabil- ranging up to about 371 °C (700°F) service tem-
ity or high temperature mechanical properties perature, but do not represent either a very
are essential. Since the out-gassing cure prod- large group of materials or a large volume of
ucts usually require venting or perforating the usage. In addition to categorizing the available
core material and a number of non-out- adhesives by chemical type, they can be
gassing, high temperature adhesives have grouped by the form in which they are avail-
become available, their use as sandwich adhe- able. Generally these are as follows.
sives has sharply declined in recent years.
Light liquids, heavy liquids, thixotropic
12.4.8 EPOXIES MODIFIED WITH NYLON OR liquids, pastes, putties, or syntactic foams
OTHERPOLYANITDEPOLYMERS
Only a few are used as a core-to-face bond, but
These adhesives were the first to have excel- many such materials are used in sandwich
lent filleting and controlled flow along with construction to splice pieces of core to each
both high strength and high toughness, other in order to provide high strength edges,
although they are somewhat moisture sensi- areas, or surfaces, or to carry shear loads from
tive. Some versions are provided as one side of fittings, inserts, or end ribs. Most of the mate-
a two-sided tape adhesive, in which the other rials so used are epoxies, modified epoxies,
side is a rubber or vinyl-phenolic, to provide epoxy polyamides or epoxy polyimides.
both excellent peel and durability at the skin Curing temperatures vary from as low as
side with excellent peel at the core side. 4.4°C (400 P) for some two-part systems up to
276 Sandwich construction

216°C (420°F) for some of the materials All the above forms of adhesive are in cur-
intended for service at elevated temperatures. rent use at substantial volume and most are
available from many sources.
Supported films
12.5 DESIGNING A SANDWICH
Films or tapes having a carrier of light glass
fiber, cotton, nylon, or polyester fabric, or The usual objective of a sandwich design is to
spunbonded synthetic fiber are provided save weight or to increase stiffness or to use
either dry or with slight to moderate 'tack' or less of an expensive skin material, or perhaps
stickiness, so that the parts of the assembly all three. Sometimes other objectives, such as
stay in place as they are being assembled. reducing tooling or manufacturing costs,
achieving aerodynamic smoothness, reducing
reflected noise, or increasing durability under
Unsupported films, containing only the
exposure to acoustic energy, are also involved.
adhesive
The designer's problems sift down to rela-
The very low weight films are nearly always tively few, such as getting the loads in, getting
furnished without a carrier, as the weight of the loads out and attaching small or large
the carrier itself becomes quite appreciable in load-carrying members, under constraints of
very light sandwich structures. They are often deflection, contour, weight and cost.
hard to handle and sometimes have bond line
1. Understand the fabrication sequence and meth-
control problems.
ods. The cost of a sandwich structure is
fundamentally fixed at the design stage and
Reticulating films a considerable difference in cost can result
from alternate solutions to the design prob-
These are intended for use at very low
lem. Both of the edge close-out details
weights, with the adhesive being melted by
shown in Fig. 12.12 perform essentially the
hot air after placing on the core, so that it
same job at the same weight. Placing the
draws back to the cell edge and provides
legs of the channel facing outward instead
material to form the largest possible fillet
of inward saves the cost of two relief cuts
without wasting any on the inside facing sur-
into the core and the very difficult step of
face in the middle of the cell.
sliding the edge of the core and adhesive
into the channel. Another alternative at
Cell-edge adhesive even lower cost for either fixed or simply
supported edges is shown in Figs.
This is a material pre-placed on the cell edge
12.13-12.16.
by the honeycomb manufacturer to provide
2. Use the right core. Several densities of core
the same results as those produced with retic-
can be used in a single panel, each appro-
ulating films.
priate to the load carried in the area and
adhesively bonded to its neighbor, as
Self-adhesive skins shown in Fig. 12.17. In many cases, how-
ever, the weight saved in lower density
These skins are usually structural fabrics of
areas of core is added back in the form of
glass, graphite, quartz, or aluminum coated
core splice adhesive weight. Core splices,
glass fibers, pre-impregnated with a resin,
such as those shown in Fig. 12.18(b) or (c),
which is then cured so that the fiber-filled
have been used to produce ablative matrix
resin becomes both the face structure and the
structures for large re-entry heat shields,
attaching material.
 Designing a sandwich 277

::mIIJIIJl~ GOOD 1[[01111 IIl[ ~

Fig. 12.12 The square edge close-out shown here using a channel may result in a neat, clean edge, but
requires machining both the top and bottom of the core and squeezing adhesive and core into the channel
during assembly. The alternative shown on the left would be much better.

1111111111111111 III jJ,!J ~l! illill[il 111111111

Fig. 12.13 Densified core edge treatment.

,111111111111 '''R£INFGOCEMENT
!III II 11111 11111 III
DENSIFIED CORE
WI
} IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII@IIDII
HIGH-STRENGTH INSERT
1111111111111111_
EOGE CELLS FILLEO

llllllllllllllllll]: METAL CHANNEL

11 III I I'I 111I111 III 00
FACINGS FORMED

HIGH-STRENGTH INSERT

11111111111111111
METAL CHANNEL
'=1 \111111 ~':::::17'"
CR(JSHEO ANO BONDED

HIGH-STRENGTH INSERT, f'EINFORCEMENT

\1IIIlllIIlIIIJllF METAL Z·

Fig. 12.14 Several common edge treatments.

278 Sandwich construction

Strong. Uses standard Strong. Special extru-

angle. Pop rivets to sion. Seals can be
locate and apply pres- incorporated.
sure during bonding.

Very strong. Special

extrusion. Difficult to
apply adhesive uniformly
and assemble

Very strong with inside

Low strength. Very low tie-bar. Can include
cost. Inside facing and external seal or gasket.
core scarfed then bent.
Fill corner with epoxy or
foam to stiffen.

Fig. 12.15 Several suggestions for comer designs, edge close-outs and splices.

Potting

..
Locking la.

Fig. 12.17 Typical core splice using a foam-tape

adhesive. Foaming of the tape adhesive permits a
less-than-perfect fit of core details, but requires that
the core be precisely fixed in position during the
Fig. 12.16 Additional joints and comer treatments. cure to avoid a step in the surface at the splice line.
 Designing a sandwich 279

ACROSS RIIJ80N (WI DIRECTION IN RIBBON (LI DIRECTION

c
Fig. 12.18 Joint A may be formed by simply crushing one piece of glass fabric honeycomb into the adjoin-
ing section. This method will work to some extent with some aluminum honeycombs, but not with most
other core materials. Joints B and C require a perfect match of cell shape and cell pitch and are very diffi-
cult to produce on any realistic and cost-effective basis.

but become prohibitively expensive to pro- 5. Use doublers where needed, instead of a heavier
duce for splices more than a few inches facing over the entire part. The use of doublers,
long. although adding labor cost in assembly,
3. Do not hesitate to use several joining methods often improves the part quality. Where
in the same part. Fittings to be included in a skins are formed of glass or graphite
bonded sandwich may be produced from prepreg, the problem is even simpler, since
weldments, forgings or riveted assemblies, extra plies can be added to carry extra loads
or may themselves be bonded assemblies. exactly where and as needed.
Available adhesives permit secondary 6. Use external doublers rather than internal dou-
bonding to be performed at temperatures blers wherever possible. The use of internal
from 16°C (60°F) up to 177°C (350°F) with- doublers usually means that a relief cut
out degrading the integrity of the must be made in the thickness dimension of
previously bonded sub-assemblies. the core to prevent bridging and a conse-
4. Use bolts and rivets for carrying loads (not for quent unbonded area where the doubler
soothing fears). Where space is not available ends. Figure 12.19 shows a panel where the
for progressive doublers or wide-area loads which can be carried are the same at
bonded overlaps to carry high loads, the each end of the panel. The design detail on
addition of rivets or bolts is sometimes the the left end can cost substantially more to
only solution. Their use, however, often manufacture than that on the right end.
results in lower (sometimes dramatically Figure 12.20 shows the same panel with
lower) fatigue life of the structure, in addi- both ends produced at low cost, while still
tion to increased weight. The use of achieving an unbroken outer skin line on
Ichicken rivets', added for the sole purpose one side. In the case of some skin materials,
of appearance, is to be particularly avoided, such as 0.25 mm (0.010 in) aluminum, or
since they often defeat much of the advan- most weights of pre-impregnated glass or
tage which would otherwise result from use graphite cloth, it is feasible to use thin dou-
of the bonded structure. blers without a relief cut in the core, since
280 Sandwich construction

Fig. 12.19 Internal and external doubler treatment.

Fig. 12.20 Low cost doubler treatment.

11111111111\\lj\\\\IIIII\IIIIIIII!j11111111111111

Fig. 12.21 Doublers at a skin splice.

the gap caused by bridging is small enough 12.6 STRUcrURAL A~ALYSIS FOR SPECIFIC
to be within the capacity of the adhesive to CASES
fill. Sometimes an extra layer of adhesive
The following notations are used in sandwich
film is added to help. An example of a dou-
design formulas.* This chapter's formulas are
ble skin splice using this method is shown
only for honeycomb beams and columns
in Fig. 12.21.
which have the same facings on each side of
the core.
 Structural analysis for specific cases 281

E t h2 12.6.1 DESIGN REQUIREMENTS

D = flexural stiffness' D = I L
, 2A.B Sandwich structures should be designed to
Ec = modulus of elasticity of the honey- meet the basic structural criteria listed below,
comb, Pa (psi) when these criteria pertain to the type of load-
Ef = modulus of elasticity of facing material, ing under consideration.
Pa (psi) 1. The facings should be thick enough to with-
Gc = shear modulus of rigidity of the hon- stand the tensile, compressive and shear
eycomb, Pa (psi) stresses induced by the design load.
~ = bending deflection constant 2. The core should have sufficient strength to
Ks = shear deflection constant withstand the shear stresses induced by the
L = beam span length or column height, design loads.
mm (in) 3. The core should be thick enough and have
b = beam width, mm (in) sufficient shear modulus to prevent overall
M = maximum moment, kg/m width (lb/ buckling of the sandwich under load.
in width) 4. Compressive modulus of the core and the
P = load, kg (lb) compressive strength of the facings should
P cr = column critical load, kg/m (lb/in) be sufficient to prevent wrinkling of the
P = column facing yield load, kg/m (lb/in) faces under the design load.
V = maximum shear force, kg/m width 5. The core cells should be small enough to
(lb/in width) prevent intracell dimpling of the facings
d = sandwich total thickness, mm (in) under design load.
h = distance between facing centroids, mm 6. The core should have sufficient compres-
(in); h = tc + tf sive strength to resist crushing by design
s = core cell size, mm (in) loads acting normal to the panel facings or
t = core thickness, mm (in) by compressive stresses induced through
t; = facing thickness, mm (in) flexure.
w = uniform beam load, Pa (psi)
~ = maximum beam deflection, mm (in)
12.6.2 MODES OF FAILURE
AB = 1- Poisson's ratio of the facing material
squared = 1 - p,2 Typical modes of failure are shown in Fig.
P, = facing material's Poisson's ratio 12.22.
Of = maximum facing stress, Pa (psi)
o = yield stress of facing material, Pa (psi)
T ~ = maximum core compressive stress, Pa
12.6.3 DESIGN STEPS
(psi) 1. Define loads. For multi-point loadings, use
T cs = maximum core shear stress, Pa (psi) the formulas in Roark's Formulas for Stress
and Strain. 16
2. Define beam type. The values of Fig. 12.23
The need to be able to accurately calculate the provide the simple starting point for these
exact performance for many forms of sand- calculations. Some care in using the fixed
wich structures had led to the development of end type of support is needed, as in actual
a substantial body of literature on the subject. practice total fixity is not realized and the
This chapter will cover in detail only the com- resulting deflection is greater than that cal-
monly used analyses and will provide culated.
reference sources for a number of others. 3. Determine deflection limitations. For most
applications, the allowable deflection of the
282 Sandwich construction

Facing failure

T~~~~
Initial failure may occur in either
compression or tension face. caused
by insufficient panel thickness, facing
thickness, or facing strength.
in facing

Transverse shear failure

Caused by insufficient core strength 'l'= V
chli
or panel thickness.

Local crushing of core

Caused by low core compression p
strength (]<If
=

General buckling
Caused by insufficient panel
thickness or insufficient core rigidity.
=pI! un RI II II UotE
Shear crimping
Sometimes occurs following, and
as a consequence of, general buckling.
Caused by low core shear modulus,
~lIlIlIlIhIlIlIlD~ Pcr = tG
c <
or low adhesive shear strength.

Face wrinkling Adhesive bond

failure
Facing buckles as a 'plate on an
elastic foundation'. It may buckle ~lllIlIlllirnllllll~
inward or outward, depending on
relative strengths of core in Core compression
compression and adhesive in ~ failure ~
flatwise tension. 1111111101111111

Intracell buckling (dimpling)

Applicable to cellular cores only. Faces buckle into
core cell
Occurs with every thin facings and
large core cells. This effect may cause
failure by propergating across c~ [1]2
adjacent cells thus inducing face
wrinkling.
~Irll ~ P = 4tf Ef
cr Af S

Fig. 12.22 Modes of failure in sandwich structures. Sandwich structures must be designed to resist these
modes of failure. Failures may occur which combine more than one of the modes shown.
 Structural analysis for specific cases 283

Maximum Maximum Bending Shear

shear bending deflection deflection
Beam type
force moment constant constant
V M ~ K.
Simple support Uniform load
P= ql
P PL 5 1
i I I
=~,E I 7±im
2 8 384 8

Both ends fixed

I I i I I I I I
Uniform load

P= ql P
2
PL
12
1
384
1
8

Simple support Center load

( P
2
PL
4
4
192
1
-
4
J?:;;~777 ~m
Both ends fixed Center load

I (
I
P PL 1 1
-
2 8 192 4

Cantilever P = ql Uniform load

PL 1 1
jllliII! P
2 8 2

Cantilever End load

I
P 1
t P PL 1
3

Cantilever Triangular load

~='1,ql P
PL
3
1
15 "3
1

One end simply supported

one end fixed Uniform load

I
5P PL 1 1
I I I I I f[=ql 2 8 185 14.2

Fig. 12.23 Loaded beam chart, where P = total load (per unit width), L = span, crf = facing stress, tf = skin
thickness, h = centroid distance, Tes = core stress, shear, Tcc = core stress, compressive). = (1-Ji)2 facing prop-
erty, Ef = modulus of elasticity of facings, Gc = modulus of elasticity of core in shear, lsw = moment of inertia,
sandwich, S =cell size, Ec = modulus of elasticity of core in compressions, FS = factor of safety; T = total
sandwich thickness (note that P must be determined for a beam unit width). If deflections are critical,
actual deflections should be verified by tests.
284 Sandwich construction

structure is usually limited to L!360. In the shear component should be calculated

some cases, greater deflections may be and the core selection may be influenced by
used, or, as in the case of snow skis, very the shear modulus needed.
much greater deflections may be a normal 9. Face wrinkling and intracell dimpling. With
part of the function of the structure. thin skins, a local failure of the skin in buck-
4. Select skin material. Skin considerations ling may be encountered. A check on the
include the weight target, possible abuse 0fcrit will determine whether this may be a
and local (denting) loads, corrosion or dec- design consideration.
orative constraints and costs. Select 10. Other considerations. Often, honeycomb
standard thicknesses and make the initial panels are supported on more than two
calculation as outlined below. The facing sides. If the ratio of length to width is
thickness directly affects both the skin stress greater than 3: 1, the calculations using the
and the deflection. shorter span and designing as a unit beam
5. Calculate first approximation. After the first are quite adequate. The formulas in Roark16
sandwich thickness, h, is determined, are useful where the shear deflection may
another selection of tf or Ef may be made to be ignored, using the following formulas.
arrive at more desirable or practical values t h2b bf3
of h. Most sandwich structures in ordinary [ =_f_.[ = __
sandwich 2' solid 12
usage are in the thickness range of
1.5-150 mm (0.06-6.00 in). So, for plate calculations:
6. Select skin thickness. Keep in mind that mate-
[solid = 6tf h2
rials such as fiberglass cloth and aluminum
are available in specific, standard thick- Use of these formulas for deflections may give
nesses. After the skin thickness for lower values than actually experienced, since
deflection is selected, it should be checked the shear deflection may be important. Table
for stress. The formula for Of is used and a 88 of Roark16 gives some approximate multi-
factor of safety determined. pliers to use for plates when supported as
7. Select core. Calculate the core shear stress, tc-c' noted.
Note that the core strength is not the same in
the L and W directions. Refine the selection,
12.6.4 SIMPLE FORMULAS
including considerations of material compat-
ibility, cell sizes and types. Determine the Bending stress in facings:
corrections needed to account for the effects
M
of thickness on strength, as shown in Fig. Of = IT
12.5. Check the factor of safety using the cal- f

culated stress and the corrected allowable where M is determined by Fig. 12.23.
stress. Other considerations include crushing
Core shear stress:
and compression strengths, modulus in
shear, weight and costs. For rolling wheel V
t =h-
loadings, the crushing strength and the skin cs

thickness are often the most important con-

where V is from Fig. 12.23.
siderations.
8. Check deflection. For many applications, the Deflection:
calculation of the expected deflection may
omit the shear deflection portion. With a
very small deflection limitation, with a very
thick sandwich, or with a very short span,
 Structural analysis for specific cases 285

(Kb and K. from Fig. 12.23). material is to be woven roving, polyester and
core to be KP-3/8-60(25).
For most beams, the second term is relatively
small, but should be checked if deflection is
critical or span is short. Load, P': P' = 120/144 =0.833 psi

Moment of inertia, sandwich: Span, L: L = 8 x 12 = 96 in

Isw =
t h2b
T ~ Ks' M, V from Fig. 12.23:

Face dimpling: Kb = 0.013, K. = 0.125, M = 8, V= 2.

a ferit 2E£[ st£]2

=T Skin, t f: Try t f = 0.090 in

Face wrinkling: A, Ef for fiberglass use:

Ect ]1/2
a fcrit 0.82 = E£[ E/: A=0.98, E£ =1.85 X 106•
Calculate h:
Factor of safety:
. ~P'L42').}/2
FS = Allowable or typical stress L1 = t~£
calculated stress
0.013 x 0.833 x 964 x 2 X 0.981 / 2
=----~--------,--,cc---
0.090 x 96/270 x 1.85 x 106
12.6.5 SAMPLE PROBLEM: ANALYSIS OF FLAT
RECTANGULAR SANDWICH BEAMS h =5.518 (round out to 5.5 = panel total thick-
ness, h = 5.41)
Design a flat roof panel for a bus stop. Use a
Try thicker skin t£ = 0.150
snow load of 120 Ib/ft2. Use a simple panel
with a simply supported span of 0.203 m (8 ft). h = 0.013 x 0.833 x 964 x 2 X 0.981/ 2
Deflection is to be limited to L/270 and the fac- 0.150 x 96/270 x 1.85 x 106
tor of safety is to be greater than 2.0. Skin
= 4.274
Use 4.00 overall thickness, h = 3.850. Since
either construction is practical check out the
skin and core stresses:
P'U 0.833 x 962
af = t/tMe = 0.090 x 5.41 x 8 = 1971 psi
or,
0.833 x 962
a£ = 0.150 x 3.85 x 8 = 1662 psi
Note that the skin stresses are quite close;
therefore, the factors of safety would be simi-
lar.

Fig. 12.24 Schematic diagram of a flat sandwich FS = 38 000 = 19; FS = 38000 = 23

panel. 1971 1662
286 Sandwich construction

For core: 12.6.6 ANALYSIS OF FLAT RECTANGULAR

SANDWICH COLUMNS: COLUMN DESIGN
P'L 0.833 x 96 EXAMPLE
f
cs
= -hV = 5.41 x 2 = 739 psi

or, L=8ft
....·-----(20.32cm}----~·
0.833 x 96 10 38 .
3.85 x 2 = . PSt
Note that the core stresses are quite low and ~IIIIIIIIIIIIII~-p
. No end fixity
there is not much difference in the stresses for
the two thicknesses chosen. For KP-3/8-60(25),
W shear strength = 60 psi. From Fig.12.5: thick-
ness factor = 0.42, W shear modulus = 5800. Facings:
Tempered hardboard, Gf = 3600 psi
W, shear, corrected = 60 x 0.42 = 25 psi
# = 0.99
FS = 25/ 10.4 = 2.4.
Ef = 0.65 X 106 psi
The use of KP-3/8-60(25) with a factor of
safety of 2.4 could be marginal, which may Core:
vary from lot to lot of material. The other
properties, compression strength and density Urethane foam, 6Ib/ft3, tc = 3 in
are acceptable. Note that if the core is oriented From Table 12.1:
to utilize the L shear properties, KP-l /2-80(11),
with fes = 70 x 0.42 = 29.4 might be satisfactory. Aes = 90 psi

Calculate deflection: Ac_c = 170 psi

K"P'L42 KrU
For 5.50 T, A = tf h2Ef + GJl Gc = 1500 psi

0.013 x 0.833 x 964 x 2 Check facing yielding:

=~0~.~~x~5~.47.12~x-l~.~85~x~I~= P y = 2tf ay = 2(0.25)(3600) = 1800 lb/in

+ 0.125 x 0.833 x 962 Check general buckling:

5800 x 5.41
Pcr =
= 0.377 + 0.032 = 0.0409 in
Note that the added shear deformation is only
9% of the total deflection.
D = 0.65 x 1()6 x (0.25)(3.25)2
For 4.00 T A = 0.013 x 0.833 x 964 x 2 2(0.99)
, 0.150 x 3.851 x 1.85 x 1()6
= 866 8721b-in/in of width
+ 0.125 x 0.833 x 9@
5800 x 3.85 x2(866872)

= 0.447 + 0.042 = 0.489 in (96)2 + 1t2(866 872)

3.0(1500)
 Manufacturing sandwich structures 287

= 112Ib/in or 13521b/ft temperature (both pressure and temperature

in the precise amounts, at the precise time
Check shear crimping:
required for cure of the adhesive being used);
Pex = tcGc = (3.00)(1500) and the provision for tooling and fixtures to
hold the assembly in the desired shape and
= 3214 1b in/in of width keep all the details in their proper positions
during cure. Many different ways of providing
Check dimpling and wrinkling: these conditions are currently used, from vac-
Since facings are relatively thick and continu- uum bags or simple presses to autoclaves and
ously supported by foam. core, dimpling or unit tools, where volume and complexity can
wrinkling will probably not occur. justify them. Most of the equipment is similar
to equipment used in producing bonded struc-
12.6.7 DESIGN CONDmONS tures or reinforced plastic parts where no
sandwich structure is involved. However,
In-depth treatments for the design conditions bonding of sandwich structures is nearly
listed in Table 12.7 can be found in MIL- always performed at lower pressures than is
HDBK-23l, available from the US Government the bonding of structures which do not have a
Printing Office. low density core and tooling is sometimes
lower in cost as a result. Aside from the need
for lower maximum pressure, there is little
12.7 MANUFACTURING SANDWICH
noticeable difference between a sandwich
STRUcruRES
bonding facility and one which only handles
The manufacture of sandwich structures non-sandwich bonding.
requires three conditions to be met: the appli- A few suggestions can be offered to aid in
cation of pressure; the application of living with the problems of sandwich bonding.

Table 12.7

Subject MIL-HDBK-23 *
CHAPTER

Wrinkling of sandwich facings under edgewise load 3

Dimpling of sandwich facings under edgewise load 4
Design of flat, rectangular sandwich panels under edgewise compression load 5
Design of flat, rectangular sandwich panels under edgewise shear load 6
Design of flat, rectangular sandwich panels under edgewise bending moment 7
Design of flat, rectangular sandwich panels under combined loads 8
Design of flat sandwich panels under uniformly distributed normal load 9
Design of sandwich cylinders under external radial pressure 10
Design of sandwich cylinders under torsion 11
Design of sandwich cylinders under axial compression or bending 12
Design of sandwich cylinders under combined loads 13
Design of sandwich strips under torsion load 19
Design of flat circular sandwich panels loaded at an insert 20
.. MIL-HDBK-23 is revised from time to time, with new chapters sometimes added and older material updated. A check
with the Plastics Technical Evaluation Center, US Army Armament Research and Development Command, Dover, New
Jersey, USA, can verify that you are in possession of the most recent revision.
288 Sandwich construction

1. Make sure the core is properly sized to fit much more slowly, resulting in warpage
the space it is intended to occupy. If it has problems upon cool-down. Warpage on
been stretched a little, to make the distance very light parts can also be caused by one
from one edge member to the opposite one, side cooling down too fast as a result of
it will probably shrink back as the cure having one side removed from the still-hot
cycles starts, leaving mysterious voids next tooling, while the other side continues to
to an edge member. If it is undersize in stay at the temperature of the tool. Also one
thickness at an edge, the adjoining edge side, next to the bag may be heating faster
member or fitting will hold the facing away or to a higher temperature than the oppo-
from the core and result in an unbonded site face, which is in contact with a massive
area. and still cold tool. Slower heat-up rates or
2. If a honeycomb core is being used, remem- better heat distribution in the tool design
ber that the adhesive between the core and will help prevent these problems.
the faces will end up much thinner than the 6. Be sure to provide a route for the escape of
same adhesive between the edges or solid trapped air and gases from a totally enclosed
inserts and the facings. For this reason, it is part while it is being cured. This is particu-
common to require the core to be as much larly important in parts which are vacuum
as 0.25 mm ( 0.010 in) thicker than adjoining bagged to a female tool and cured in an auto-
solid parts in the same assembly. clave. A coarse cloth 'breather' should be
3. The elevated temperatures which most enclosed inside the bag to prevent the bag
core-to-facing adhesives require for curing from sealing off portions of the assembly as
are often inaccurately measured. A good pressure is being applied. Critical or expen-
point to remember is that only the adhesive sive assemblies should have several vacuum
being cured can give you the cure tempera- lines attached at different points of the bag,
ture yOll' are trying to measure. Some shops with each monitored separately by a pres-
insert thermocouples directly into the bond sure recorder.
line to determine temperature and then 7. Caul plates should be carefully matched to
leave the thermocouple permanently in the the job they are expected to perform. These
part after cure is completed. tooling aids are often used to cover the top
4. Most adhesives flow at an early point in the of an assembly containing several different
cure cycle. At this time, the bond lines will pieces of core, inserts, edges, etc., so that a
change in thickness by substantial amounts. thin skin will not push each detail to the
The tooling employed to establish the shape minimum bond line thickness and result in
of the part and hold details in place must an uneven outer surface. When the caul
also allow the details to move into their plate is moderately stiffer than the top skin,
final cured position. Simple examples are a the bonding pressure is transmitted more to
hot platen press, in which the platens close the thicker inserts and less to the under-
on the sandwich as the bond lines grow sized inserts, allowing all of the details to
thinner, or an autoclave, in which a flexible 'float' in the adhesive before cure, resulting
bag follows the details as the adhesive in optimum relative placement of all the
flows, continuously transmitting the auto- internal details in the sandwich. If the caul
clave pressure to the shrinking assembly. plate is extremely stiff or thick, this effect is
Keep in mind that most adhesives are very changed to one of simply bridging over the
weak and crack-prone as they go through most oversized details and the danger of
the gel point. producing voids or unbonded areas over
5. Inserts or heavy members being cured as a the thinner details is substantially
part of a very light assembly will heat up increased. Generally, the caul plate should
 Manufacturing sandwich structures 289

not be more than two or three times the

thickness of the sandwich facing material.
Where thicker caul plates are used, the
dimensional control over the size of detail
parts in the assembly must be correspond-
ingly better. The advantage of using such a
thick caul plate derives from the ability to
make both sides of a sandwich part have
the smooth appearance usually associated
only with the 'tool side'.
8. Make sure that core, pre-cured or rigid
edges, inserts, skins and other relatively
unyielding details assembled in the lay-up
have close enough dimensional control to
allow adhesives or resins to achieve the tar-
get strengths. In simple bonded assemblies,
a tolerance of ± 0.1 mm (± 0.005 in) is neces- Fig. 12.25 Honeycomb carving bits employing a
sary, while assemblies having multiple slitting saw 0.254 mm (0.010 in) thick x 12.5 teeth
layers of prepreg or many layers of thin per cm (32 teeth per in), 50.8 mm (2 in) in diameter
at the cutting edge. Turning at 12000 to 30 000 rpm,
metal doublers can sometimes be success- these tools leave a smooth, burr-free surface on
fully produced with much less demanding nearly any core material. The coarse teeth on the
dimensional control. tool in the foreground are for the purpose of break-
ing up and removing the excessive amounts of core
in cut depths of 5.08-50.8 mm (0.2-2.0 in).
12.7.1 CORE SHAPING

When core materials must be cut, trimmed,

carved, or shaped, many special purpose tools
are available. Sawing is the most common
machining method, using either conventional
blade tooth patterns, or, for some trimming
operations, a special 'honeycomb band', in
which the blade appears to be running back-
ward, with the teeth sharpened on the back
side, so that each tooth acts as a slicing knife
blade. A different type of saw is also used as a
mandrel-mounted router bit. Such tools,
shown in Fig. 12.25, are very common where
sculpturing of honeycomb or foam is to be
accomplished. Router speeds vary from
1 000-30 000 rpm for blade diameters of
1.8-10 cm (0.75-4 in). Roll forming can be
accomplished on metal cores, as shown in Fig. Fig. 12.26 Metal honeycomb may be roll-formed
12.26, while non-metal cores must usually be using ordinary slip rolls. The surface usually must
be protected during the operation by inclusion of a
heat formed. In either case, forming can be loose sheet of thin sheet metal between the core and
much easier if an inherently formable cell con- the outer forming rolls. The tool being used is a
figuration, such as that shown in Fig. 12.8 'Farnham Roll', commonly used in sheet metal
view H, is used. shops.
290 Sandwich construction

REFERENCES

1. MIL-HDBK-23, US Government Printing Office,

Washington, DC*.
2. MIL-HDBK-17, US Government Printing Office,
Washington, DC.
3. MIL-HDBK-5, US Government Printing Office,
Washington, DC.
4. MIL-A-132, US Government Printing Office,
Washington, DC.
5. MIL-A-25463, US Government Printing Office,
Washington, DC.
6. MIL-STD-401, US Government Printing Office,
Washington, DC.
7. Adhesive Bonded Aerospace Structures Standard
Repair Handbook, US Government Printing
Fig. 12.27 Nose radome core assembly, assembled
Office, Washington, DC.
by edge-bonding together several post-formed sec- 8. Hexcel Corporation, TSB-120.
tions of glass fabric-phenolic honeycomb. Nomex 9. Hexcel Corporation, TSB-123.
core may also be formed in this manner. 10. Hexcel Corporation, TSB-124.
11. Aleore, TR-112.
12. American Cyanamid, Handbook of Adhesives.
13. Baltek Corporation, Baltek Catalog.
14. Plantema, Frederic J., Sandwich Construction,
John Wiley & Sons, New York. 1966.
15. American Plywood Association, Plywood Design
Specification.
16. Roark, R.J., Formulas for Stress and Strain,
McGraw-Hill, New York, NY., 5th edn, 1975.
17. Tunoshenko, 5., Woinowsky-Krieger, 5., Theory
of Plates and Shells, McGraw-Hill, New York, NY.
2nd edn, 1959.
'Publications of the US Government may be updated
and revised from time to time. Be sure you have the
most recent edition. This can be checked by con-
tacting the Plastics Technical Evaluation Center, US
Army Armament Research and Development
Fig. 12.28 Effect of roll-forming on aluminum hon- Command, Dover, New Jersey. The publication
eycomb. The core on the left has been roll-formed in MIL-HDBK-23 was abandoned some years ago.
sheet metal forming rolls, while the piece on the However, because the information it contained con-
right has not been pre-formed at all. Note the anti- tinues to be needed by designers of spacecraft
clastic, or 'saddle shape', which the unformed piece structures, the entire publication will in future be
assumes when forced to a cylindrical form. included within MIL-HDBK-17.
METAL MATRIX COMPOSITES 13
V. 1. Kostikov and V. S. Kilin

13.1 INTRODUCTION 3. providing the correct fiber length, greater

than the critical length;
Industrially developed countries are sucess- 4. even distribution of fibers in the matrix;
fully producing materials and shapes based 5. orientation of fibers in the direction of the
on metal matrix reinforced with carbon fibers applied load;
(MMC). Metal matrix, ceramic and carbon 6. achieving the required shape and dimen-
composites reinforced with both discrete and sions of the MMC;
continuous reinforcements have been 7. obtaining a MMC strength reasonably near
describedH . The present chapter will to theoretical.
describe the results of investigations of the
properties and preparation of aluminum/ car- Contemporary processes for obtaining an alu-
bon composites minum/ carbon MMC can be divided into
MMC production deals with the solution of three main types: solid stage, liquid stage, and
specific problems which involve the formation solution sedimentation. In the solid stage, the
of useful structural components, with the greatest effort has been spent in obtaining the
attainment of their potential properties. The MMC by hot extrusion. In this case, a foil, used
first problem is to fill the interfiber space of the as a matrix, is interspersed with carbon fibers
carbon fibrous framework with matrix metal or to form a preform. The preform can then be
alloy. Several molding methods are available. subjected to optimum pressure, time and tem-
Optimization of the materials and the perature in air, vacuum or inert gas medium.
processes for obtaining usable MMC may fur- Considering metal bonding in the liquid stage,
ther the use of MMCs in industry. When the chemical bonds predominate; rarely are there
MMC is reinforced with high molecular and mechanical bonds and there are no physical
high strength fibers the composite may be very bonds7 • Strong chemical bonds are possible
brittle and can have thermal dynamic incom- because the atoms (or surface) of one sub-
patibility with many metals used as matrices. stance come close enough to the surface of the
MMC production technology is compli- other (1.5-3 A) to enter the zone of the surface
cated and requires satisfaction of the following atoms' field-of-forces effectivity.
conditions, of which the most significant are as When MMCs are produced by solid stage
follows: methods (diffusion welding, rolling, extru-
sion), it is practically impossible to provide
1. maintaining the reinforcing fibers' strength; full convergence of fiber and matrix surface
2. ensuring a strong bond of fibers with matri- due to uneven surface contours. Only applica-
ces and between the matrix layers; tion of extreme external forces makes the
convergence possibleS. This pressure increase
Handbook of Composites. Edited by S.T. Peters. Published leads to brittle fiber breakage. Also, some
in 1998 by Chapman & Hall, London. ISBN 0412 54020 7
292 Metal matrix composites

fibers inside the bundle may not be bonded takes place only after the surface catalysis by
with the matrix because of a shadow effect. treatment in an oxidizing medium, sensitizing
Due to the small diameters of carbon fibers and activation of the carbon fibers prior to
and the short distances between them in the proper metallization. The method is complex
tow (about several microns), poor bonding of because of the requirement for a strong, thin,
matrix to the uncoated fibers located deep in well-adhering bond to the carbon fibers with-
the bundle prevents development of their out compromising their mechanical or
potential strength. Also, during the produc- physical properties. A typical treatment of car-
tion of composite material, MMCs based on bon fibers in an oxidizing medium of 65%
metals such as AI, Mg, Ni, Ti, Cu, etc. and rein-
solution of nitrous acid for 5 min does not lead
forced with carbon fibers, the components to fiber strength loss. Ions and radical groups
show high reactivity affecting composite per- that have affinity for metal are attached to the
formance. Protective coatings (silicon, degraded surface of the carbon fibers.
titanium, zirconium carbides, and nitrides) on When silicon carbide is used as a barrier
the carbon fibers can prevent these undesir- layer, an oxidizing treatment of the carbon
able effects9,1O. fibers is not necessary because the barrier
Many MMCs can be made by soaking the coating itself is a good adhesive layer with a
reinforcing fibers in the molten metal matrix. rough surface structure which bonds to fibers
The problems can be solved by application of and nickel coating. In this case, preliminary
a thin, strongly-bonded metal film that allows treatment for all types of carbon fiber tows
the melted matrix material to completely wet and tapes includes only sensitizing with stan-
the fiber. There are several ways to apply thesenous chloride solution at 80°C for 10 min and
coatings: stretching fibers through the melt; activation of palladium chloride at 80°C for 5
spraying the molten matrix metal; depositing min. During the sensitizing process,
a metal film from the gaseous phase. However, hydrolized Sn2+ ions which have high absorp-
application of a coating at constant thickness tion properties are strongly attracted to the
on the carbon fibers in tows is not always fea- carbon fiber surface. During activation, palla-
sible because of fiber mutual screening. Fiber dium chloride is reduced to metal by the
coating by a chemical method may be more ionic tin bonds formed at sensitization.
effective. Washing of fibers with water follows. The
activated fiber surface is then dried at
60-70°C for 15-20 min.
13.2 CARBON FIBER COATING METHODS
Nickel coatings are applied from a solution
Chemical methods are used to deposit thin of nickel chloride, 50 g I-I, sodium hypophos-
coatings and improve wetting on filaments in phite, 20 g 1-1, ammonium chloride, and 50 g I-I,
carbon tows and tapes. Because of low depo- trisodium citrate at 80 C at pH 8-9. A lustrous
D

sition rates, chemical methods cannot be used nickel coating is obtained which varies in
for deposition of thick layers of matrix. There thickness from 0.05-2!lm at holding time of
are two widely-used methods of deposition. 0.1-4 min. An estimate of coating continuity,
and indirect data on coating adhesion to car-
bon fiber by means of a scanning electron
13.2.1 CHEMICAL PRECIPITATION
microscopy, has shown that the coatings are
(ELECTROLESS DEPOSITION)
applied evenly and that the nickel penetrates
Chemical deposition of nickel and copper on the microrecesses of the fiber surface, fully
carbon tows and tapes with various textile replicating it and filling the grooves and irreg-
structures is based on the reduction of metal ularities. In carbon tows every individual fiber
ions from a water solution. Nickel deposition is coated with nickel.
 Carbon fiber coating methods 293

The textile structure of a braided fabric 13.2.2 THERMAL DECOMPOSmON OF

becomes more complicated. Coating thickness CARBONYLS
discontinuities may be considerable because
Coatings of refractory metals applied to car-
of reduced clearances between fibers in zones
bon tows and tapes by thermal decomposition
difficult to access and stagnation and poor
of volatile carbonyls in the gaseous phase have
solution exchange take place. In these zones,
the following advantages: low temperature of
the solution is fully depleted, while in other
decomposition of carbonyls of some commer-
zones, the nickel plating may continue. The
cially important metals (Cr, Mo, W, Rhs etc.);
addition of 0.01 g 1-1 lead sulfide to the solu-
high reaction velocity; uniformity of coating
tion slows the process of electroless nickel
thickness on each individual carbon fiber; ease
plating and eases the conditions of the forma-
of adjustment of structures and properties of
tion of uniform coating thickness. Reducing
coatings as a result of change of the deposition
the number of individual fibers in carbon tows
conditions.
to 100 and the number of ends in tows also
The thermal decomposition of the metal
he.lp to provide better uniformity of coating
carbony is carried out as follows:
thickness. When 300-fiber, 6 J.IlIl equivalent
diameter, twisted tows of bean shaped cross Me(CO)n =Me + nCO
section were coated, the coating thickness on
the fibers was uneven and varied from In addition to the main reaction, side processes
0.2-0.6 11m up to 0.5-1.0 I1ID at nickel gain in that cause the presence of admixtures of metal
weight from 0.5 to 1.0 g. (Editor's note: Carbon carbides in the coating can take place. As the
fiber tows are not generally twisted and are all temperature is increased, the possibility of
single end in the USA and other countries.) obtaining pure metal layers without admix-
The thickness of nickel coatings, used for tures of free and fixed carbon occursll • To
subsequent impregnation of carbon tows and obtain pure metals with a minimum content of
tapes with aluminum melts, is usually about admixtures, the process should be run at the
0.5-1.0 J.lm. The nickel coating obtained by ~ghest possible temperature and sufficiently
chemical deposition from solutions does not high vacuum. Some characteristics of coatings
diminish the fiber strength. by precipitation from the gaseous phase are
The chemical deposition of copper is shown in Table 13.1. The type of substrate
accomplished at 20°C in alkaline solution that heating is usually determined by temperature
contains sodium sulphate, formalin, caustic requirements, and may be performed by con-
soda, potassium sodium tartrate, and diethyl duction, induction, radio frequency or
sodium carbonate. At a holding time of 3--6 infra-red radiation.
min, the coating thickness reaches O.l-OAI1ID. During the vapor phase of carbonyl decom-
Before application of copper coatings, the position there are a number of subsequent
fibers are subjected to the same preliminary processes: vaporization (sublimation) of car-
treatment as the nickel coating. bonyls, heat transfer, chemical process in the
The more rapid electrolytic method of gaseous phase and substrate, adsorption-des-
deposition should be a more efficient tech- orption on the substrate, formation of nuclei
nique to obtain thick nickel or copper coatings. and growth of coating crystals12 • Successively
However, the electrolytic method may be used changing the carbonyl vapor feed velocity
only for building up of the coating layer on (growth of the coating layer from 0.3 to 6-811ID
previously electrolessly coated carbon fibers. h-1) and the temperature of substrate heating
Otherwise, the fiber tows will be coated with a from 300 to 900°C can result in denser coatings,
crusty layer from the outside. which will act as barriers excluding direct inter-
action of the matrix and the fiber, with a more
294 Metal matrix composites

Table 13.1 Conditions for production of metal coat- At the initial deposition stages, thin smooth
ings by thennal decomposition of carbonyls from films reproducing the fiber surface are formed.
the gas phase At this point, one of the most important factors
is the selection of an optimal coating thickness.
Metal Initial Heating
temperature,OC
This depends substantially on intended use of
coating substance
(carbonyl) the material. Extremely thin coatings may be
of carbonyl of base entirely soluble in the matrix and incapable of
Cu Cu(CO)CI 20 250-400 preventing counterdiffusion of the reacting
Au Au(CO)Cl 20 120-150 components and thus may fail to ensure the
V V(CO)6 20 70-100 barrier effect. Thicker coatings reduce the car-
Cr Cr(CO)6 40 350-700 bon fibers tensile strength, as shown below.
Mo MO(CO)6 50 450-700
The estimated strength of carbon fibers
W W(CO)6 70 450-700
Mn ~(CO)10 70 110-300 coated with silicon, titanium and zirconium
Te Te2(CO)10 20 60-70 carbides as a function of the corresponding ele-
Re Re2(CO)10 70 400-600 ment content is shown in Figs. 13.1 and 13.2.
Fe Fe(CO)s 25 100-350 The element content is assumed to be equiva-
Ni Ni(CO)4 20 100-250 lent to the carbide coating thickness, provided
Co CO(CO)3NO 20 180-220 the coating is uniform. The strength change of
Co Co2(CO)g 20 180-200
carbon fibers with a silicon carbide coating
Ru Ru(CO)s 20 200-300
Pt Pt(CO)2C~ 100 5()(Hi00 derived from the substrate carbon is of particu-
Pt [Pt(CO)2]X 20 210-220 lar interest [see Fig 13.1, coating deposition
reaction (1)]. 'The strength of carbon fibers (ini-
tial strength 2000 MPa and thickness 25 nm) at
first increases at a coating thickness of 2-5 nm.

,
porous upper layer, providing the capillary
MPa
effect for wetting and improving the fiber
impregnation.
~~
250 ~ 2500
•
-
13.3 APPLICATION OF COATING IN GAS
'"E 200 2000
PHASE E
-0

\
;;::
A conventional method for the effective pro- ~

t :\
- 150 1500
tection of carbon fibers to avoid reaction with

".
metals is chemical vapor deposition (CVD) or
refractory coatings. (IJ 100 1000
.!!! j
For example, when carbon fiber materials in
the form of VMN grade tow and a carbon tape
'Kulon' are used as a substrate for the coat-
~ 50 50

ings, the coating is deposited from an

appropriate gas mixture by one of the follow- o o
ing reactions: o 2 4 6 8 10
Si Content, Wt. %
(1) MeCl4 + C ~ MeC+2C~
(2) MeCl4 + CH4 ~ MeC+4HCI Fig. 13.1 Variation in strength of CF silicon carbide
(3) 2MeC14 + N2 + ~ ~ 2MeN +8HCl coating relative to thickness and composition of
steam-and-gas mixture: (1) • deposition reaction:
The coating thickness is controlled by the SiC14 + CH4 -7 SiC + 4HC1; (2) 0 deposition reac-
deposition time. tion: SiCl4 + C -7 SiC + 2C1 2"
 Application of coating in gas phase 295

1.2

1.0 "\
\ -r-- I---
.s:::
.§~

----
:::l
f!!
~~ 0.8
ZrC
~
1\
§~
.c:t::
_... ....c:
(1j'-
00
0.6
~
0 Q) "'-

"'" ~
.s::: .~
'5«s TiC
~e
Ui Ci 0.4
~:S
'iii 0co
c: 0
Q)
I-
0.2

o
1 2 4 6 8 10 12 14
Me (Ti, Zr), Wt. %

Fig. 13.2 Relation between tensile strength of titanium carbide and zirconium carbide coated CF and coat-
ing thickness.

This increase may be a result of healing of fiber

zrC
surface defects. Then the strength of carbon
f----
fibers sharply decreases with increasing coat-
ing thickness. In the case of carbides obtained V'
-----
by the reaction scheme (2), the carbon fiber
strength is stabilized at a certain level when a
/' VsiC

definite thickness of the coating is attained.

The Young's modulus of carbon fibers with
a barrier is a structurally less sensitive factor; it o
changes quite insignificantly and is usually
o 4 8 12 16
Si,Zr, wt. %
somewhat higher than that of the fiber, irre-
spective of the initial coating thickness and 1:
type. The shear strength of the carbon fiber
E f..---TiC
------
~ 40

/
:::.::
surface increases, apparently by surface defect .c
healing and the strong bonds between the car- "&
c:
bon fibers and the coating (Fig. 13.3) based on ~
Ui 20
the results of torsion testing. 1U
Q)
To determine the barrier properties, the .s:::
CJ)
compatibility of carbon fibers having a refrac- o
tory coating with a metal matrix (Al, Ni) was o 2 4 6 8
Ti,Wt. %
investigated by an internal friction method9,12
using suitable micro composites in the form of Fig. 13.3 Relation between torsional strength of
carbon fibers with a double coating. The nickel coated CF and coating thickness.
296 Metal matrix composites

coating was deposited on the carbon fiber sur- not subject to noticeable structural changes
face by CVD from an aqueous nickel salt connected with interactions of the carbon
solution, and samples of the aluminum matrix fibers with nickel or the coatings. For carbon
microcomposite were cut from plates obtained fibers with and without silicon carbide or zir-
by vacuum hot molding or vacuum aluminum conium nitride coatings, the maximum
sputtering. Internal friction is the logarithmic internal friction background is observed at
decrement divided by 1t: 600°C which indicates the structural changes
to the carbon fibers. X-ray analysis showed the
presence of nickel carbide. For the same car-
(13.1)
bon fibers having no contact with nickel, the
curve was similar to a curve for internal fric-
where Al and An+I are the amplitudes of the tion change for fibers with titanium and
first and the nth vibration. zirconium carbide coatings.
Compatibility of carbon fiber and barrier The advantages of carbide coatings over
coatings based on silicon, titanium, and zirco- nitride with thicknesses less than 300 nm were
nium carbides, and titanium ~c!~irconimn shown when studying the compatibility of the
nitrides with nickel additions (coating thick- coated carbon fibers with an aluminum matrix
ness 50--100 nm at up to 1200°C) is shown in (Fig. 13.5).
Fig. 13.4. It can be seen that up to llOO°C, the Direct investigations confirming the barrier
internal friction background is negligible in coating efficiency have been conducted with
the case of titanium and zirconium carbides. the systems C-TiC-AI, C-SiC-Ni-AI,
This means that in the absence of pronounced C-SiC-Si-Al. The effect of carbide coatings on
maximum and minimum, the carbon fibers are the carbon fibers-aluminum interaction was

120~--~--~--~---r---r---'

.,. 100 .,.

0 SiC 0
x x 240
~b
b"C
"ti c:
c:
e
:::I
e
:::I

OJ
-" !2
0
til 15
m 160
m
c: c:
0 0
U
.c: ~
LL
LL
"iij zrC "iij
c:
E 80
S TiC $
E E

o 200 400 600 800 1000 1200 200 400 600 800
Temperature, °C Temperature, °C

Fig. 13.4 Relation between internal friction back- Fig. 13.5 Relation between internal friction back-
ground of barrier coated CF and temperature in ground of barrier coated CF and temperature in
contact with nickel. contact with aluminium.
 Application of coating in gas phase 297

investigated with nickel and silicon coatings. Deposition of nickel and aluminum coat-
The degree of aluminum-carbon interaction ings on a SiC barrier layer substantially affects
was estimated in terms of quantity of the the carbon fibers-aluminum interaction. These
obtained aluminum carbide, using the alu- coatings, dissolving in the aluminum melt,
minum carbide-water reaction. ensure impregnation of the carbon fiber tows.
The results of determining the aluminum However, in the presence of nickel, the resid-
carbide content on carbon fibers with various ual aluminum content on the fibers after
coatings, and their tensile strength after annealing at 720°C is increased. This increase
annealing (the aluminum matrix is etched obviously is associated with a nickel-silicon
away) are shown in Table 13.2. carbide interaction that results in opening the
These results suggest a strong alu- fiber surface.
minum-carbon fiber interaction. The However, the surface layer increases the SiC
quantity of aluminum carbide is sharply barrier properties. The quantity of aluminum
increased with annealing temperature and carbide measured after annealing under the
exposure time. However, a SiC coating with a same conditions is decreased down to the level
thickness of only 28 nm decreases the alu- which can be obtained for TiC coating. This is
minum carbide formation by a factor of 1/5. associated with the fact that silicon dissolving
Doubling the SiC coating thickness inhibits in aluminum saturates it and suppresses the
the reaction resulting in the aluminum car- following reaction:
bide formation by an order of magnitude.
(4) SiC + Al ~ (SiC) + Al4C3
Subsequent increases of the coating thickness
affect the aluminum carbide formation at a To summarize, based on the estimated
lesser degree. strength of carbon fibers with barrier coatings
and their compatibility with aluminum and

Table 13.2 The effect of coating thickness on the aluminium carbide content in the microcomposite Al-C
and CF strength after annealing

Coating appearance and Temperature, Al4C3 content, mg m-2 Tensile

thickness °C Time, min. strength"
CPa
5 10 20
Without coating 670 10.5 17.9 35.1
720 22.2 44.8 50.9 1.43 (2.80)
770 28.3 47.0 54.5
SiC,28nm 670 2.2 2.8 4.3
720 4.9 5.8 6.5 1.96 (2.90)
770 6.2 10.9 17.7
SiC,56nm 670 0.6 0.9 1.1
720 0.8 0.8 1.7 2.34 (2.53)
770 1.0 1.4 2.6
SiC,97nm 670 0.5 0.7 1.0
720 0.7 0.8 2.0 2.28 (2.43)
770 0.7 1.4 2.7
SiC, 56 nm + Ni, 200 nm 720 1.2 2.7 3.1
SiC, 69 nm + Si, 230 nm 720 0.4 0.5 0.8
TiC,53nm 720 0.3 0.5 0.6
a Initial CF (without coating) strength shown in brackets
298 Metal matrix composites

nickel, it can be concluded that titanium and shapes, by rolling in dies and extrusion. The
zirconium carbides are preferred for a nickel hot extrusion, and consolidating in autoclaves
matrix, and that silicon and titanium carbides are used for manufacturing cylindrical cases
are preferred for an aluminum matrix with the and tubes with longitudinal and circumferen-
coating thickness of no more than 100 nm. tial reinforcement.

13.4 TYPES OF COMPOSITE STRUCTURES 13.5 CONTINUOUS CASTING

Typical cross sections of MMC composite Preformed composite MMC with constant
structures are shown in Fig. 13.6. Many types cross section can be manufactured by contin-
of continuous length composite structure can uous casting or drawing of fibers through
only be produced by continuous casting. In melt. The principal advantages of continuous
some cases, it requires a preliminary applica- casting are continuity, reduced time of con-
tion of a barrier coating on the fibers by means tact of the fibers with molten metals and
of chemical deposition from solution or minimum unstable period. This method also
vapor-gaseous mixtures. It is convenient to results in a high production system, effi-
prepare a unidirectional tape for composite ciency and reduced capital investment. The
structures by plasma spraying of matrix alloys basic expenditures are for control equipment
on rows of monofibers or taped fibrous since the process only requires a common
prepegs. In some cases, use is made of a varia- melting furnace and a foundry crucible,
tion of a hot extrusion treatment compression molding fixture, a magazine of spools with
method to obtain a taped foil composite struc- fibers and a system for transportation of
ture reinforced by one or several rows of fibers fibers and handling products. During opera-
and infiltrated tows, as well as of plasma tion the melting furnace must be filled with
sprayed or infiltrated tapes. Sheets and plates argon or evacuated. As a result, the cost of the
are obtained by rolling in a die and extrusion. MMC produced by continuous casting is

_·9
The simplest sections can be produced by close to the cost of the initial fibers and
bending of rolled strips, and more complicated matrix.

•••
•
2 3 4

5
¥6
t'---_ _-----o>I
Fig. 13.6 Typical sections of semifinished items of composite materials; (1) filament with metal coatings
and multifilament braids infilitrated with metals; (2) bars reinforced with braids or filaments; (3) bands
with single-layer or multi-layer reinforcement; (4) pipes and cylindrical housings with longitudinal rein-
forcement; (5) sections; (6) bands and plates.
 Infiltration under pressure 299

The continuous casting method has a num- operations with fibers and the procedure for
ber of advantages from a metallurgical testing of fiber volume fraction are shortened.
standpoint since the composite is prepared The production of aluminum carbon fiber
without formation of typical oxide inclusions for MMC is complicated due to several fac-
which are common in plasma spraying and tors associated with the carbon fibers. Their
hot molding. Most investigators have used small diameter (6-10 nm), friability, high
vertical techniques of continuous casting, but reactability and poor infiltration characteris-
there are versions of horizontal drawing of the tics with aluminum melts at temperatures
fiber through the melt. Before being sub- below 1030°C are all contributory factors. The
merged into the melt, each fiber is detached latter two are most important. The danger of
from the other fibers to ensure intimate contact chemical degradation and reduction of fiber
with the molten metal. Then the fibers pass strength due to graphitization and formation
through the drawing die that determines the of carbides increases with greater contact
cross section of the preform. The fibers are time of the fiber with the melt and with
drawn with a rate necessary for solidification higher temperatures.
of the matrix to fill up the inner space of the
composite preform. For example, in produc-
13.6 INFILTRATION UNDER PRESSURE
tion of cast boron aluminum cable, the
drawing rate could reach 900 m/h, but cannot Infiltration of fibrous preforms makes it possi-
exceed 450 m/h due to fiber defects. ble to obtain intermediate preforms and
The continuous casting process can consist composite structures of limited size.
of two stages: the first stage involves the pro- Infiltration is accomplished by various means:
duction of the composite preform with a small by heating the mold containing the fibers and
cross section, essentially a wire, bar or band; in matrix material in the form of a foil, powder
the second stage, the previously produced or coating on the fibers up to above the melt-
composite preform is integrated into a more ing temperature of the matrix material with
complicated structure with a larger cross sec- the subsequent shaping of the mold; by sub-
tion. For example, in the course of production merging the fibers into the melt or filling
of a boron-aluminum MMC, the most popular molds with the melt in vacuum, protective
form of the initial composite structure is a bar atmosphere or air; by vacuum suction of the
of a circular section, containing 16-19 fibers. In melt; by feeding the melt into the mold under
the second stage, the amount of matrix added pressure; and by centrifugal casting. In all
into the MMC can be very small to keep a high types of infiltration, the common denomina-
volume fraction of reinforcing fibers in the tor is the presence of the mold with the fixed
final product. The two stage process is used fibers. Because the process is long compared
for the production of such profiles as rods, with continuous casting, the fibers should
angle bars, T-beams, I-beams, and lenses. have high thermal stability in metal melts.
The composite resulting from cast lamina Experiments on infiltration with uncoated
may be obtained in one, two or more steps. To carbon fibers and with nickel plating by suc-
form profiles, particular attention must be tion or by filling with melted aluminum or
focused on the first stage since it determines magnesium in vacuum have been unsuccess-
the basic physicomechanical properties, espe- ful due to the failure of the fiber to be
cially specific strength. Because of the infiltrated at the melt temperatures without
organization of the liquid phase process by extensive carbide formation. Common draw-
stages, the amount of manual labor can be backs of vacuum infiltration of carbon fibers
reduced, the equipment simplified and the with aluminum or magnesium alloys are
dead time due to fiber breaks shortened. Also, incomplete infiltration because of nonuniform
300 Metal matrix composites

penetration of the matrix material into the braids of carbon fibers with aluminum melts
interfiber spaces, nonuniform package config- at lower temperatures and higher pressures. It
uration and touching and burning of the has been reported that a two layer coating pro-
fibers where there is no protective coating. duced the best results l4; the first layer of
These drawbacks are especially evident in silicon carbide protects and the second layer of
infiltration of multifilament braids and bands. such transition metals as chromium, molybde-
Uneven distribution of fibers results in defects num, tantalum or tungsten which are properly
in the form of isolated groups of filaments not infiltrated with aluminum improves infiltra-
separated from each other by matrix, thus tion. They also have a lower affinity for carbon
serving as stress concentrators. The result is compared to silicon.
that with an increase in volume fraction of
carbon fibers to more than 30-35%, which
13.7 ROLLING IN VACUUM
does not increase the strength of the compos-
ite material but starts to reduce it, due to the Nondeforming process methods, including
greater number of groups of filaments not ionic and plasma spraying and spontaneous
infiltrated with the matrix. infiltration make it possible to obtain the high
To improve infiltration of the carbon fibers physical and mechanical properties of car-
with aluminum melts with metal coatings bon-aluminum MMC. However, these
(especially nickel plating), conditions must be methods are characterized by low productiv-
chosen to ensure a high rate of flow and to ity and high labor input. The formation of the
considerably shorten the duration of time of structure and properties of carbon aluminum
contact of the coating with the meltl3 . The rela- that has been vacuum rolled within the solidi-
tionship between the capillary and excessive fication range of the matrix aluminum-
pressure and the infiltration angle and volume magnesium alloy has also been studied.
fraction of carbon fiber in the molten alu- Rolling of the initial billets of carbon alu-
minum matrix on the other, taken at a minum was performed on a vacuum rolling
temperature range 925-1083°C, shows that the mill with residual pressure P resid "" 6.5 X 10-1 Pa.
melt penetrates the braids of the fibers only at Powder alloys were used as the matrix, the
a pressure less than 1.1 x 105 Pa. Experimental carbon band, grade 'Kulon' with two-layer
testing shows that, under appropriate pres- coating (silicon carbide and free silicon) was
sure, the rate of infiltration of fibers with used as reinforcement. The temperature condi-
aluminum melt with an 1% (mass) titanium tions of the rolling were characterized as
additive moves at 0.5 m/ s. follows:
The technology of infiltration of carbon
fibers and fabrics with melts using aluminum (13.2)
has been advanced with application of exter-
nal pressure. However, the best tensile where T is the rolling temperature within the
strength values of the carbon/aluminum solidification range of the matrix alloy; TUq and
MMC with V f = 40--60% are not in excess of Tsol are the temperatures of the liquidus and
700-800 MPa, whereas the higher values solidus of the matrix alloy.
(1000-1200 MPa) would be expected from rule The amount of reduction of porosity in the
of mixtures. initial porous billets of carbon aluminum was
To eliminate the strength drawback, it is determined by considering the maximum
necessary to determine the nature of barrier spread and extraction of the part of the matrix
ceramic and process metal coatings which alloy from the billet in compliance with the
ensure protection (compared with the nickel following formula:
plating) and allow effective infiltration of
 Rolling in vacuum 301

= 1-
Ep e [me
/p + (m m - m'm)
(13.5)

where Vf is the volume fraction of fibers in the

me and mm are the masses of the carbon
fibers and matrix alloy in the billet;
Vr
material;
is the volume fraction of the fibers
which are not surrounded with matrix
Pc and Pm are the densities of the carbon
alloy;
fibers and matrix alloy;
10 bo ho are the initial values of the length, Vl is the volume fraction of the fibers
within the volume of the fibrous material
width, and thickness of the billets;
limited by the contours of the fiber braid.
m'm is the mass of the matrix metal extruded
from the billet in rolling. When the metal completely fills the fiber inter-
stices and the fibers are completely well
Since maximum properties of the composite
distributed in the matrix K and R are equal to 1.
material can only be reached with a 100% fill-
Figures 13.7 and 13.8 show the relationship
ing of the interfiber spaces inside the braids of
between factors K and R and the modes of liq-
the carbon fibers and the fibers must be evenly
uid phase rolling for carbon-aluminum. The
distributed within the matrix, the filling (K)
and structural evenness (R) factors of the processes of filling up of the interstices and
the formation of a uniform structure are
fibrous material are:
highly dependent on the temperature and
degree of each reduction in rolling. When
K= (13.4)
there is a low amount of liquid phase, the

0.8r------+---+----+---l---~

::.t:: 0.6
~
c:
(Il
e
(Il
Co 0.4
c:
0
'00
~
0.2

o 0.2 0.4 0.6 0.8 1.0

Temperature Coefficient, 'I'

Fig. 13.7 Infusion percentage, K of braids CF MCM Al-C for various rolling modes: (1) E = 1.0 E . (2) E =
0.75 Epi (3) E = 0.5 Ep; growth = 6.65 X 10--1 Pa. P'
302 Metal matrix composites

II:
0.6
~
CD

~CD
0
()
~ 0.4
E...
~
c:
:::l
...!..
!!!
:::l
U 0.2
2
en

o 0.2 0.4 0.6 0.8 1.0

Temperature Coefficient, 'If

Fig. 13.8 Uniformity, R of structure of MCM Al-C for various rolling modes: (1) E = 1.0 Ep; (2) E = 0.75 Ep;
(3) E =0.5 Ep; growth =6.65 X 10-1 Pa.

braids of the carbon fiber are sealed without N-NR

CR= (13.6)
filling up of the interfiber spaces. Thus, 1/Jn > N
0.6 due to an avalanche-like decreasing level
of hydrostatic pressure at the deformation N-N"
Cv= (13.7)
point and efficiency of filling up of the braids N
is reduced. The latter is associated with phys-
ical changes in the billet during deformation where:
from the solid body with limited inclusions of
the solid phase. Increasing the rate of rolling CR is the relative number of the kinked
and increasing the number of the steps to the fiber"s;
same total reduction improves the conditions CV is the number of the broken fibers;
for forming a uniform structure in carbon N is the number of the fibers in the specimen;
aluminum, and the factors K and R approach NR is the number of the fibers that have lost
unity. their original cross-sectional shape
The highest quality composite must be (kinked);
sealed. During rolling, within the solidifica- N" is the number of broken fibers.
tion range of the matrix alloy, the number of During these experiments, it was found that
shrinkage pores obtained in the composite the probability of fibers breaking during solid
requires seal rolling of the matrix in the solid phase rolling can be reduced by increasing
phase. However, the probability of breakage number of steps before sealing.
and kinking of fragile carbon fibers also When these process requirements are satis-
increases. The degree of breakage and kinking fied by rolling under vacuum, carbon
of fibers has been estimated using relative fac- aluminum composite with the following
tors during microstructural research of the structural characteristics can be obtained:
rolling specimens: fiber braids are filled to 95-97%; structural
 Rolling in vacuum 303

uniformity is 78--94%; and the number of the rolling must be accomplished at a vacuum
defective fibers is 4-10%. High quality rolled of 6.65 x 10-1 Pa, and compacting of the initial
carbon aluminum is most dependent on the powder-like billets of the carbon aluminum
percentage of defective fibers. In a 45% Vf must be performed. with rigid, plastically
composite, the properties in the direction of undeformed casings.
reinforcement are as follows: tensile strength, However, this technology is unsuitable for
825 MPa; bending strength, 1300 MPa; and production of combined composites. For
the modulus of elasticity is 200 GPa. A two example, during production of Ti-Al-C by this
layer protective coating on the carbon fiber, technique, intimate contact of the layers of
with a small amount of aluminum carbide titanium with carbon-aluminum is not real-
(~0.1%), retains the transverse strength of the ized; whereas, during rolling of fibrous
reinforced composite (at = 55 MPa). material AI-B-C, an intensive interaction of
Application of a protective chromium coating the boron fibers with the molten aluminum
on the carbon aluminum by condensation ion takes place. Thus, in the production of com-
bombardment also increases its resistance to bined composites, the criterion requirements
corrosion. are necessary with at least an additional
However, carbon aluminum is somewhat requirement for intimate connection of the
unsatisfactory in economic terms because of its separate hybrid components. To produce a
inherent problems with soldering, welding and diverse combination, such as MCM AlB-MCM
mechanical connection with other materials. Al-C, the number of requirements doubles.
These drawbacks can be eliminated by However, research shows the requirements
using combined composites of the following can be met by non-traditional manufacturing
systems: boron-aluminum - carbon-alu- approaches.
minum (high compressive strength, good The combination Ti-AI-C is an example of
resistance to erosion); titanium - carbon-alu- the product of such an approach. The bimetal
minum, dispersively reinforced aluminum titanium-aluminum composite can be success-
(filling agent: SiC, Al4C3) carbon-aluminum fully produced by rolling at temperatures of up
(high hardness, transverse strength). to 500°C (932°F), with high degrees of defor-
There are specific criteria for production of mation. The requirement for preservation of
such composites using the method of rolling fiber continuity and the elimination of the car-
along the fibers. It is advisable to use some of bon-aluminum reaction decreases the
these criteria to optimize the manufacturing probability of formation of an intimate connec-
process 15• The basic requirements for carbon- tion between titanium and aluminum layers
aluminum are: sealing of the porous billets during production of the combined composite.
during extrusion; 100% filling of the interfiber Also, during compaction of carbon-aluminum
spaces with matrix alloy; provide for compati- within the solidification range of the matrix
bility; and prevent component oxidation. alloy, the probability of formation of friable
Research considering the influence of intermetallic layers between titanium and alu-
rolling manufacturing parameters on meeting minum is very high. Thus, the application of
criteria requirements makes it easier to select liquid phase methods (infiltration, soldering,
an optimum technology for production of car- and rolling above the alloy solidus tempera-
bon aluminum. The main feature of this ture) makes it possible to obtain a high quality
technology is that rolling is performed over combined composite. For the combined com-
10-12 steps at a temperature corresponding to posite, the following manufacturing methods
the median value of the solidification range of were used. The pack billet consisted of layers
the matrix alloy (~ 40-50 vol. % of the liquid of the matrix alloy, titanium foil, and carbon
phase). After solid phase additional sealing, bands and was heated to a temperature
304 Metal matrix composites

exceeding the liquidus temperature in the vac- the braids of the carbon fibers, as well as the
uum chamber of the rolling mill. Preferred infiltration of the porous monolayer of boron
evaporation of magnesium from the aluminum aluminum with molten melt, but without direct
alloy with subsequent deposition on the con- contact with the boron fiber, takes place. After
tact surface of titanium, as well as cooling below the solidus point of the carbon-
simultaneous infiltration of the carbon bands aluminum additional reverse rolling results in
took place. After treatment, the billet was preferred sealing of the boron-aluminum.
cooled down to the carbon-aluminum com- Analysis of the microstructure and mechan-
pacting temperature and was vacuum rolled. ical properties of the obtained material shows
Deposition of the protective magnesium coat proper weldability and filling of separate
does not form chemical compounds with tita- monolayers, with a minimum number of the
nium. The infiltration of the carbon framework broken fibers. This combined composite
with subsequent welding-soldering of tita- shows better mechanical properties than car-
nium with matrix alloy during rolling ensures bon-aluminum. Thus, for a one step filling
that bonding bridges are formed during the technique, the combined composite has
liquid phase high temperature contact of com- 10-15% greater tensile strength than the car-
ponents at the moment the infiltration is bon- aluminum composite and the transverse
completed. strength was increased by a factor of 2 (20% B).
Plate and layer reinforced carbon aluminum Combined fibrous dispersively-reinforced
with high resistance to corrosion, with trans- composites AI-C-SiC and Al--C-Al4C3 have
verse strength exceeding that of the aluminum been obtained by careful selection of the matrix
by 1.5-2.0 times, has been obtained by this tech- alloy or by using an alternative manufacturing
nology. This composite can also be successfully processes. For these composites, adherence to
attached to titanium structures by soldering or the criterion requirements is relatively simple.
welding. The strength of the titanium to matrix It is possible to produce sealed monobands
alloy bond is higher than the strength of the from the dispersively reinforced composites
bond between the carbon fibers and the matrix (AI-SiC, AI-C-Al4C3) with subsequent com-
(attempts to peel the titanium result in fracture pacting of the composite using the techniques
between the aluminum and the carbon). developed for compacting carbon-aluminum.
To obtain the combined composite Al-B-C, Data developed on compatibility of carbon
the following ground rules have been estab- fibers with barrier (protective) coatgtgs and
lished: compacting of plasma porous aluminum and nickel matrixes, as well as data
semifinished boron aluminum items should be on strength as a function of a barrier coating
performed in the solid phase at temperatures thicknesses, have facilitated the development
exceeding 560°C (1040°F)lS; and carbon-alu- of a number of MMCs.
minum should be compacted in the middle of Properties for some of these MMCs are pre-
the solidification range of the matrix alloy. An sented in Table 13.3. These materials were
alloy of AI with 10% Mg with a solidus temper- made with 'Kulon' carbon fibers with Russian
ature of 510°C (950°F) was used as a matrix for production barrier coatings. Data on these
carbon-aluminum, but a higher temperature fibrous materials both coated and uncoated
alloy was used as a matrix in the boron-alu- are shown in Table 13.4. Figures 13.9 and 13.10
minum. The assembled pack billets of show articles made of aluminum--carbon and
boron-aluminum and carbon-aluminum uni- nickel-carbon composites. Although MMC
directional1amina were rolled at a temperature properties are excellent, higher strength car-
of 540-550°C (1004-1022°F) for 9-10 steps (the bon fibers will permit further increases of both
compacting mode for carbon-aluminum). With specific and absolute strength of the newly
this procedure, filling of the interfiber space in created MMC.
 Rolling in vacuum 305

Table 13.3 Some properties of carbon reinforced metal-matrix composites

Property Matrix
AL-9 AMG-6 Ni Cu
Filler Kulon Kulon VMN-4 VMN-4
Tensile strength Mroin' GPa 2.5 2.5 2.0 1.5
Elastic modulus, min, GPa 40.0 40.0 25.0 23.0
Coating TiC SiC Ti(Zr)C
Metal-matrix composite

Density, p, kg m-3 2000-2300 2200-2300 5200-5400 5200-5500

Tensile strength, MFa
at 20°C 900-1000 700-900 450-500 400-500
at 400°C 800-900
at 800°C 400-500
Bending strength, MPa 900-1050 800-1100 700-900 600-700
Compressive strength, MPa 800-1000 500-700
Elastic modulus, E, GPa 270-300 240-270 140-180 120-170
Specific modulus, 135 109 26 23
E/p, m 2 s-2 x 106

Table 13.4 Some properties of carbon fibers without coating and with barrier coating

Properties VMN-4 VMN-4SiC VMN-4TiC Kulon-SiC

Density, g cm-3 1.72 1.72 1.72 1.91
Ib in-3 0.062 0.062 0.062 0.069
Filament diameter, 11m 6.5 6.5 6.5 6.5
Elongation at break, % 1.10 1.10 1.10 1.10
Carbon content, wt. % ~8 ~5 ~95 ~94
Coating w/c SiC TiC SiC
Coating thickness, nm 60-80 50-80 60-80
Specific surface area, m 2 g-l g.O ~0.5 ~0.5 ~0.5
Longitudinal tTE x 1O~ 2.25 (1.2) 2.25 (1.2) 2.25 (1.2) 2.25 (1.2)
at 21°C (70°F)
Tensile strength, MFa (psi) ~2000 ~300 ~2300 ~500
(290000) (334 000) (334000) (363000)
Tensile modulus, GPa (psi) ~230 ~30 ~230 ~400
(33 x 106) (33 X 106) (33 X 106) (57 X 106)
Typical Strand Properties:
Threads / strand 24 24 24 24
Filaments/thread 300 300 300 300
Twist, tpm 100 100 100 100
(tpi) (2.54) (2.54) (2.54) (2.54)
Twist / strand 2
Filaments / strand 7200 7200 7200 600
306 Metal matrix composites

Fig. 13.9 Various types of shaped articles made of alu- Fig. 13.10 Ring made of the composite
minocarbon composites material: (1) shell; (2) tube; (3) material nickel-carbon fiber.
gas-turbine blade; (4) T-profile.

REFERENCES 10. Clement, J.R., Kock, H.J., Wu, K.T., Spenser,

H.G. Interfacial modification in metal matrix
1. Lee, S.M. International Encyclopedia of Composites. composites by sol-gel process. Mater. Manu!
New York: VCH, pp. 187-358. Process, 1990, 5(1), 17-33.
2. Fridlyander, LN. Metal Matrix Composites (Soviet 11. Syrkin, v.G. Chemistry and Technology of Carbonyl
Advanced Composites Technology Series). London, Materials. M. Chemistry, 1973, p. 185.
Chapman & Hall. 12. Moskalenko, A.G. et al. Kinetics of carbon fiber
3. Metal matrix composites. The next generation and aluminium contact interaction. Mekhanika
of high-performance materials. Emerging Polymerov. 1977, No 4, 42-43.
Technologies, No 20 Technical Insight Inc., 13. Varenkov, A.N., Kostikov, V.L Vliyuanie
Englwood/Fort Lee, NF., May 1986, p. 190. cmachivania grafitovykh volokon na ysloviya
4. Bunsell A, Gorg, M. Les matriaux composites a izgotovleniya materialov metodom vakuum-
matrice metallique an Japan. Industrie noy propitki. (Influence of wetting graphite
Ceramique. 1988, No 830, 642. fibers on conditions of manufacturing of com-
5. Zimick D.C., Koik B.M. Design of thermally sta- posite materials by using vacuum impregnation
ble graphite/aluminum tubular structures for method). In Voloknistye i dispersno-uprochnenye
space applications. SAMPE Q., 1990, 21(2), 11. Twmpozitsionnye materialy (Fibrous and Disperse/y-
6. Strength optimization of C/ AI metal matrix Reinforced Composite Materials). M. Science, 1976,
composites produced from prepregs. Ph.D. pp.25-28.
Thesis. Masson J.-J. DIR-FB-92-39, ETN-93- 14. Shorshorov, M. Kh. et al. Methods of improve-
94871. 1992. p. 110. ment of infiltration of carbon fibers with
7. Ivanov, VS. (ed.). Aluminium and Magnesium aluminium. Adhesion Melts Soldering Mater.
Alloys, Reinforced Firbres. M. Science, 1974, p. 202. 1976, No 1, 86--89.
8. Fukuda, Sh., Matsubara, T., Tokao, Yo. A fabri- 15. Shorshorov, M. Kh. et al. Electronic and micro-
cation process for fiber-reinforced composites scopic researching of carbonyl metal coatings
by using a wire explosion spraying method. carbon fibers for reinforcement of composite
Carbon fiber reinforced aluminum preform materials. Phys. Chem. Metal Treatment. 1976, No
sheets. Yosha, 1990, 27(1), 35-40. 4,141-143.
9. Dergunova V.s. et al. Protective coatings for car-
bon fibrous materials. In: Anticorrosive Coatings.
L. Nauka, 1983. pp. 164-168.
CERAMIC COMPOSITES 14
M. F. Amateau

14.1 INTRODUCTION The mechanical characteristics of ceramic

composites are profoundly influenced by the
Ceramics generally are compounds of metallic
mechanisms of strengthening and toughening.
or non-metallic elements and other non-metals
Successful design of ceramic composites
such as oxygen, nitrogen, carbon and boron.
requires an understanding of the role of the
Compared to metals these compounds have
constituents, including reinforcement, matrix
higher melting temperatures, higher Young's
and interphase, on the these mechanisms. In
moduli and hardness, lower densities and
addition to the properties and form of the con-
lower electrical and thermal conductivities.
stituents, the characteristics of the processing
Engineered ceramics are used in thermal and
method can also have a major impact on
structural applications requiring high temper-
mechanical and physical properties.
ature resistance, high hardness and chemical
inertness. Applications that exploit the ther-
mal structural properties of ceramic 14.2 CONSTITUENT MATERIALS
commonly include cutting tool inserts, wear
14.2.1 REINFORCEMENI'S
resistant components, ballistic armor, heat
exchangers, burner tubes, prosthetics, dental Fibers used in ceramic matrix composites fall
implants, heat engine components and ther- into three general categories based on their
mal barrier coatings. The dominating diameter: monofilaments, textile fibers and
characteristics of ceramics that limit these and whiskers. In addition, reinforcements in the
other engineering applications are their lack of form of particulates and platelets are used in
plastic behavior at room temperatures and ceramic composite designs. The strengthening
their low tolerance to flaws, i.e. low fracture role of the reinforcements in ceramic compos-
toughness, that lead to catastrophic failure. ites is significantly different from that in
Reinforcing ceramics with particles, polymer and even metal matrix composites. In
whiskers, platelets, discontinuous fibers and metal and polymer composites the reinforce-
continuous fibers significantly improves their ments contribute directly to the increase of
strength, toughness and apparent ductility. strength and stiffness by carrying a significant
Composite design can also be used to tailor portion of the load. In ceramic composites the
other important properties such as high tem- reinforcements usually increase the strength
perature strength and thermal shock indirectly by increasing toughness of the
resistance, wear resistance and low friction, matrix. The load carrying capabilities of the
thermal and electric conductivity and thermo- reinforcements are of secondary interest at
elastic properties. most. This places special importance on the
whiskers as reinforcements for ceramic com-
Handbook of Composites. Edited by S.T. Peters. Published posites.
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7
308 Ceramic composites

Whisker reinforcements that can profoundly influence their interfacial

properties and hence effectiveness as a rein-
Whiskers are a very fine filamentary form of a
forcement in a ceramic matrix.
material. They are usually I jlffi or less in
Surface layers of the whiskers can contain
diameter and up to 200 !lm in length. They are
crystalline Si02 or amorphous Si-O-C (so-
nearly perfect single crystals with strength
called 'black glass')1. Surface compositions can
properties approaching the theoretical
also be modified by chemical treatments such
strength of the material. Figure 14.1 is a scan-
as HF washing.
ning electron microscope image of silicon
carbide whiskers. As a reinforcement for
ceramic matrix composites it is their size and Ceramic textile fiber reinforcements
aspect ratio that is most significant to their
Although significant improvement of mechan-
strengthening effect.
ical properties of ceramic materials can be
The whiskers that have recently been com-
achieved by whisker reinforcement, truly
mercially available are listed in Table 14.l.
mechanically robust and reliable structures can
Silicon carbide whiskers obtained from differ-
only be approached with continuous reinforce-
ent sources have unique surface compositions
ments. Continuous ceramic fibers of sufficient
strength and optimum stiffness and diameter
are now commercially available. They are
capable of being filament wound, woven or
braided into textile fonns. They are composi-
tional variations of three systems: AIP3' Si-C
and Si-N. Table 14.2 lists the composition and
selected properties of fibers that are commer-
cially or near commercially available.
The usefulness of these fibers can be mea-
sured by their high temperature properties and
their ability to be shaped into complex pre-
forms for subsequent infiltration. A convenient
measure of the ability to handle these fibers for
processing is their critical bend radius, Pc' i.e.
Fig. 14.1 Scanning electron microscope image of
SiC whiskers.
the smallest radius that the fibers can be bent
before they fracture. A simple beam strength
Table 14.1 Selected properties of commercially available whiskers1,2

Whisker Crade Source Diameter Length Young's Coefficient Density

(mm) (mm) modulus of thermal (g cm-3)
(CPa) expansion
(1 (T6jOC)

SiC SC-9 Advanced Composite 0.6 10--80 600 4.5 3.2

Materials Corp.
SiC Tokamax Toaki Carbon 0.3--0.6 50-200 400-700 5 3.2
Company
SiC XPW-2 Huber Company 0.05-1.0 5-100 600 4.5 3.2
Si3N4 SNWB UBE Industries 0.05-0.5 5-100 370 2.5 3.2
A1 20 3 Catapal XW Vista Chemical 4-7 40-100 400 7 4.0
Company
 Constituent materials 309

Table 14.2 Commercially or near commercially available continuous ceramic fibers

Manufac- Designation Composition Tensile Tensile Density Diameter Critical

turer strength modulus (g cm-3) ()lm) bend
(MPa) (CPa) radius
(mm)
DuPont FP >99 a-AI:p3 >1400 385 3.9 20 2.75
Sumitomo Altex 85 Al:P3' 15 Si02 1800-2600 210-250 3.2 9-17 0.53
3M Nextel 73 AI:P3' 27 Si02 2000 193 3.03 10-12 0.48
550
3M Nextel 62 AlP3' 14 Bpy 15 Si02 1750 154 2.7 11 0.48
312
3M Nextel 70 AlP3' 2 Bpy 28 Si02 2100 189 3.05 10-12 0.045
440
3M Nextel 70 AlP3' 2 Bpy 28 Si02 2275 224 3.05 10-12 0.049
480
DuPont FP AlP3' 15-25 Zr02 2100-2450 385 4.2 20 1.83

°
166
Nippon Nicalon 50 Si, 31 C, 10 2520-3290 182-210 2.55 10-20 0.36
Carbon
Ube
Dow
Tyranno
MPS
Si, TI,C, °°
69 Si, 30 C, 1
>2970
1050-2450
>200
175-210
2.4
2.65
8-10
10-15
0.27
0.83
Corning
Dow
Corning
MPDZ 47 Si, 30 C, 15 N, 8 ° 1750-2450 175-210 2.3 10-15 0.50

Dow
Corning
HPZ 59 Si, 10 C, 28 N, 3 ° 2100-2450 140-175 2.35 10 0.33

Tonen Silicon 60 Si, 37 N, 30, 0.4 C 2500 300 2.5 10 0.60

nitride

analysis will show that the p cr can be calculated studies are critical to developing proper com-
by multiplying the failure strain of the fiber by posite fabrication methods since almost all
the radius of the fiber. Thus high strength, low processing techniques require high tempera-
elastic modulus and fine diameters produce the ture processing. Both FP and PRD166 fibers
most robust fibers. The first successful alumina can be heated to lOoooe before any serious
based textile fiber was DuPont FP fiber. This reduction in strength occurs3. The boria modi-
fiber has a very high elastic modulus but its low fied alumina-silica (Nextel 312) and mullite
strain to failure results in very large critical compositions (Nextel440 and Nextel480) are
bend radius which has limited its application. A also degraded with high temperature heat
modification of the all alumina fiber by adding treatments as seen in Fig. 14.24. The high tem-
partially stabilized zirconia, the DuPont FP 166 perature degradation in strengths and elastic
fiber, has reduced the critical bend radius by moduli of the boria modified alumina silica
one third. However, only the silica modified bases fibers are summarized in Table 14.3. The
alumina based fibers such as the 3M Nextel retention of properties at high temperatures
fibers and the Sumitomo Altex fiber have criti- following the relative degree of crystallinity
cal bend radii of about 0.5 mm. among the three grades with Nextel312 being
Numerous studies have been performed on the least crystalline and Nextel 480 being the
these fibers to determine their strength and most crystalline.
modulus retention after heat treatment. These Strength and Young's modulus at tempera-
ture of the Sumitomo alumina silica fibers !tas
310 Ceramic composites

1000 •
312

.
::::;:; . ..
~.-­
~

440

---:'----...
.---=~~~::'i. :~J.---
~
.---. ~ ~

1000 • air
• hydrogen
~vacuum

700 800 900 1000 11 00 1200 1300 1400

HEAT TREATMENT TEMPERATURE. °C

Fig. 14.2 The effect of heat treatment temperature on the strength of alumina and mullite fibers (440 and
4800)4.

Table 14.3 High temperature degradation of boria modified alumina silica fibers4

Property Benchmark Comparative temperature (OC)

Nexte1312 Nextel440 Nexte1480
Strength after heat treatment 50% of initial
in air >1300 1400 >1400
in hydrogen 1200 1100 1200
in vacuum 1200 1350 1400
Hot strength 50% of initial 1000 1120 1200
Hot elastic modulus 50% of initial 950 1000 1250
Creep rate lO-4/h at 69 MPa 800 1010 1190

been measured by Bunse1l5 up to 1200°C. Both of a silicon carbide based textile fiber for com-
the strength and elastic modulus begin to posite reinforcement. All commercially
degrade at BOO°C while above 1000°C the available fibers in this category contain oxy-
properties drop significantly. gen but can also contain nitrogen and
During the last 20 years there have been titanium. Nicalon fiber manufactured by
significant advancements in the development Nippon Carbon Company and marketed in
 Constituent materials 311

the USA by Dow Coming Corporation is by Monofilament reinforcements

far the most commercially developed. X-ray
Monofilament SiC and boron fibers are pro-
diffraction analysis indicates that Nicalon
duced by chemical vapor deposition onto a
consists of ultra fine ~-SiC particles dispersed
fine substrate filament. For the case of SiC
in a matrix of amorphous Si02 and free car-
fibers the core is 37 !lm amorphous carbon fil-
bon6 • Nicalon has excellent resistance to
ament, while for boron a 13 Jlm tungsten wire
thermal degradation in argon and air expo-
has been commonly used. The principal
sure at temperatures to 1000°C for as long as
advantage of monofilament reinforcements
100 h7. The loss of tensile strength for Nicalon
are their ability to tolerate some degree of sur-
by exposure to temperatures to 1400°C in both
face reaction with the matrix during
air and <'ITVnn <'ITf> nrf>"f>ntf>C1 in Fiv_ 14_~_
fabrication or high temperature service. These
fibers can be infiltrated by a number of pro-
cessing methods including powder sintering,
l 3
• 10--...,• BOOoC powder hot pressing, plasma spraying and
melt infiltration. These fibers are limited to

.
~
•
'" structures with relatively simple shapes such
-
t;)
1000"C
J! as sheet, plates and large diameter cylinders
!:II 2
c -.", because of their large critical bend radius.
! Table 14.4 lists the properties of these fibers.
-; 1200°C
.! .~.
'ii 1
c Table 14.4 Properties of monofilament
~
-OrIginal fiber
• 14OO"C reinforcements

Boron SCS-6
0
Cal Manufacturer Textron Textron
Composition B SiC
Tensile strength (MPa) 2.5 4.3
Tensile modulus (GPa) 400 427
Density (g cm-3) 2.5 3
l 3
*
0
0
Diameter (Jlm) 140 140
~

~
Critical bend radius (mm) 11 7
t:>
J! 1000°C ~

0""
til
c 2
0 -0
~~12000C LI
14.2.2 MATRIX MATERIALS
!
ii
.! o The selection of matrix materials for ceramic
"ii o-~ 1400"C 0
c 1 composites is strongly influenced by thermal
~ stability and processing considerations. The
properties of matrix materials commonly used
in ceramic composites are shown in Table 14.5.
o These include oxides, carbides, nitrides,
10 1~ 1tr
borides and silicides. The first indication of the
Heat treatment time, t (h)
(b) ability of a material to resist high temperature
service is melting temperature. With the
exception of glass all these materials have
Fig. 14.3 Loss of strength of Nicalon™ after expo- melting temperatures above 1600°C. As the
sure to (a) argon and (b) air at temperature to melting temperature increases the ease of pro-
1400°C7. cessing decreases.
312 Ceramic composites

Table 14.5 Properties of typical ceramic matrix materials

Materials Young's Poisson's Modulus of Fracture Density Thermal Melting

modulus ratio rupture toughness (g cm-3) expansion point (OC)
(CPa) (MPa) (MPa m1!2) (lo-6/"C)
LAS 117 0.24 138 2.42 2.61 5.76
Pyrex 48 0.20 55 0.08 2.23 3.24 1252
~03 345 0.26 483 3.52 3.97 8.64 2050
Mullite 145 0.25 186 2.20 3.30 5.76 1850
Zr02 PS 207 0.23 648 8.46 5.75 7.92 2760
Zr02 FS 207 0.23 248 2.75 5.56 13.5
Ti02 283 0.28 83 2.53 4.25 9.36 1849
S~N4SN 310 0.24 496 5.60 3.18 3.06 1870
Si3N4RB 165 0.24 303 3.41
Si~4HP 310 0.24 827 5.60 3.19 3.06 1870
Si02 76 0.16 0.77 2.20 0.54 1610
SiCSn 331 0.19 386 4.94 3.21 4.32 1980
SiCHP 414 0.19 462 4.94 3.21 4.32 1980
B4C 290 310 2.41 3.06 2350
TiB2 552 0.20 896 6.92 4.62 8.10 2900
TiC 427 0.19 248 4.92 8.46 3140
TaC 283 0.24 200 14.50 6.66 3880
BeO 359 0.24 234 3.00 5.76 2530
WC 669 0.20 15.80 4.50 2870
Cr20 3 103 262 3.85 5.21 7.56 2435
Cr3C2 386 0.20 6.70 9.67 1890
BNL 34 76 1.94 6.66 2982
76 110 1.94 0.36 2982
:~ 448 0.21 7.82 6.66 3499

Mechanical and chemical compatibility of the 14.3 PROCESSING METHODS

matrix with the particular reinforcement ulti- Processing of ceramic composites can be per-
mately determines whether a useful formed by solid, liquid, or gas phase
composite can be made. For the case of processing of the matrix material to achieve
whisker reinforced composites the chemical infiltration of the matrix around the reinforc-
reactions with matrix are particularly critical
ing phase. The goals in processing ceramic
since even minor reactivity can consume the
composites are to achieve minimum porosity
entire reinforcement. Large differences in the with a uniform dispersion of the constituents
coefficients of thermal expansion between and controlled bonding between the reinforc-
reinforcement and matrix can lead to large
ing phase and the matrix.
residual stresses during the fabrication and
ultimately result in serious degradation of
mechanical strength. Small or optimum differ- 14.3.1 POWDER PROCESSING
ences can be beneficial to mechanical
Fundamental steps in processing ceramics
properties by placing the weaker constituent composites from powdered constituents are:
in compression or by inducing crack deflection
between reinforcements. • powder selection;
• powder characterization;
• agglomerate reduction;
 Processing methods 313

• constituent mixing; To minimize voids and interfacial weakness

• green body fabrication; and maximize the toughening effect of the
• green dressing (machining and gate re~forcing phase, a uniform finely dispersed
removal); mIxture must be produced. Arranging the con-
• binder removal; stituents to minimize free space between them
• consolidation and densification; is referred to as 'packing'. When the con-
• final dressing (burr removal); stituents are not effectively packed,
• inspection. subsequent densification becomes difficult
requiring higher pressing temperatures, pres~
The selection of constituent powders is the
sures and duration. Both constituent shape
first step in composite design and considera-
and particle size difference can affect packing.
tion must be given to chemical, mechanical
~ptim.urn. pa~king occurs when the particle
and thermoelastic compatibility between the
SIze diStri?utton contains 30 vol. % of very
constituents as well as the desired final
small particle and 70 vol. % of large particles9 .
mechanical and physical properties of the
If uniform round fibers (or whiskers) are per-
composite. In addition to the obvious problem
fectl~ align~d in a closed packed array then
of reaction between constituents, other incom-
matrIX partIcles approximately 0.15 times the
patibilities such as large differences in melting
fiber diameter would theoretically fit in the
temperature of matrix and reinforcement can
interstices.
preclude successful processing. Thermal
Most ceramic powders can be comprised of
expansion mismatch between constituents can
a mixture of primary particles and agglomer-
cause premature failure in them or at their
ates. Agglomerates are primary particles
interface.
bonded by surface chemical forces, electrosta-
The rational selection of constituents usu-
tic ~orces o~ solid bridging. In order to produce
ally requires knowledge of certain physical
a fmely dIspersed, homogeneous mixture of
characteristics such as particle size distribu-
matrix and reinforcement successfully, the
tion, ~hape, specific surface area, bulk density,
agglomerates must be reduced. A typical
electncal charge, impurities, etc. The ultimate
agglomerate (mullite in this case) is shown in
aim of such constituent characterization is to
Fig. 14.5. The agglomerate is 8-9 Jllll in diame-
predict the final characteristics of the ceramic
ter while the constituent primary particles
composite, as shown schematically in
range from 0.1 to 1.5 Jllll. To uniformly incor-
Fig. 14.48 •
porate 0.5-1 Jllll diameter whisker or particles,

IIlll TillE
--~MIIIIf

'---~""MI'" ,---."Ad}_
GRINDING AIDS FORIlING
P..R..IIEYERS

Fig. 14.4 Use of powder characterization in process controlS.

314 Ceramic composites

Fig. 14.5 Typical agglomerate found in mullite powders.

the large agglomerates must be broken down (5 % or lower) sintering may produce satisfac-
by mechanical action or chemical treatments if tory results. For complete densification of
the agglomeration is due to surface forces. If even low fiber volume fraction composites,
the whiskers are robust or some degree of sintering may impractical due to excessive
whisker breakage can be tolerated, both temperatures and durations. For high fiber or
agglomerate reduction and constituent mixing whisker volume fractions, hot pressing and
can be accomplished simultaneously by ball hot isostatic pressing are the only effective
milling. methods for densification. Table 14.6 shows
Organic binders are usually mixed with the the effect of hot pressing time and temperature
particle-whisker mixture for near-net-shape at 31 MPa pressing pressure on the theoretical
processing by a variety of cold forming opera- density of SiC whisker-Al20 3 composites for
tions including uniaxial pressing, cold various vol.% whiskers. Theoretical densities
isostatic pressing, tape casting, extrusion, of over 99 vol. % can be achieved in unrein-
compression molding and injection molding. forced AIP3 at 1500°C, in lO%SiC whisker
The ceramic preform after cold consolidation composites at 1650°C and in 20%SiC whisker
is referred to as the green' form. The part in
I composite at 1800°C10.
the green form can usually be machined with-
out damage. In this state additional
14.3.2 LIQUID PROCESSING
near-net-shape processing can be applied such
as gate removal and machining. When high temperatures and mechanical
Final consolidation and densification is per- forces are used to consolidate composites from
formed at high temperatures. Three of the the powder, the optimum strength properties
most common methods are sintering, hot (uni- can be sacrificed. Reducing processing tem-
directional) pressing and hot isostatic perature, time and pressure can minimize
pressing. For low fiber or whisker contents damage to the reinforcements but fully dense
 Processing methods 315

Table 14.6 Effect of processing parameters on the theoretical densities of SiC

whisker-Alp3 composites pressed at 31 MPa pressure lO

Vol. % Pressing temperature Pressing time Density Density

whiskers (OC) (min) (g cm-3) (% theoretical)
0 1500 60 3.95 99.1
10 1500 35 3.78 96.7
10 1650 60 3.89 99.5
20 1500 120 3.68 96.1
20 1650 60 3.72 97.1
20 1725 60 3.78 98.6
20 1800 25 3.81 99.5

or near-fully dense composites cannot gener- SrO·AIP3·Si02 were infiltrated into SiC pow-
ally be produced. Processing by infiltration der preforms with and without SiC whiskers
with a molten matrix would be an ideal way of with resulting open porosity of about 1%12.
minimizing mechanical damage and still Recently considerable attention has been
achieve fully dense structures. The melting applied to directed melt gas-metal reactions
temperatures of ceramics used as matrices in which produce ceramic matrix composites
composites limits the general use of melt infil- directly from the liquid metaP3, 14. Both metal
tration as a viable processing route. However, oxide matrix and metal nitride matrix com-
by careful tailOring of the matrix and the use of posites have been produced by this technique.
innovative in situ reaction techniques, melt Net shape composites can be processed at tem-
infiltration has been successfully utilized to peratures of the melting temperature of the
fabricate ceramic composites. metal. The commercial development of this
Glass and glass-ceramic matrices have been processes is called the DIMOXTM process of the
successfully infiltrated in the liquid form into Lanxide Corporation. In this process oxidation
fiber preforms by 'matrix transfer mold- or nitridation occurs on the surface of the
ing'll.The high temperatures required to molten metal forming a layer of solid ceramic.
achieve the appropriate fluidity of the matrix The layer thickens as the molten metal wicks
limits the available fiber-matrix compositions up between the grains of the ceramic. A
to only those with low mutual reactivity. Other schematic representation of the process is
matrix materials such as CaSiOy SrSi03 and shown in Fig. 14.614 • The phenomenon is made

Oxidant gas
~~".-+--- Reinforcement preform

- + - - - Reinforcement preform
entrapped in solid
reaction product

Fig. 14.6 Directional metal oxidation method for processing ceramic composites4 •
316 Ceramic composites

possible by dopants which modify the surface 14.3.3 VAPOR PROCESSING

energy between the phases. For instance, if the
The infiltration of the reinforcing phase by a gas
grain boundary energy, YB' is less than twice that decomposes to form the solid matrix phase
the energy of the solid-liquid interface, YSL and
is generally referred to as 'chemical vapor infil-
the surface energy of the solid-liquid inter-
tration' (CVI). Various carbides, nitrides, oxides
face, Y is greater than surface energy of the
and borides17 as well as unreacted carbon18
solid-~~por interface, Ysv' thickening of the have been deposited on silicon carbide-based
oxide (or other reaction compound) layer does
yam fibers (e.g. Nicalon and Tyranno fibers),
not take place. By reversing the relative values
oxide based fibers (e.g. Nextel fibers) carbon
of the surface energies, (i e, YB > 2YSL and YSL < yam fibers and SiC whiskers. Silicon carbide is
Y ) as illustrated in Fig. 14.7 the unstable grain
one of the most commonly applied matrices
b~undary permits wicking of the liquid metal using CVI. Methyltrichlorosilane is reacted
through the grain boundaries of the reaction
with hydrogen on the surface of the fiber to
product phasel4 . •
deposit silicon carbide. A typical reaction for
Application of this technology to cerar:ruc
this process is19 •
matrix components is achieved by allowmg
the reinforcement preform to float above the CH3SiCl 3 + H2 ~ SiC + 3HCl +H2
liquid metal bath for infiltration of the molt.en
This reaction can take place by conventional
metal. A growth barrier can surround the rem-
chemical vapor deposition (CVD) at tempera-
forcement preform to produce practical net or
tures of 1000-1400°C. Silane-hydrocarbon
near net sh ape component s15' 16.
(SiH -C H ) mixtures can be used to deposit
4
SiC at x y
temperatures below 5000Cby PI asma
assisted chemical vapor deposition (PACVD).
Table l4.7lists some of the more commercially
important matrix materials that can be ap~lied
Air by conventional CVD. A more complete hst of
It"fsv ceramic materials produced by both conven-
Oxide
tional CVD and PACVD can be found in
reference (20).
~"fSL
Molten Metal Table 14.7 Ceramic materials formed by CVI
processes21
(a)
Matrix Reactant Reaction
ceramic gases temperature (0C)
Air
ItYsv TiC TiCr CHr H 2 900-1000
SiC CH3SiCI3-H2 1000-1400
~.... ,SL O~de B4C BCI3-CH4-Hz 1200-1400
TiN TiClr N 2-H2 900-1000
Wlckln
Molten Metal
Si3N 2 SiCI-NH
4 3
-Hz 1000-1400
BN BCI3-NH3-H2 1000-1400
(b) AIN AlCI3-NH3-H2 800-1200
AIP3 AlCI3-C°2-H2 900-1100
Si02 SiH-C°2-H2 200-600
Fig. 14.7 Mechanism of directed metal oxidation Ti02 TiCI3-Hp 800-1000
growth14 • (a) no growth due to ~table ~rain bound- Zr02 ZrCI4-C°2-H2 900-1200
aryi (b) oxide growth mecharusm with unstable TiB2 TiCI4-BCI3-Hz 800-1000
grain boundary. WB WC16-BBr3-Hz 1400-1600
 Design considerations 317

The main drawbacks for processing com- gases are forced through the fiber preform
posites by CVI are the high processing time retained in a graphite holder with a sharp
and costs. Since the deposition occurs most thermal gradient maintained by water cooling.
rapidly on the outer surfaces, the internal pas- A schematic diagram of the forced flow /ther-
sages can be blocked off long before full mal gradient method is shown in Fig. 14.8.
densification is complete. It is usually neces-
sary to interrupt the infiltration process to
14.4 DESIGN CONSIDERATIONS
grind the surfaces in order to reopen the gas
access to the fibers or preform in the center of The approach to designing ceramic matrix is
the part. Residual porosity of 10-20% with less constrained by the brittle nature for both the
than 10% open porosity are typically matrix materials and reinforcements used.
obtained21 • Unlike polymer matrix composites and even
Two basic methods of CVI are isothermal metal matrix composites, the rationale for
processing and forced flow / thermal gradient design of ceramic composites is to impart
processing. In isothermal processing the fiber toughness in a structure that would have unac-
preform is heated by radiation from the walls ceptable toughness as a monolithic ceramic22 •
of the furnace (so called 'hot wall reactor') or Design methods are unique to the form of the
by inductively heating a carbon mandrel on composite, depending on whether continuous
which the preform is placed. In both cases the unidirectional reinforcements, discontinuous
decomposing gases are allowed to diffuse reinforcements or multi-layer, multi-directional
through the fiber preform. In the forced reinforcements are being considered. As a start-
flow / thermal gradient method the reactant ing point in the design of continuous

Hot zone

Heating element - - -..

Perforated lid - - - -

Graphite holder----..... ---

- ----
""'-- ...............
..--........,..----~

- -Cold surface
~tt--+---
Infiltrated preform
Fibrous preform

Reactant gases

Fig. 14.8 Forced flow thennal gradient method for CVI processing20.
318 Ceramic composites

unidirectional ceramic composites, the rule-of- strength over the range of reinforcement vol-
mixtures can be used to calculate elastic and ume fractions will depend on the relative
thermoelastic properties of the composite. fracture strain, strength and stiffness of the
Strength properties of the ceramic matrix com- constituents. The relative fiber and matrix
posites are poorly predicted by the stress-strain curves and strength prediction of
rule-of-mixtures since flaw sensitivity and rein- a composite consisting of a high stiffness, high
forcement-matrix bond strength are not strength fiber in a lower stiffness, low strain to
addressed by these techniques. Rule-of-mix- failure matrix is represented in Figs. 14.9(a)
tures properties are less important in and 14.9(b), respectively. There are many such
discontinuously reinforced ceramic composites fiber-matrix combinations that have this rela-
since toughness is strongly controlled by the tive behavior as the examination of Tables 14.2
interfacial properties. and 14.5 will reveal. For instance if Nicalon is
selected as the fiber then the selection of mul-
lite, lithium alumino silicate (LAS) or Pyrex
14.4.1 DESIGN OF CONTINUOUS
glass, for the matrix meets the requirement.
UNIDIRECTIONAL REINFORCEMENT
Selecting SiC monofilament produces this case
COMPOSITES
for almost all matrix materials listed in Table
The Young's modulus of unidirectional contin- 14.5 with the exception of TiB2 and TiC. The
uous fiber ceramic composites Ec is composite strength in such a system should
satisfactorily predicted by rule-of-mixtures: increase at a rate predicted by the linear rule-
of-mixtures based on the strength of the
matrix and the stress on the fiber at the frac-
ture strain of the matrix, a/. At fiber fractions
where Ef and Em are the Young's moduli of the to V crit failure of the matrix constitutes failure
reinforcement and matrix respectively and VI of the composite. The behavior of the compos-
and Vm are the respective volume fractions. ites with fiber fraction below Vcrit have simple
When there is a high bond strength between linear stress-strain behavior to failure. Above
the fiber and matrix, prediction of composite this fiber fraction the matrix breaks before the

i
-------a--
!
I
<1mu I
-----+----------
I
I
I
I
I
I

Strain
o Verit 1
Fiber Fraction
(a) (b)

Fig. 14.9 Strength prediction for high stiffness, high strength fiber and a lower stiffness, low strain to fail-
ure matrix.
 Design considerations 319

fibers. The fibers can retain the broken matrix X', is between the range given by
in place before the fibers break at a higher
VmrO'" VmrO'" I
load. Composite strength above Vcri! depends --<x<--
2Vf't - Vf't
upon the fiber strength. A typical stress-strain
curve for such a system above Vcri! is shown where 't is the interfacial shear stress and r is
the fiber radius. The value of strain at the end
of this process, E.max' is23

E.mu(1 + ~~~~) ~ Emax ~ Emu(1 + ~~~)

The final linear rising portion is the curve is
the elastic response of the fiber. Continuous
+ ! - - - matrix cracking fiber breakage and fiber pull-out can produce
the pseudo-ductility of the final portion of the
curve.
There are many potential continuous
fiber-matrix combinations in which the matrix
has a higher elastic modulus than the fiber. For
Strain ) instance, matrix materials such as titanium
diboride, titanium carbide, silicon carbide and
Fig. 14.10 Stress-strain behavior for composite alumina with most of the continuous fibers
with high stiffness, high strength fiber and a lower listed in Table 14.2 would have the relative
stiffness, low strain to failure matrix. stress-strain behavior shown in Fig. 14.11(a).
The strength of predictions of such systems is
schematically in Fig. 14.10. The elastic portion shown schematically in Fig. 14.11 (b). In this
of the curve is followed by a serrated, constant case the predicted strength of the composite
stress portion induced by a matrix failure would decrease with increasing fiber fraction
process. During this process the matrix contin- until a minimum fiber fraction, Vmin is reached.
ues to crack until the spacing between cracks, This behavior is similar in appearance to the

------------- --------- ---------------------- a w

iiaI

IO'mu
---------, - -..................J
I
__ ~----------- a~
I
I
I
I

Vmln
Strain
(a) (b) Fiber Fraction

Fig. 14.11 Strength prediction for a high strength fiber and a higher stiffness, low strain to failure matrix.
320 Ceramic composites

case of a high modulus, low failure strain fiber transverse elastic moduli EL and ET, respec-
in a lower modulus, high failure strain matrix tively, composites with aligned short fibers can
as is typical of many metal matrix composites. be made by using the Halpin-Tsai relation25:
However the cause of the minimum behavior
is quite different. Below V nUn failure of the EL = (1 + ~L 1JL Vf) / (1 -1JL Vf)
matrix still constitutes composite failure
where the rule-of-mixtures strength is com- ET = (1 + ~T 1JT V f) / (l-1J T Vf)
posed of the matrix ultimate strength, a mu and
where
the stress on the fiber at the matrix failure
strain, a/. Unlike the case for the lower modu- 'fIL = (EJE m -1) / (EJE m + ~L)'
lus matrix, the stress a/ is lower than the amu
increasing fiber fraction lowers the rule-of- 1JT = (EJE m - 1) / (EJE m + ~T)'
mixtures strength. Above V nUn the fracture of
the matrix no longer constitutes composite ~L =2l/d,and~T=2
fracture as the fibers alone are able to carry the
The coefficients of thermal expansion in the
load after matrix failure.
longitudinal and transverse directions, a L and
The above description applies to compos-
aT respectively, can be estimated from26
ites with a high fiber-matrix bond strength
and neglects the effect of fracture surface a L = (alfVf + amEmV m) / EL
energy. These conditions are not typical in real
composites and the simple rule-of-mixtures aT = (1 + Vf) afVf + (1 + Vrn) am Vm -aLvLT
predictions must be modified to account for
where
these effects. Aveston et al. 24 accounted for the
effect of the fiber-matrix bond strength on the
matrix failure strain as follows:
and vf and vm are the Poisson's ratio for fiber
, [ 6'trE f V/ ]1/3
£ f = E2 V rE and matrix respectively. These calculations will
m m c usually overestimate the value of these proper-
where r is the fracture specific fracture energy ties because of ineffective bonding between
of the matrix. fiber and matrix and deviation from ideal fiber

14.4.2 DESIGN OF DISCONTINUOUS

REINFORCEMENT COMPOSITES

In polymer and metal matrix composites it is

usually desirable to design the fiber lengths to
exceed the critical length, Ic' given by al / r to
allow the fiber to carry its full load prior to
fracture. In ceramic composites, fiber breakage
is rarely the design goal. Instead the role of the
fiber is to provide toughness by a combinations
of fiber pull-out, crack deflection and crack
bridging. Nevertheless, the designer may want
to predict the elastic and thermoelastic proper-
ties of the discontinuous reinforcement
composites. An estimate for longitudinal and Fig. 14.12 Microstructure of 20 vol. % SiC-AlP3
composite fabricated by tape casting.
 Design considerations 321

alignment. Figure 14.12 shows the in-plane laminate fracture theories must be employed.
microstructure for a 20 vol% SiC whisker-alu- Examples of material designs that can make
mina composite fabricated by tape casting and use of laminated-composite concepts for
hot pressing. This processing method promotes improved performance are illustrated in Fig.
fiber alignment in the tape cast direction, how- 14.1332 . The magnitude of the surface compres-
ever there is still a considerable deviation from sive stress can be calculated from laminate
the predominant fiber direction27 • theory. Figure 14.13(a) shows a laminate
For randomly oriented fibers or whiskers the design intended to produce surface compres-
elastic modulus, Er' of the composites can be sive stresses. In this design the layers toward
estimated from the results of the Halpin-Tsai the mid-plane gradually increase in coefficient
method using the empirical relation: of thermal expansion. The outer layers, con-
Er = (3/8)EL + (5/ 8)Er

14.4.3 DESIGN OF MULTILAYER, MULTI-

DIRECTIONAL REINFORCEMENT LAMINATES

The concept of a laminated composite is used

effectively in the design of polymer-matrix
composites to achieve the high degree of
strength, elastic and thermoelastic tailoring.
Polymer-matrix composites reinforced with (a) Design with graded composition
either continuous or discontinuous fibers are
fabricated by stacking layers with specific char-
acteristics and orientations in a predetermined
sequence to achieve desired mechanical or
physical properties. As with polymer compos-
ites, the ceramic composite layer properties
may be calculated using theoretical and semi-
empirical methods 28-29 from the constituent
properties such as the elastic modulus of the
fibers and matrix respectively, Ef and Em' the (b) Design with toughening layers
orientation factor, f, the volume fraction of the Oxidation
fiber, VI' the fiber aspect ratio, lid, and the coef- resistant layer
ficients of thermal expansion for the fibers and Wear
matrix, a f and am' By selection of the sequence resistant layer
of layer orientations and compositions, various High
toughness
elastic, thermoelastic, strength, physical and core
chemical characteristics can be produced.
Classical laminate plate theory3°-31 can be used
to accurately predict the elastic and thermoelas-
(c) High temperature wear design
tic properties of laminated composites from the
layer properties. The strength properties, on the
other hand, cannot be readily determined by Fig. 14.13 Typical laminate design concepts for
commonly used laminate failure criteria since ceramic matrix composites. (a) with graded compo-
fracture of these laminates is still strongly con- sition; (b) with toughening layers; (c) for high
trolled by the presence of flaws. Modified temperature wear.
322 Ceramic composites

taining increasing amounts of low-expansion stresses that are likely in such a construction.
whiskers generate compressive residual Differences in elastic modulus and coeffi-
stresses as a result of the differential contrac- cients of thermal expansion for layers
tion during cooling after the high-temperature containing different volume fractions of rein-
densification process. A major advantage of forcing whiskers can be used to generate
laminated-composite processing is that it pro- favorable residual stress patterns in fabricated
vides the engineering flexibility to use laminates. The thermal stresses 0' T, 0' Tand 't T
innumerable material and property combina- in each layer of the laminate at xany positidn
tions that would be impossible with through the thickness, z, measured from the
traditional methods involving thermal or midplane, caused by the restraint of the neigh-
chemical tempering. This concept also allows boring layer can be determined by Hooke's
the use of non-equilibrium compositions for Law,
greater degree of stress profile variation. For
instance, the depth and magnitude of the
stress gradient can be independently con-
trolled by selection of layer composition and
properties. Maximizing the stress gradient by
the introduction of a high-expansion material
where E~ E~ and yO are the mid-layer strains in
in the interior of the composite would be
the arbitrary di~tions of interest and Kx, K_y
impossible by conventional chemical temper-
ing but is quite feasible by lamination.
and Kxy are the curvatures for th~ laminate. gj
are the transformed reduced stiffness coeffi-
Strengthening can also be achieved by ren-
cients for each layer in the laminate.
dering surface flaws ineffective through the
A symmetric cross-ply laminate is the most
introduction of a tougher ceramic layer below
practical design for many applications. In this
the surface (Fig. 14.13(b». This design miti-
case, the constitutive equation for midplane
gates surface damage in the outer layers by
laminate strains reduces to
blunting the cracks when they reach the under-
lying toughened layer. This layer may contain
whiskers, a toughened ceramic, or metallic
particles. The use of a toughened ceramic layer
as the outer layer would not be as effective
since abrasion or impact could produce flaws
through its entire depth, thus permitting the where A/ are the in-plane extension flexibility
crack to propagate through the lower-tough- coeffici~ts and N}, N/ and NxyT are thermal
ness interior layers with minimum resistance. forces generated in the laminate as a result of
In addition to increased strength and tough- cooling from the temperature where stress
ness, high-temperature corrosion resistance can relieving is insignificant, I1T. For the case of
be designed into a composite material by using ceramic composites this may be very close to
a corrosion-resistant layer on the exterior sur- the hot pressing temperature. For a design in
face (Fig. 14.13(c» and layers tailored for which all of the reinforcements are aligned in
high-temperature strength in the interior. A one direction the thermal forces are
similar concept may be employed for a material n

designed as a high-temperature heat exchanger N; =k=l

~ (QIPI + Q12a2)t~T
by grading the interior layers for high thermal
conductivity. Using composite laminate theory, n

a materials designer can tailor the grading to ~ (QIPl + Q22a 2)tkI1T

Wy = k=l
minimize the deleterious residual tensile
 Properties 323

w=o
xy

where tk is the thickness of the kth layer, Qij are

the untransformed stiffness coefficients and a i
are the coefficients of the thermal expansion in
the principal material directions. The thermal
moments M , M and M are zero. The resid-
ual stresses, "aL a~d aT' inXYthe longitudinal and
transverse directions respectively are
(a) Bridging
aL = Qll (A'll N! +A'12 ~)
+ Q12 (A'12 N! + A'22~)

or = Q12 (A'll N! +A'12 ~)

+ Q22 (A'12 N! + A'22~)

For this laminate geometry the residual shear

stress, r LT =O.
(b) Pull-out
The compressive residual stresses thus
induced in the outer surface of the ceramic
composite raise the fracture strength by that
amounfl3.

14.5 PROPERTIES
The principal objective in design of ceramic
composites is to produce enhanced toughness
and mechanical reliability. Various energy (c) Deflection
absorbing mechanisms are produced by the
reinforcement depending on the relative ther- " \ , /
",l- ""\-I}..'; 1: ~,
mal expansion coefficients, relative elastic
moduli and interfacial bond strength between
the reinforcement and matrix. In addition the
size, shape, distribution and volume fraction
of the reinforcement plays a strong role in con-
=1: -{ ~ :-'Y "{ I
trolling the effectiveness of the toughening. -< Y, k"'t:( 1--<.,..)...;\
).. 1'< ~ H I
14.5.1 MECHANISMS OF STRENGTHENING 1\/
The four principal mechanisms of toughening (d) Microcracking
(crack bridging34, 35, fiber pull-out, crack
deflection and matrix microcracking) are
shown schematically in Fig. 14.14. More than
one of these mechanisms can be operative at Fig. 14.14 Toughening mechanisms for ceramic
the same time in a ceramic composite but there matrix composites.
324 Ceramic composites

is usually a dominant one depending on the resistance. A quantitative treatment of the

constituent and interfacial properties. effects of crack deviation on toughness have
Bridging, pull-out and deflection are most been provides by Faber and Evans37 .
effective where the fibers are generally aligned In certain composites the conditions can be
normal to the crack surface. favorable to allow the stress field of the prop-
In the crack bridging mechanism (Fig. agating crack to interact with the stress field
14.14(a)) the fibers remain intact for some dis- around the reinforcements to produce local
tance behind the crack front, thus restraining matrix cracking around the reinforcement.
the crack opening displacement and reducing Maximum effectiveness of this mechanism
the stress intensity at the crack tip. The energy requires a fine dispersion of many reinforce-
absorbing processes include fiber fracture, ments as illustrated in Fig. 14.14(d). An
fiber-matrix friction and elastic strain energy appropriate mismatch in thermal expansion
of the fiber. Thus, strong fibers, high strain to coefficients between reinforcement and
fracture fibers and strong fiber-matrix bond- matrix provides the local stress field around
ing promote this mechanism. The toughening the reinforcement. An analysis of this mecha-
produced by a uniform closure stress of the nism of toughening has been provided by
bridging fibers has been estimated by Becher Hutchinson38 .
et al. 35 to be Typical fracture in whisker reinforced
ceramic (SiC/ AIP3) shown in Fig. 14.15 has
dKc = 0' [V rE G /6(1 - v 2 ) EIG.]1/2
fulcm I
elements several mechanisms of toughening.
Whisker pull-out and crack defection are evi-
where Gm and Gi , are the strain energy release dent in this example.
rate for the matrix and interface respectively.
If the fiber-matrix interfacial bonding is
weak the fibers will be pulled out of the matrix
as the crack propagated through the compos-
ites, as illustrated in Fig. 14.14(b). The energy
absorbing mechanism in this case is the friction
between the fiber and the matrix. For fibers of
critical length the increment in toughness is
given as (see Reference 36, for example):

Crack deflection mechanism (Fig. 14.14(c))

forces the crack to deviate out of the normal
stress plane as it negotiates around the rein-
Fig. 14.15 SEM photograph of fracture path in SiC
forcements. The driving force for this deviation whisker-A120 3 composite.
is the residual stress distribution produced by
the mismatch in thermal expansion between
14.5.2 TYPICAL PROPERTIES
the fiber and matrix. Reinforcements with
higher coefficients of thermal expansion than The selection of materials for component
the matrix will cause the matrix to be in com- design requires accurate and reliable mechani-
pression near the reinforcement. This state will cal property data. Because of the large
tend to deflect a crack as it approaches the variations in processing characteristics and
vicinity of the reinforcement which is a higher starting material forms, such data are generally
crack resistant area to a region of lower crack sparse. The mechanical properties most signifi-
 Properties 325

cantly affected by reinforcements in ceramics 9.-----------------.

are fracture strength and fracture toughness.
8
Vaughn et al. 39 examined the effect of processing
temperature and reinforcement for alumina 7
and SiC whisker/alumina composites. Table
14.8 shows that processing temperature and
reinforcement have little effect on elastic modu-
lus, but a significant impact on strength,
fracture toughness and work of fracture. The
noncompliance with the rule of mixtures for
Young's modulus is evidence of the lack of
bonding and hence load transfer between fiber
and matrix. The toughening of various ceramic
matrix materials with increasing SiC whisker
content is shown in Fig. 14.1635 •
A compilation of fracture strength and frac- 0.1 0.2 0.3 0.4
ture toughness for SiC whisker (Silar-SC9) Whisker content (Volume fraction)
/ AlP3 composites is given in Table 14.92 • The Fig. 14.16 Increment in fracture toughness of SiC
fracture strength of the composites for the whisker composites with various matrices35•

Table 14.8 Mechanical properties of polycrystalline AI:P3 and SiC whisker/ AI:P3 matrix composites39

Alumina Alumina Alumina Composite Composite

(1500°C) (1659°C) (1900°C) (Silar SC-1) (Tateho SCW-1-S)
Young's 371 380 375 375 393
modulus (GPa)
Fracture 456±40 385±18 253±8 641±34 606±146
strength (MPa)
Fracture 3.3±O.2 5.0±O.2 3.7±O.1 8.7±O.2 4.6±O.2
toughness(MPa m 1/2 )
Work of 10 20 39 67 21
fracture Gm-2)

Table 14.9 Room temperature strength and fracture toughness of SiC whisker (Silar-SC-9)/ AIz03
composites (Adapted from Reference 2)

Whisker content Fracture strength Fracture toughness Reference

(Vol. %) (MPa) (MPam lll)
0 150 4.3 40
253 3.7 39
5 391 3.6 41
475 4.0 42
10 540 4.8 42
15 652 4.6 41
20 675 6.1 42
30 680 8.7 40
641 8.7 39
720 7.0 42
40 640 7.9 42
850 6.2 43
326 Ceramic composites

same whisker content can vary fairly signifi-

cantly among the different sources while the SiCw - mull ite
fracture toughness is relatively consistent. ~7
Above 30% whiskers some composites can o
Q.
exhibit either a strength or fracture toughness ::l6
that actually decreases because of the decrease fit
III
in homogeneity of the whisker distribution. G>
.E 5
The fracture strength and fracture toughness co
::I
o
for SiC whisker reinforced mullite are shown I- 4
in Figs. 14.17 and 14.18 respectively«. These ~

composites were fabricated by tape casting -=u

~ 3
and the L-type designation indicates that the IL

crack propagation direction is normal to the

tape casting direction and the T-type designa-
tion indicates that the crack propagation SiC Whisker Fraction (vol %)
direction is parallel to the tape casting direc-
tion. Greater strength and toughness are Fig. 14.18 Fracture toughness of SiC whisker-
mullite composites with the crack propagating nor-
achieved for the case of fiber orientation pre- mal (L-type) and parallel (T-type) to the
dominantly normal to the whisker axis predominant whisker direction44 •
direction. The maximum strength and tough-
ness for this system is at 40% whiskers. Table 14.10 Room temperature mechanical
The effect of SiC (Silar 5-9) whisker content properties of SiC whiskers/Si3N4 matrix
on strength and toughness of Si3N4 matrix composites45
composites is given in Table 14.1()45. For this
material the maximum toughness occurs at Whisker Fracture Fracture
30% whiskers while the strength maximum is content strength toughness
(Vol. %) (MPa) (MPa ml(2)
0 375 4.0
10 395 4.9
20 550 7.0
30 970 6.4
~ 800
:E
apparently above this fiber content. An alterna-
01
c: tive form of SiC whiskers was used to fabricate
~ 600 S4N4 matrix composites by Shalek et al. 46 • This
t;
whisker is formed by the Vapor-Liquid-Solid
process47,48. The effect of whisker content and
processing temperature on the elastic modulus,
fracture strength and fracture toughness are
shown in Figs. 14.19, 14.20 and 14.21 respec-
tively. There is an apparent critical processing
SiC Whisker Fraction (vol %) temperature above 1600 C to achieve maxi-
D

mum attainable strength and toughness

Fig. 14.17 Fracture strength of SiC whisker-mullite
properties in these composites.
composites with the crack propagating normal (L- Continuous fiber ceramic matrix compos-
type) and parallel (T-type) to the predominant ites are more likely to obey the rule of mixtures
whisker direction44 • prediction for elastic and strength properties
 Properties 327

especially where minimum fiber damage and

SiC Whisker-hot-pressed porosity occurs. The predictable increase in
Sia N4 matrix composite flexural strength for a Nicalon fiber Iborosili-
cate glass composite is seen in Fig. 14.2249•

14r-------------------~--~
SiC Whisker-hot-pressed
13· Sia N4 matrix composite
•
12

.: Hot-pressed at 1750°C
• : Hot-pressed at 1850°C
• : Hot-pressed at 16OO°C

22110 II 10 III 20 25 30 " 40

Volume.% SiC whiskers

Fig. 14.19 Elastic modulus of SiC whisker-S~N4

composites processed at various pressing tempera-
• : Hot-pressed at 175O"C
tures46 • • : Hot-pressed at 185O"C
4 A: Hot-pressed at 1600"C

30 5 10 15 20 25 30 35 40
Volume % SiC whiskers
SiC Whisker-hot-pressed
Sia N4 matrix composite
Fig. 14.21 Fracture tougness of SiC whisker-Si3N 4
composites processed at various pressing tempera-
• : Hot-pressed at 1750°C tures46 •
• : Hot-pressed at 1850"C
.. : Hot-pressed at 1600°C

• •
•
3OO0~------------------------~

• • 2500
Rule of mixtures prediction

•
1 for SiC/Borosilicate class ~ ,,/
~ ... '"
• "i2oo0
...;'"
......

• ~
~"
:g,1500
c: //'
~ 1000

500
",.r
...
... "
O~""""
o~~--~----~~~--~--~~~
o • 10 ~ ~ " 30 " 40
Volume. % SiC whiskers .1.2.3.4 . .5.6.7.8.9
Fibre volume fraction
Fig. 14.20 Fracture strength of SiC whisker-Si3N4
composites processed at various pressing tempera- Fig. 14.22 Rule of mixtures strength for SiC fiber
tures46 • (Nicalon™)-borosilicate glass49 •
328 Ceramic composites

14.6 APPLICATIONS
Crater Wear
14.6.1 HIGH TEMPERATURE STRUCTURES
Depth of
The most advanced demonstration of ceramic Cut Notch

matrix composites for high temperature struc-

tures and components has been for SiC CVI
infiltrated carbon and SiC (NicalonTM) fiber.
Extensive application of these composites
have been made by Societe Europeenne de
Propulsion (SEP)50. A demonstration two-
stroke 100 cc engine consisting of SiC-SiC pis-
ton, cylinder and cylinder head has run for ten
hours at full load and a speed of 1500 rpm
without lubricant.
A list of fabricated and tested high temper-
ature ceramic composite structures is given in Nose Wear
Table 14.11. These components operated suc-
cessfully under actual or simulated service
Fig. 14.23 lYrical wear pattern in cutting tool
conditions. inserts.

14.6.2 TRIBOLOGICAL COMPONENTS

(CUTTING)
of the chip. The abrasion of the rake by the
The application of ceramic composites to cut- broad face of the chip produces a crater.
ting tool inserts has made a significant impact Material removal from the rake face can be by
on machining of difficult-to-cut metals. The dissolution, adhesion and chemical reaction.
typical wear pattern in a cutting tool insert is Low chemical reactivity, hot hardness and
shown in Fig. 14.23. Uniform wear due to the wear resistance allows ceramics to minimize
rubbing action of the metal work piece pro- all these forms of wear.
duces the flank and nose wear. Deeper wear Unreinforced ceramics, while capable of
patterns at the depth-of-cut and trailing edge high cutting speeds, suffer from unpre-
region of the insert are due to the sharp edges dictable life due to low impact load tolerance

Table 14.11 Typical high temperature applications of ceramic matrix composites

Application Composition Operating Environment Test condition Reference

temperature
(OC)

Combustion SiC-SiC 1200 Oxidizing 14h running 50

chamber
Turbine blade SiC-SiC 1200 Air-kerosene Thermal cycle 400-1200°C 50
Turbine wheel SiC-SiC 1150 Air-kerosene 1 h up to 55 000 rpm 50
Turbine wheel C-SiC 1150 Air-kerosene 70000 rpm 50
Leading edge C-SiC 1400 Oxidizing Several hours 51
Nozzle C-SiC 1550 Oxidizing Maximum thermal gradient 52
60°C/mm
Radiant Nextel 312-SiC 1150 Air-natural gas 18 month operation 52
burner tube
 Applications 329

or poor thermal shock properties. Reinforcing reinforced alumina and cemented tungsten
ceramics especially by whiskers is an effec- carbide as seen in Fig. 14.2555•
tive technique to defeat these limitations. The
best known example of this application is the
SiC whisker reinforced Al20 3 insert. The com-
mercially available composition contains 40~--~w---------------~10
about 30% whiskers and is designated WG-
300 by Greenleaf Corp., Saegertown, PA,
USA.
The range of machining parameters for the
SiC whisker alumina compared to other con-
ventional and advances cutting tool materials
is shown in Fig. 14.2454. Carbide cutting tools
::i 20
are limited to cutting speeds below 100 m/min <.i
while the ceramic cutting tools range from .!:
over 100 m/min to 450 m/min. The SiC ~ 15
whisker reinforced composite can be seen to E
~ 10
provide the largest range of machining para-
meters for the machining of nickel based ~
:!: 5
superalloys compared to other advanced cut-
ting materials such as Sialon and TiC
particulate reinforced alumina.
The SiC whisker reinforced alumina (WG-
300) also has significantly greater tool life and Fig. 14.25 Comparison of tool life and metal
allows much greater rates of metal removal removal rates between various cutting tool materi-
for Inconel 718 compared to TiC particulate als55•

0.25

I
Sialon
AIPa-SiC(w)

0.20 C
a
:2 r
E b
g"C 0.15 i
CIl d
If e AIP3-TiC

0.10

0.05
0 100 200 300 400 500
Cutting speed (mlmin)

Fig. 14.24 Approximate range of machining parameters allowed by various cutting tool materials54 •
330 Ceramic composites

REFERENCES 15. Mecholsky; J.J., Engineering research needs of

advanced ceramics and ceramic-matrix com-
1. Karasek, KR, Bradley; S.A. Donner, J. T., Yeh, H.
posites, Am Ceram. Soc. Bull., 1989, 68(2),
C, Schienle, J.L. and Fang, H.T.,
367-375.
Characterization of silicon carbide whiskers, J.
16. Maloney, L.D., Make way for 'Engineered
Am. Ceram. Soc., 1989,72,1907-1913. Ceramics,' Design News, March 13, 1989, 45(5),
2. Homeny, J., Whisker reinforced ceramics, in:
64-71.
Ceramic Matrix Composites (ed R Warren), New 17. Naslain, R, CVI Composites, in: Ceramic Matrix
York, Chapman and ~all: 1992, p.245. . Composites (ed. R Warren), New York:
3. Romine, J.C, New high temperature ceramIC
Chapman and Hall, 1992, p. 199.
fiber, Ceram. Eng. Sci. Proc., 1987, 8, 755.
18. Buckley, J.D., Carbon-carbon: an overview, Am.
4. Holtz, AR and Grether, M.E, High temperature
Ceram. Soc. Bull., 1988, 67(2) 364-338.
properties of three Nextel ceramic fibers,
19. Besmann, T.M., Sheldon, B.W. and Kaster, M.D.,
Presented at 32nd International SAMPE
Temperature concentration dependence of .SiC
Symposium and Exhibition at Anaheim
deposition on Nicalon fibers, Surf. Coatzngs
Convention Center, April 6-9, 1987.
Technol., 1990,43/44,167-175.
5. Bunsell, AR, Development of fine ceramic
20. Stinton, D.P., Besmann, T.M. and Louden, RA,
fibers for high temperature composite, Materials
Advanced ceramics by chemical vapor deposi-
Forum, 1988, 11, 78. tion techniques, Am. Ceram. Soc. Bull., 1988, 67
6. Ishikawa, T., Recent developments of the SiC
(2), 350-355.
fiber Nicalon and its composites, including
21. Warren, R and Lundberg, R, Principles of
properties of the SiC fiber Hi-~icalon for ultra-
preparation of ceramic composites, in: Ceramic
high temperature, Compos. Scz. Tech., 1994, 51,
Matrix Composites (ed R Warren), New York:
135-144.
Chapman and Hall, 1992, p35.
7. Okamura, K, Ceramic fibers from polymer pre-
22. van Konijnenburg, J.T., Siskens, CAM. and
cursors, Composites, 1987, 18(2), 107-127.
Sinnema, S, Practical designing aspects of engi-
8. Flock, W.M., Characterization and processing
neering ceramics, in: Designing with Structural
interactions, in: Ceramic Processing Before Firing,
Ceramics, (ed RW. Davidge and H.M. Van de
(ed. George Y. Onoda, Jr. and Larry L. Hench)
Voorde), New York: Elsevier Science Publishing
New York: John Wiley and Sons, 1978, p.31.
Co, Inc., 1991, p.98.
9. Laynge, EF., Lam, D.CC, Sudre, 0., Flinn, B.D., 23. Pigott, M.R, Load Bearing Fibre Composites,
Folsom, c., Velamakanni, B.V., Zok, EW. and
Oxford: Pergamon Press, 1980, p.l06.
Evans, AG., Powder processing of ceramic
24. Aveston, J., Cooper, G. and Kelly, A, Single and
matrix composites, Mater. Sci. Eng., 1991, Al44,
multiple fracture, in: The Properties of Composites,
143-152.
Guildford: IPC Science and Technology Press,
10. Kragness, E.D., Processing and mechanical 1971, p.15.
behavior of tape cast and laminated silicon car-
25. Halpin, J.c. and Kardos, J.L. The. Halpin-Tsai
bide whisker/alumina composites, M.Sc.
equations: a review. Polym. Eng. Scz., 1976, 16(5),
Thesis, Pennsylvania State University; 1988.
344-352.
1l. Prewo, K.M., Brennan, J.J. and Layden, G.K,
26. McCullough, RL., Wu, CT., Seferis, J.C and
Fiber reinforced glasses and glass ceramics for
Lindenmeyer, P.H. Predicti~~s_. of _limiting
high performance applications, Am. Ceram. Soc.
mechanical performance for anisotropic crys-
Bull. 1986, 65(2), 305-322. talline polymers. Polym. Eng. Sci. 1976, 16(5),
12. Hillig, W.B., Melt infiltration approach to
371-387.
ceramic matrix composites, J. Am. Ceram. Soc.,
27. Schapery, RA Thermal expansion coeffici.en~
1988, 71(2), C96-C99.
of composite materials based on energy pnnCl-
13. Newkirk, M.S., Lesher, H.O., White, D.R,
pIes. J. Compos. Mater. 1968, 2(3), 280-404.
Kennedy; CR, Urquart, AW. and Claar, T.D.,
28. Kragness, E.D., Amateau, M.E and Messing,
Formation of Lanxide™ ceramic composite G.L., Processing and characterization of lami-
materials, Ceram. Eng. Sci. Proc., 1987,8,879.
nated SiC whisker reinforced Al203, J. Compos.
14. Newkirk, M.S., Urquart, AW., Zwicker, H.R
Mater., 1991,25(4),416-432.
and Brevel, E., Formation of LANXIDETM
29. Paul, 8., Predictions of elastic constants of mul-
ceramic composite material, J. Mater. Res., 1986,
tiphase materials, Trans. Met. Soc., AIME, pp.
1(1),81-88.
 References 331

36-41 (February 1960). Vol. 7, Composites, Impact, Statistics and High-

30. Pister, K.S. and Dong, S.B., Elastic bending of Temperature Phenomenon, (ed. KC. Bradt, D.P.H.
layered plates, J. Eng. Mech. Div., ASCE, 1-10 Hasselman, A.G. Evans and EE Lange), New
(October 1959). York: Plenum Press, 1986, pp. 61-72.
31. Reissner, E. and Stavsky, Y, Bending and 44. Wu, M., Messing, G.L. and Amateau, M.E,
stretching of certain types of heterogeneous Laminate processing and properties of oriented
aelotropic plates, J. Appl. Mech., 402-408 SiC whisker-reinforced composites, Ceramic
(September 1961). Transactions, VoL 19, 1991, Westerville, OH:
32. Amateau, M.P., Properties of laminated ceramic American Ceramic Society, 665-676, 1991.
composites, The 37th Sagamore Army Materials 45. Bulgan, S.T., Baldoni, J.G. and Huckabee, M.L.,
Research Conf Proc., (ed. D.J. Viechnicki), Si3Nr SiC whisker composites, Bull. Am. Ceram.
Watertown, Mass: Materials Technology Soc., 1987,66,347-352.
Laboratory, 317-338, October, 1990. 46. Shalek, P.O., Petrovic, J.J., Hurley, G.E and Gac,
33. Evans, AG. and Davidge, RW., A biaxial stress ED., Hot-Pressed SiC Whisker/Si3N 4 Matrix
method for the determination of the strengh of Composites, Am. Ceram. Soc., 1986, 65(2),
sections cut from glass containers and the size 351-356.
of critical Griffith flaws, Glass Tech. 1971, 12(6), 47. Milewski, J.V., Gac, ED., Petrovic, J.J. and
148-154. Skaggs, S.R, Growth of beta-silicon carbide
34. Evans, AG. and McMeeking, RM., On the whiskers by the VLS process, J. Mater. Sci., 1985,
toughening of ceramics by strong reinforce- 20,1160-1166.
ments, Acta Met., 1986, 34, 2435-2441. 48. Petrovic, J.J., Milewski, J.v., Rhor, D.L. and Gac,
35. Becher. P, Hsueh, e.-H., Angelini, P. and TIegs, ED., Tensile mechanical properties of SiC
T.N., Toughening behavior in whisker-rein- whiskers, J. Mater. Sci., 1985, 20, 1167-1177.
forced ceramic matrix composites, J. Am. Ceram. 49. Dawson, D.M., Preston, KE and Purser, A,
Soc., 1988, 71, 1050-1061. Fabrication and materials evaluation of high
36. Warren, R, Fundamental aspects of the proper- performance aligned ceramic fiber-reinforced,
ties of ceramic-matrix composites, in: Ceramic glass-matrix composite, Ceram. Eng. Soc. Proc.,
Matrix Composites (ed by R Warren), New York: 1987, 8, 815-821.
Chapman and Hall, 1992, p. 82. 50. Heraud, L.P. Spriet, 'High toughness C-SIC and
37. Faber, K.T and Evans, AG., Crack deflection SiC-SiC composites in heat engines', in Whisker-
process, Part I and 11., Acta Met., 1983, 31, and Fiber-Toughened Ceramics, Proceedings of an
565-584. International Conference, (ed. KA Bradley,
38. Hutchinson, J.w., Crack tip shielding by micro- D.E. Clark, D.e. Larsen and J.O. Stiegler),
cracking in brittle solids, Acta Met., 1987, 35, International Metals Park, OH: ASM, 1988.
1605-1619. 51. Melchior, AB., Pouliquen, M.E, Soler, E.,
39. Vaughn, W.L., Homeny, J. and Ferber, M.K., Thermostructural composite materials for liq-
Mechanical properties of silicon carbide uid propellant rocket engines, Paper
whisker / alumina oxide matrix composites, AIAA-87-2119, AIAA/SAE/SME/ ASEE, 23rd
Ceram. Eng. Sci. Proc., 1987, 8(7-8), 848-859. Joint Propulsion Conference, June 29-July 2,
40. Tiegs, T.N. and Becher, P.E, Sintering of 1987, San Diego, CA, American Institute of
Al z03-5iC whisker composites, Bull. Am. Ceram. Aeronautics and Astronautics, Washington, DC.
Soc., 1987, 66, 347-352. 52. Suacereau, D. and Beaurain, A, Demonstration
41. Porter, J.R, Langej, EE and Chokshi, AH., of carbon-silicon carbide Novoltex reinforced
Processing and creep performance of SiC composite nozzle on a LH2-LOx, Engine, Paper
whisker reinforced AlZ03' Bull. Am. Ceram. Soc., AIAA-90-2180, AIAA/SAE/SME/ ASEE, 26rd
1987, 66, 343--346. Joint Propulsion Conference, July 16-18, 1990,
42. Lio, S., Watanabe, M., Matsubara, M. and Orlando FL, American Institute of Aeronautics
Matsuo, Y., Mechanical properties of alu- and Astronautics, Washington, DC.
mina/ silicon carbide whisker composites, J. 53. Richards, RE., Bodkins, D. W. and Copes, S. J.,
Am. Ceram. Soc., 1989, 72, 1880-1884. Progress toward a cost effective thin wall RBT,
43. Becher, P.E, TIegs, TN., Ogle, J.e. and Warwick, in Energy Technology: Processings of The Energy
W. H., Toughening of ceramics by whisker rein- Technology Conference, Vol. 15, Energy
forcement, in: Fracture Mechanics of Ceramics, Technology Conference, February 17-19, 1988,
332 Ceramic composites

Washington, DC, Government Institute Inc. 55. Rhodes, J.F., Whisker reinforced ceramic com-
Washington DC, 749-756, 1988. posites, in Proc. Fifth Ann. Conf Materials
54. Billman, E.R., Mehrota, P.K., Shuster, A.F. and Technology, Materials Technology Center,
Beegley, C.W., Machining with A1203-SiC Southern Illinois University, Carbondale, IL,
whisker cutting tools, Bull. Am. Ceram. Soc, 205--219,1988.
1988,67,1016-1019.
CARBON-CARBON COMPOSITES 15
John D. Buckley

15.1 INTRODUCTION
Carbon--carbon (CC) materials are a generic
class of composites similar to the
graphite/ epoxy family of polymer matrix
composites. These materials can be made in a
wide variety of forms, from one-dimensional 1-0 2-D
to n-dimensional, using unidirectional tows,
tapes, or woven cloth (Fig. 15.1). Because of
their multiformity, their mechanical properties
can be readily tailored (Table IS.I). Carbon
materials have high strength and stiffness
potential as well as high thermal and chemical
stability in inert environments. These materi- 3-D n-O
als must, however, be protected with coatings
and/or surface sealants when used in an oxi- Fig. 15.1 Multiformity and general properties of
dizing environment. carbon-fiber and carbon-matrix composites.

The development of CC materials began in

Table 15.1 General properties of carbon-carbon 1958 and was nurtured under the US Air Force
composites space plane program, Dyna-Soar and NASA's
Apollo projects. It was not until the Space
Ultimate tensile >276MPa >40 000 psi Shuttle Program that CC material systems
strength were intensively researched. The criteria that
Modulus of >69 GPa >107 psi led to the selection of CC composites as a ther-
elasticity mal protection system were based on the
Melting point >4100°C 7412°F following requirements:
Thermal ",11.5 W m- K-l 6.64 BTU
1

conductivity hft OF 1. maintenance of reproducible strength levels

",1.1 X lO~oC 6.1 X 10-7 of
at 1650°C (3002°F);
Linear thermal
expansion 2. sufficient stiffness to resist flight loads and
large thermal gradients;
Density <2990 kgm-3 186.6Ib/fi3
3. low coefficient of thermal expansion to min-
imize induced thermal stresses;
Handbook of Composites. Edited by S.T. Peters. Published 4. oxidation resistance sufficient to limit
in 1998 by Chapman & Hall, London. ISBN 0412540207 strength reduction;
334 Carbon-carbon composites

5. tolerance to impact damage; plane or basal plane (a-b direction) have a

6. manufacturing processes within the state of covalent bond strength of ",524 kJ / mol
the art. (Kanter, 1957), while the bonding energy
Carbon-carbon composites consist of a fibrous between basal planes (c direction) is ",7 kJ / mol
carbon substrate in a carbonaceous matrix. (Dienes, 1952). The result is a crystal that is
Although both constituents are the same ele- remarkable in its anisotropy, being almost
ment, this fact does not simplify composite isotropic within the basal plane but with c
behavior because the state of each constituent direction properties that differ by orders of
may range from carbon to graphite. magnitude. On a larger scale, carbon, in addi-
Crystallographic carbon, namely graphite, tion to its two well-defined allotropic forms
consists of tightly bonded, hexagonal (diamond and graphite), can take any number
arranged carbon layers that are held together of quasicrystalline forms ranging continu-
by weak van der Waals forces. The single crys- ously from turbostratic (amorphous, glassy
tal graphite structure is illustrated in Fig. 15.2 carbon) to a highly crystalline graphite (Fig.
(Bokros, 1969). The atoms within the layer 15.3) (Bokros, 1969).

c -----r
Interlayer
- d 002

Reference directions

Fig. 15.2 TIghtly bonded, hexagonally arranged carbon layers held together by weak van der Waals forces.

La ·1
d 002 ~ 3.440 A d 002 = 3.354 A
(a) Lc:S; 50 A (b) Lc ~ 300 A

Fig. 15.3 Comparison of (a) carbon turbostratic structure with (b) 3-D graphite lattice (Bokros, 1969).
 Carbon fibers 335

The anisotropy of the graphite single crystal three precursor materials: rayon, polyacryloni-
encompasses many structural forms of carbon. trile (PAN) and petroleum pitch. Fibers having
It ranges in the degree of preferred orientation a low modulus (27.6 GPa (4 x 106 psi» are
of the crystallites and influences porosity, formed using a rayon precursor material that
among other variables. A broad range of prop- may be chemically pretreated by a sequence of
erties is the result of this anisotropy. In CC heating steps. First, the fiber is heated to
composites, this range of properties can >400°C (752°F) to allow cellulose to pyrolyze
extend to both constituents. Coupled with a (decomposition or chemical change during
variety of processing techniques that can be thermal conversion of organic materials to car-
used in the fabrication of CC composites, great bon and graphite). Carbonization (continued
flexibility exists in the design of and the resul- heating of organic material to >1000°C (1832°F)
tant properties obtained from CC composites. to initiate ordering of the carbon structures
The wide range of properties of carbon produced by pyrolysis) is completed more
materials can be shown when comparing the rapidly at >1000°C (1832°F). Upon completion
tensile modulus of commercially manufac- of carbonization, the fiber is graphitized (con-
tured carbon fibers that range from 27.6 GPa tinued heating of carbonized organic materials
(4 x 106 psi) to 690 GPa (100 x 106 psi). In fabri- to the 200D-3000°C (3632-5432°F) range of
cation, the fibers can be used in either produce 100% graphite-ordered crystal struc-
continuous or discontinuous form. The direc- ture) by heating to >2000°C (3632°F); the fiber
tionality of the filaments can be varied ranging is now, for all practical purposes, 100% carbon.
from unidirectional lay-ups to multidirec- High-modulus carbon fibers from rayon pre-
tional weaves. Fiber volume fraction cursors are obtained by the additional process
constitutes another variable. The higher the of stretching the carbon fibers at the final heat
volume fraction of a specific high-strength treatment temperature. High-modulus (344
fiber in a matrix, the greater the strength of the GPa (50 x 106 psi», high-strength (2.07 GPa
composite. The matrix can be formed via two (300 x 103 psi» carbon fibers are typically
basic approaches: (1) through the carboniza- made from PAN or, in some cases, mesophase
tion of an organic solid or liquid, such as a pitch precursors. These fibers are processed
resin or pitch, or (2) through the chemical similarly in a three-stage operation (Fig. 15.4)
vapor deposition (CVD) of carbon from a (Diefendorf, 1987). The PAN fibers are initially
hydrocarbon. A range of carbon structures can stretched from 500-1300% and then stabilized
be obtained by either approach. Finally, heat (cross-linked) in an oxygen atmosphere at
treatment of the composite material at graphi- 200°C (392°F) to 280°C (536°F) under tension.
tization temperatures offers additional Carbonization of the fibers is conducted
variability to the properties that can be between 1000°C (1832°F) and 1600°C (2912°F).
obtained. Typically, there is an optimum Finally, graphitization is accomplished at
graphitization temperature at which the high- >2500°C (4532°F). Mesophase pitch fibers
est strength can be obtained for a given undergo the same processing procedure as
composite composition of fiber and matrix PAN fibers but do not require an expensive
(Edie et al., 1986; Stoller et al., 1974). stretching process during heat treatment to
maintain preferred alignment of crystallites
(Fig. 15.4) (Diefendorf, 1987). Control of fiber
15.2 CARBON FIBERS
shape has resulted in improved fiber strength
The properties of carbon fibers can vary over a (4.1 GPa, 600 000 psi) (Cogburn et at., 1987)
wide range depending on the organic precur- when produced from melt-spun, mesophase
sor and processing conditions used. At petroleum pitch (Fig. 15.5) (Cogburn et al.,
present, graphite fibers are produced from 1987). Round fibers using the same method
336 Carbon-carbon composites

Stretch
PAN Thermoset 'G h't'
p~~proce~carbonoze rap ooze

~
Petroleum pitch Melt
~:
Thermoset Carbonize
H f Graphitize

Spool Surface treatment

Fig. 15.4 Carbon fiber production using PAN and pitch processes (Diefendorfer, 1987).

Cartridge
temperature controller
and indicator Hydraulic piston

Cartridge housing
.YJ---- Heating collar
Melted-pitch precursor
Spinnerette

u---+-- Melt-spun carbon

Melt-pressure indicator filaments

Spinning t----+Wind-up bobbin

chamber

Fig. 15.5 Melt spinning apparatus used to produce noncircular carbon fibers (Cogburn et al., 1987).
 Carbon fibers 337

had a strength of 2.1 CPa (300 x 103 psi) (Edie 2. achieve a more nearly isotropic material;
et al., 1986). Of the shapes studied, the C-shape 3. increase the composite interlaminar tensile
and hollow fibers were found to be superior in strength;
strength to round solid and trilobal cross sec- 4. along with continuous filament substrates,
tions (Edie et al., 1986; Cogburn et al., 1987). obtain a stronger composite by providing
additional nucleation sites that serve to
reduce composite porosity.
15.2.1 CARBON FIBERS IN CARBON MATRIX
The most widely used starting materials are a
Addition of a matrix to carbon fiber, either
carbonized, rayon felt substrate with a
through the carbonization of an organic pre-
pyrolytic carbon matrix and short, chopped
cursor or by the deposition of pyrolytic
fibers in a pitch-based matrix. Felt is produced
carbon, is conducted at 800°C (1472°F) to
through the mechanical carding of viscous
1500°C (2732°F). Subsequent heat treatment of
rayon fibers to produce a continuous web of
the composite material may involve tempera-
fibers. The webs are folded one on top of
tures to 3000°C (5432°F).
another to produce a batt. The batts are then
cut, stacked and needled to produce the
15.2.2 DISCONTINUOUS FIBER COMPOSITES required felt. The rayon felt is subjected to a
controlled carbonization cycle in an inert
Fabrication of discontinuous fiber composites
atmosphere or vacuum; the maximum temper-
uses short carbon fibers combined with either
ature determines such factors as shrinkage,
a pyrolytic carbon or pyrolyzed organic
weight loss and chemical composition of the
matrix. This approach to CC composites gen-
felt. A maximum carbonization temperature of
erally does not have true fiber reinforcement
1200°C (2192°F) is a nominal standard; the
as an objective. Rather, discontinuous fiber
length of the carbonization cycle and rate of
substrates have been used to:
temperature rise are dictated by the thickness
1. increase fabrication capability of large-scale of the felt. Carbon content in the fibers is
structures; ",98%. Carbon-carbon composites have also

(a) (b)

Fig. 15.6 Models of fiber arrangements for four short-fiber fabrication techniques: (a) flocking lay-up, (b)
pulp molding, (c) isotropic casting, and (d) spray lay-up (Cook, Lambdin and Trent, 1970; Lambdin, Cook
and Marrow, 1969; Lambdin and Cook, 1971).
338 Carbon-carbon composites

been fabricated from short carbon fibers using oriented substrate, usually with no interlock-
isotropic casting, flocking lay-up, spray lay-up ing between layers. Woven fabrics are used to
and pulp-molding techniques (Fig. 15.6) form a two-dimensional laminate with no
(Cook, Lambdin and Trent, 1970; Lambdin, interlocking between layers. Helical filament
Cook and Marrow, 1969; Lambdin and Cook, winding, which is directional, results in con-
1971). The rationale for using these short fibers tinuous, adjacent layer interlocking.
is to reduce composite anisotropy (Lambdin, Multilayer locking is achieved through com-
Cook and Marrow, 1969). plex weaving patterns or yarn placement
resulting in 'multidirectional' substrates (Fig.
15.7).
15.3 CONTINUOUS FIBER COMPOSITES
Continuous filament substrates reflect the
15.4 CHEMICAL VAPOR DEPOSITION
properties of high-strength filaments or
achieve a high degree of preferred orientation The CVD of carbon from a hydrocarbon gas
on the macroscale of the matrix. The fabrica- within a substrate is a complex process.
tion complexity for continuous-filament Various techniques have been applied to infil-
substrates is determined by two parameters: trate various fiber substrates including
(1) the directionality of the filaments, and (2) isothermal thermal gradient (Pierson, 1968),
the amount of layer interlocking achieved in pressure gradient (Kotlensky and Pappis,
the substrate. Filament winding of unidirec- 1969) and pressure pulsation (Beatty and
tional tapes can be used to achieve a highly Kipplinger, 1970). The first two have been the

Fig. 15.7 Interlocking approaches of continuous filament substrates: (a) tape wrapped, shingle; (b) filament
wound, helix; and (c) multidimensional.
 Carbonized organic composites 339

most extensively used. The isothermal tech-

nique is illustrated in Pig. 15.8. The substrate is Jacket
radiantly heated by an inductively heated sus- Carbon
ceptor so that the gas and substrate are substrate
maintained at a uniform temperature. Induction
Infiltration is normally accomplished at coils

1l00°C (2012°P) and at reduced pressures Graphite

(6 kPa (50 torr)) with the flow rates primarily susceptor
determined by the substrate surface area. This
technique produces a crust on the outer sur-
faces of the substrate, thus requiring
machining and multiple infiltration cycles.
In the thermal gradient technique (Pig.
15.9), the part to be infiltrated is supported by
a mandrel that is inductively heated.
Therefore, the hottest portion of the substrate Hydrocarbon
is the inside surface, which is in direct contact gas

with the mandrel. The outer surface of the

low-density substrate is exposed to a cooler
environment and results in a temperature gra- Fig. 15.8 Isothermal chemical vapor deposition to
dient through the substrate thickness. Surface infiltrate fibrous carbon substrate.
crusting is eliminated because the deposition
rate is greater on the heated fibers near the
mandrel, whereas the cooler outer fibers
receive little or no deposit. Under proper infil-
tration conditions, the carbon is first deposited
on the inside surface and, in a continuous Carbon
substrate
process, progresses radially through the sub-
strate as the densified substrate itself becomes Induction
inductively heated. Infiltration is normally coils
accomplished at atmospheric pressure with a Jacket
mandrel heated to approximately 1l00°C
(2012°P) (Theis et al., 1970). Sleeve

15.5 CARBONIZED ORGANIC COMPOSITES Graphite

susceptor
Carbonized organic composites have fabrica-
tion procedures that are similar to those of
conventional fiber-reinforced, resin-laminat-
ing techniques. The starting material is usually
a prepregged fabric or yam (a fabric impreg-
nated with a matrix material in a tacky state). Hydrocarbon
These precursor materials are staged nomi- gas

nally at al?proximately 100°C (212°P) to

achieve the desired degree of tack and flow of
the resin. A laminate is then constructed and Fig. 15.9 Thermal gradient chemical vapor deposi-
cured under pressure. Curing temperatures tion.
340 Carbon-carbon composites

Cut, lay-up,
debulk & bag

Coating Sealing

Carbon-carbon TEOS
part impregnation

3 to 5 times

Fig. 15.10 Fabrication steps involved in manufacture of 2-D carbon-carbon part impregnated with
tetraethylorthosilicate (TEOS).

normally range from 125°C (257°F) to 175°C To summarize, a typical manufacturing cycle
(347°F) with curing pressures on the order of of a 2-D ee part is shown in Fig. 15.10. First, a
2.76 MPa (400 psi). The reinforced resin lami- woven graphite fabric that is preimpregnated
nate is then post-cured at 200°C (392°P) to with phenolic resin is laid up as a phenolic-
275°C (527°F). As pyrolysis is initiated, shrink- graphite laminate in a mold and is autoclave
ing occurs as the organic phase decomposes. cured. Once cured, the part is pyrolyzed to
Simultaneously the release of vapors from form a carbon matrix surrounding the graphite
pyrolysis expands the composite material. A fibers. The part is then densified by multiple
slow release of these volatile by-products is furfural alcohol reirnpregnations and pyrolyza-
required to minimize structural damage to the tions. The resulting CC part then is ready for
char. Finally, as higher temperatures are use in inert or vacuum environments. This
reached, thermal expansion of the carbon char process is very time consuming. A single pyrol-
itself occurs after pyrolysis is complete. After ysis may take >70 h in a low-temperature,
the initial carbonization, the material is then inert-atmosphere furnace.
subjected to a series of reimpregnation and Although ee materials can withstand tem-
carbonization cycles until the desired density peratures >3000 oe (5432°P) in a vacuum or in
or the maximum density is achieved. The an inert atmosphere, they oxidize and sublime
reimpregnation process is usually conducted when in an oxygen atmosphere at 600°C
under vacuum and pressure to aid in maxi- (1112°F). To allow for use of ec parts in an
mizing the pore filling. If graphitization is oxidizing atmosphere, they must be com-
desired, the high-temperature heat treatment pounded with materials that produce
may be used after each carbonization step or at oxidation-protective coatings through thermo-
the end of the reimpregnation and recar- chemical reaction with oxygen at >2000°C
bonization cycles. (3632°F) (Buckley, 1967) or they must be coated
 Manufacturing 341

and sealed to protect them (Strife and Sheehan,

1988). For applications such as the Space
Shuttle CC leading edges and nose caps, sur-
faces are converted to silicon carbide in a
high-temperature diffusion-coating process
(Fig. 15.10). Because of differences in thermal
expansion between the silicon carbide and the
CC part, the coating develops microcracks
when the part is cooled from the coating tem-
perature. To maintain oxidation protection on
space vehicles such as the Space Shuttle, cracks
are impregnated with tetraethylorthosilicate
(TEOS). The TEOS process leaves silica in all of
the microcracks, greatly enhancing the oxida-
tion protection of the CC substrate. Current

<mJE
improvements being developed for oxidation
protection of the CC Space Shuttle components
Nose cap (1)
are additions of low-temperature glass formers
that enhance the sealing capability of the exist-
Ree Seal strip
(1) LH, (1) RH and
ing coating-TEOS system. (3) Lower
Ree Expansion seal
(1) LH, (1) RH and
(3) Lower
15.6 MANUFACTURING
The fabrication process of the Space Shuttle Fig. 15.11 Leading edge structural subsystem
Orbiter nose cap and wing leading edge com- (Curry, Scott and Webster, 1979).
ponents (Fig. 15.11) (Curry, Scott and Webster,
1979) is a multi-step process typical of the
technology used to produce CC composites. furnace and submitting it to a 7-day cycle dur-
The process steps are illustrated in Figs. ing which it is taken to 260°C (500°F) very
15.12-15.16. slowly to avoid distortion and delamination
Initial material lay-up is similar to conven- (Fig. 15.12).
tional practices with fiberglass-reinforced The next step is initial pyrolysis as shown in
plastic parts. Square-weave graphite fabric Fig. 15.13 (Curry, Scott and Webster, 1979).
impregnated with phenolic resin is laid-up in Pyrolysis tooling composed of graphite
an epoxy I fiberglass mold cavity shaped to the restraining fixtures containing the part are
desired configuration (Fig. 15.12) (Curry, Scott loaded into a steel retort that is packed with
and Webster, 1979). Lay-up thickness for these calcined coke. The retort and its contents then
components varies from 19 plies in the exter- undergo a 70 h pyrolysis cycle at 8i5°C
nal skin and web areas to 38 plies in the (1500°F) converting the phenolic resin to a car-
attachment locations. Upon completion of lay- bon state. During pyrolyzation, the resin forms
up, the part is vacuum-bagged and cured in an a network of interconnected porosity for the
autoclave to 150°C (300°F) for 8 h (Fig. 15.12). escape of volatile matter. This stage is
The cured part is rough trimmed, X-rayed and extremely critical since, during controlled char-
ultrasonically inspected for irregularities fol- ring of the cured resin matrix, the parts are
lowing the cure cycle. Post-cure of the weak and delamination can easily occur if ade-
component involves placing the part in a quate escape paths and time are not ensured.
graphite restraint fixture loading it into the After this initial pyrolysis cycle, the carbon is
342 Carbon-carbon composites

~~------------------------------------------------------------------------

16~----+-------------------------------------------------------------~
UJ 1100

~ 540
485
~
0
430
370
320
!ll

1
260

~
200
1~
95 7\
40
-20 ~T\I
I I HRS1--

,,
I
I

OJJl~
,
g I

• MATERIAL ' • CUT. • BAG (MYLAR I • AUTOC LAVE • REMOVE BAG • LOAD IN • POST CURE
COLO : LAY ·UP • APPLY VACUUM CURE • ROUG" TR IM RESTRAI N-
STORAGE , CLOT" • C"ECK FOR • DRI LL HO LES ING
: 'LiES LEAKS • X·RAY APoIOIOR FIXTUR E
I • ULTRASONIC
, OEBULK
!

Fig. 15.12 Lay-up and cure cycle (Curry, Scott and Webster, 1979).

~--------------------- , 2250

'550
16~' ~:-------------------------­
~' I00,~~--------------==j-------~
1100 :

iffi $010
i

~ ~0~~~~~~~~~~~~~~~3
i
485

ffi :~ , 0
430 '
370. !
370 ' 320 :
o '"
I
.so '

l:::!
320'

'::
200 :

~:!~~~~~~~~~~~~~~
~ 200'
t!) 1~ '
j~ 1
w
~ 95 '
40 '
-20 '
,20 ,
~1HRS
10 HOURS :

8~~ ! • LOAO I~
VACUUM
CHAMBER
• AUTOCLAVE CURE • I'OST CURE

, • LOAOIN • PYROLYSIS • CLEAN :

RETORT • INSf'EC1 j • IMP.":G wITH
FURFURY L
: • P,ACKWITH
GRAPHITE ALCOHOL
CALC INED COKE IMPREGNATION eye LE ..
,, , ITHREE TIMESI
(31_ __
1 _ _ ...... _______ ___ _ _____ _ _____ _ ___ _______ _ -- - - - - - .... ,
~

:. P'VROLYSISCYCLE ..-...
,1__ ... __ ____ ___ _(FOUR (41 _____
_ _ _ _ ____ TIMESI
_ _ _ _ _____ _ ... ... j

Fig. 15.13 Initial pyrolysis (Curry, Scott and Fig. 15.14 Densification impregnation and cure
Webster, 1979)_ (Curry, Scott and Webster, 1979).
 Manufacturing 343

designated reinforced carbon-carbon-O

(RCC-D), a state in which the material is
extremely light and porous with a flexure
2250 +-1- - - - - - - - - - - - - - - '
I ,
w
C
1650
I
I ,
I

strength of 21-24 MPa (3000-3500 psi). ~ 1100

I,
I
(!) I

,
I
F 540
lD
()
485
430
I
I
15.6.1 DENSIFICATION w 370 DRV1NG,
w 320
a: 260
I

Densification for these shuttle parts is accom-

I

~ 200
150 , I

plished in three impregnation and pyrolysis 95

/
I

-
I
40
cycles (Fig. 15.14) (Curry, Scott and Webster, -20
I 16 HOURS l -
1979). Each part is loaded in a vacuum cham- I

ber impregnated with furfural alcohol

G5~
followed by a 2-hour cure period in the auto-
clave at approximately 150°C (300°F),
followed by a post-cure for 32 h to 200°C
(400°F). This cycle is followed by a 70-hour
815°C (1500°F) pyrolyzation that is shown in :_
:_
CLEAN • COATING
INSPECT PREPARATION
-
-
CLEAN
INSPECT
Fig. 15.13. After three impregnation/pyrolyza- :
:
_ X-RAV IN RETORT _ SURFACE
_ ULTRASONIC COATING
- X-RAV
- ULTRASONIC
tion cycles, the material is designated RCC-3 :_ FINAL TRIM/DRILL - DIMENSIONAL I:
:- DIM. INSPECT ,.
with an increased flexure strength of ",,124
MPa (18000 psi) at room temperature. Fig. 15.15 Coating cycle (Curry, Scott and Webster,
1979).
15.6.2 COATING

To allow for use of CC composites at elevated idation and sintering of the powders. During
temperatures above 2000°C (3632°F) in an oxi- the process, the outer layers of the CC sub-
dizing atmosphere, it is necessary to apply strate are converted to silicon carbide. The
protective coatings to structural components. silicon carbide-coated CC composite part is
The oxidation inhibition process consists of removed from the retort, cleaned and
two steps: (1) diffusion coating the CC compo- inspected. During cool down from 1650°C
nent and (2) applying a sealer to the surface. (3000°F), the silicon carbide coating contracts
The coating process (Fig. 15.15) (Curry, Scott slightly more than the carbon substrate, caus-
and Webster, 1979) used in protecting the CC ing crazing (coating fissures). This crazing
shuttle components starts with the blending of together with the inherent material porosity
the constituent powders: 10% alumina, 30% provides paths for oxygen to reach the carbon
silicon and 60% silicon carbide. This mix is substrate. To obtain increased useful life of this
packed around the CC structural component CC structural component, it is necessary to
in a graphite retort and loaded into a vacuum add an additional oxidation inhibitor. The
furnace where it undergoes a 16-hour cycle final process used to provide oxidation protec-
that includes drying at 315°C (600°F) and the tion to this type of CC structure involves
coating reaction to 1650°C (",,3000°F) in an impregnating (Fig. 15.16) (Curry, Scott and
argon atmosphere. The powder characteris- Webster, 1979) this component with TEOS. The
tics, constituents, formulations and the part is covered with a mesh, placed in a vac-
manner in which the powders are packed uum bag and the bag is filled with liquid
around the part are important factors that gov- TEOS. AS-cycle TEOS impregnation is then
ern the chemical reactions at the high performed with the bagged part. After the fifth
processing temperatures, the degree of consol- TEOS cycle, the part is removed from the bag
344 Carbon-carbon composites

~ 320
!2 260

-
1\
Z 200
J \
~ ,SO
w 95
l- I \
~ 40 1\ I \
:- I- I- 1-
~ ·20
6 .5 6 .5 2 .5 I HAS
TO TO TO TO TO HAS
110 110 10 10 2l\
MIN MIN MIN MIN HAS

~ ..'- ~~
FIVE CYCLES
• EVACUATE
AND
~ "
.

• COATED • WEIGHTI • COVER • INSTAl.l • CONNECT IlACKFILL. REMOVE • cu"e • HOE

IIECOAD WITlt -VACUUM VACUUM AND WITH TEOS IlAG _" HR. 400"F - X.AAY
".IIT
• VERI~Y MESH FlrrlNGS FILL LINES MIX • CLEAN _ IHA. 600"F - EOOY
'(11 S'EC
__ -SIN - FILL • MIX TEOS • OVEN CUIIE P.o.AT CURRENT
mO •
7~'

_ ,IN TUBU ILloolDI SOoF - ULTRASONIC

• COVER AND FILL "5~ MIN I • WEIGHI
WITH MESH RESERVOIR • FEED TEOS RECORD
• VACUUM MIX ASREO'D
IlAG TO COVER
• 5TH CYCLE CURE
\2 TO 2" HII MIN)
• COOL TO '00°.

Fig. 15.16 TEOS impregnation (Curry, Scott and Webster, 1979).

and oven cured at 315°C (600°F), liberating all labeled Space Shuttle material, is the strength
of the hydrocarbons. This procedure leaves sil- level of the reusable carbon-carbon (RCC)
ica (Si02) in all of the microcracks and fissures material used in the Space Shuttle thermal
greatly enhancing the oxidation protection of protection system. Even though this material
the CC structure. is made with low-strength carbon fibers, its
strength efficiency is superior to both superal-
loys and ceramics at >1000°C (1832°F).
15.7 MECHANICAL PROPERTIES
Development of advanced carbon-carbon
The extreme thermomechanical requirements (ACC) composites has produced a material
of the Space Shuttle have been the impetus for that is twice as strong as the CC composite
evaluating properties of low-density Cc. The first put on the Space Shuttle. The ACC mate-
use of CC on the nose cap and leading edges of rial is made using woven carbon cloth. When
the Space Shuttle makes it imperative to know unidirectional carbon fiber tapes are interplied
as much as possible about all the characteris- with woven cloth to create a hybrid ACC,
tics of this materiaL The effect of temperature strength in at least one direction can be
on the ratio of tensile strength to density for increased by >345 MPa (>50000 psi). Current
several classes of high-temperature materials data on thermomechanical and thermochemi-
is shown in Fig. 15.17. The major advantage of cal properties of some of the advanced CC
CC materials for high-temperature applica- systems show that material composition, oxi-
tions is that they do not lose strength as the dation resistance, processing, joining and fiber
use temperature is increased. This property is architecture are producing noticeable
in contrast to other materials such as superal- improvements in CC materials and structures
loys and ceramics. Figure 15.17 shows three (Curry, Scott and Webster, 1979; Buch, 1984;
levels of CC strength efficiency. The first, RummIer and Sawyer, 1984; Ransone and
 Thermal properties 345

Temperature, °C
30 0 ·550 1100 1650 22Plbo

25 160
'iii (J)~
.:;t!
"0::3
.s:: 20 CD (I)
Q.=:
g> .~~ 120 g. CD

e 15 10 ~
(D
'Iii 001
(1):;:: 80 ~ 'S.
:J
;:::!:
::'0 '< =r
CIl(1)
Co.
10
{!!.C/) s::::
-0
40 SIJ
5

0
0 800 1600 2400 3200 4000
Temperature, OF

Fig. 15.17 Strength-to-density ratio for several classes of high-temperature materials.

Ohlliorst, 1984; Webb, 1985; Gray and Engle, ulus values for the TEOS material are shown
1985; Johnson and Finley, 1985; Sawyer and in Fig. 15.19 (Curry, Scott and Webster, 1979).
Moses, 1985; Maahs and Ransone, 1985; The effect of temperature on the as-fabricated
Ohlliorst and Ransone, 1985). tension strength properties is shown in Fig.
CC components on the Space Shuttle are 15.20 (Curry, Scott and Webster, 1979). As
required to have adequate strength at design shown in Figs 15.17 and 15.20, the strength of
temperatures to withstand the aerodynamic CC composite material does not decrease sig-
loads of flight and to continue to do so for the nificantly with temperature. Typically, above
operational life of the component. Minimum 1425°C (2600°F) there is an increase in
mechanical properties are guaranteed through strength.
statistical analysis of a data sampling having The effect of substrate mass loss through
at least 99% probability and 95% confidence. oxidation on tensile strength is shown in Fig.
The primary variables affecting the structural 15.21 (Curry, Scott and Webster, 1979). Mass
design allowables are temperature, material loss results in a significant reduction in design
thickness, coating thickness, biaxial stress con- allowable stress, emphasizing the value of the
ditions and substrate mass loss due to additional oxidation protection provided by
oxidation through the mission life of the com- the TEOS treatment.
ponent (Table 15.2) (Curry, Scott and Webster,
1979).
15.8 THERMAL PROPERTIES
Figure 15.18 (Curry, Scott and Webster,
1979) illustrates the typical effect of ply thick-
15.8.1 THERMAL OXIDATION
ness on the allowable stress values for tension,
bending, compression and shear used for A critical requirement when using CC com-
design. As fabricated, room temperature mod- posites is the ability to withstand numerous
346 Carbon-carbon composites

Table 15.2

Mechanical properties test Non-TEOS TEOS

as fabricated conditioned as fabricated conditioned
Flexure 201 221 40 24
Tension 196 80 52 34
Compression 170 57 46 28
Shear 192 79 17 17
Comer flexure 20 3 3
InterIaminar tension 20 50 5 5
Interlaminar shear 30 19
Coefficient of thermal expansion 10 6 9 6
Impact, etc. 20 10
Simultaneous cycling 3 18
Mission cycling
Tension 3 36
Flexure 3 24
Compression 3 27
Total 868 609 175 135

104

.7
10

-t NOTE
BREAK
IN
SCALE
TENSION~

,.".----
IN.PLANE SHEAR ~

21
____ ______ ______ ______ ______ ______
I' 2D 2S 40
NUMBER OF PLIES

Fig. 15.18 Design allowables at room temperature as-fabricated (Curry, Scott and Webster, 1979).
 Thermal properties 347

TYPICAL SECANT MODULUS

21

~_ _ _ _ _ _ _ Eb

____________-e,

- - -_ _ _ _ _ _ _ _ _ _ 6 ••
7.0

3~~-~---~--~Z5---~~~--~~-~
NUMBER OF PLIES

Fig. 15.19 Design allowables at room temperature as-fabricated (Curry, Scott and Webster, 1979).

thermal and thermomechanical loads during flaws were performed over a wide range of
re-entries of the Space Shuttle into the earth's pressures and temperatures in both plasma arc
atmosphere. Although CC Space Shuttle com- jets and radiant-heating test facilities. AIc jet
ponents have an oxidation-inhibiting silicon tests on CC specimens ranged in temperature
carbide coating, they can lose mass over an from 81S0C (lS00°F) to 1870°C (3400°F) and
extended temperature range without apparent atmospheric pressures from 0.01 Pa to 0.10 Pa.
surface recession. Photomicrographs of CC Radiant-test conditions ranged from 420°C
specimen surfaces show minute fissures and (800°F) to 142SoC (2600°F) and pressures rang-
thermal microcracks, some of which terminate ing from 0.01 Pa to 1.0 Pa. Mass-loss data for
at the coating substrate interface. Specimens the CC shuttle specimens exposed to the arc jet
exposed to convective and radiant heat trans- and radiant-heating tests are presented in Figs
fer tests micrographically have shown the lS.22 and lS.23. Figure lS.22 (Curry, Scott and
presence of voids at the coating substrate sur- Webster, 1979) shows mass loss at 980°C
face. Tests to characterize the effects of these (1800°F) and O.OS Pa as a function of exposure

l
,
:IE 42

i
.... __ ' _----31 PlY

-----
~---- __ • _--------- 28 PLY
III ~------------
... _ _ ;::::::::::75PlY
_ _ _ _ _ _ _ _ _ _ -~-~ 19 Pl.
~3!
~

I "--
•~:lc
~
11128

R T.
C 21 ~~~~--~~~--~~----~----~~-~~--~~
- 157 - 18 ;110 538 8'8 11114 1372 18!1O
TEMPERATURE - ·c

Fig. 15.20 Design allowables as-fabricated (Curry, Scott and Webster, 1979).
348 Carbon-carbon composites

• 42
Go
:I

it;~:::==========::======3II'LV
I r-------------____________________
:z: 35

c a~
:::::::26PLV'LV 28

11~V
ti
Z
o 21
ii
i
~ 1....
~

i9 7.0
..I
.:
O~------~--------~--------L-------~--------~
o U U ~ ~ ~
MASS lOSS - kahn')( 10'

Fig. 15.21 Design allowables conditioned (Curry, Scott and Webster, 1979).

48.8 24.5
T 982·C P 0.05 x 10' Pa T 982·C P 0.05 x 10' Pa
2:1
5! 39.0 -; 19.6

E
x
...~
~ 29.3 11.0
en
~ CI)

9 9
! 19.8 m 9.8
01(
::IE
::IE

9.8 4.9

2 ~ 6 8 10 o l 4 6 8 10
EXPOSURE TIME HOURS £ XPQSURE TIME HOURS

Fig. 15.22 Mass loss comparison plasma arc jet Fig. 15.23 Mass loss comparison radiant environ-
environment (Curry, Scott and Webster, 1979). ment (Curry, Scott and Webster, 1979).

time in a plasma jet. Figure 15.23 (Curry, Scott 15.8.2 THERMOPHYSICAL PROPERTIES
and Webster, 1979) represents radiant-heating
The CC composite used on the Space Shuttle
data. A comparison of substrate mass loss with
is both inhomogeneous and anisotropic;
and without TEOS shows the significant pro-
therefore, thermal conductivity is dependent
tection afforded the CC substrate by TEOS
upon the direction of heat flow (i.e. perpen-
impregnation. dicular or parallel to the ply). In addition,
 Applications 349

O. 18 14.~
-- " ,./'"
0.8 11.8 ............T;;:L EMISSIVITY~
12

\
0.1 " 11.7
Ii
i
~ 8 I
~
~
O. " ~5.8

~ !... ~z -THERMAL CONDUCTIVITY PERPENDICULAR

i! :i:z: 4
0
IJ TO PLY
ilIo. ~
....
~2.9

•
~

~ ~.,
::II
II:
:z: I--- TOT AL EMISSIVITY

0 0
- 18 280 538 8'8 1084 '372 UI50
-296
TEMPERATURE ·C

Fig. 15.24 Reinforced carbon-carbon thermal properties (Curry, Scott and Webster, 1979).

thermal conductivity is dependent upon the CC material systems using coatings, TEOS
mass loss experienced by the CC composite, and additions to the basic CC recipe have
resulting from subsurface oxidation. improved the oxidation resistance of products
Results of thermal conductivity studies for made of CC composites by an order of magni-
shuttle CC composite shuttle materials are tude. These composites are being used in
shown in Fig. 15.24 (Curry, Scott and Webster, products such as the nozzle in the F-100 jet
1979). To simplify thermal modeling, no differ- engine afterburner, turbine wheels operating
entiation has been made for conductivity at >40000 rpm, nonwetting crucibles for
variation resulting from the number of plies in molten metals, nose caps and leading edges
the substrate. Results for conditioned speci- for missiles and for the Space Shuttle, wind-
mens having a mass loss of 0-5 Pa (O.llb/fe) tunnel models and racing car and commercial
suggest that thermal conductivity decreases disk brakes (Klein, 1986).
with mass loss. Figure 15.24 also shows that Pushing the state of the art in CC compos-
neither specific heat nor emittance was ites is the piston for internal combustion
affected by material or mass loss conditioning. engines (Miller and Grimes, 1982; Taylor,
1985). The CC piston (Fig. 15.25) would per-
form the same way as any piston in a
15.9 APPLICATIONS
reciprocating internal combustion engine
An example of the state of the art in CC com- while reducing weight and increasing the
posite applications is a one-piece, bladed mechanical and thermal efficiencies of the
turbine rotor that, in service, is coated to pre- engine. The CC piston concept features a low
vent oxidation. The rotor offers higher piston-to-cylinder wall clearance; this clear-
temperature performance without cooling; ance is so low, in fact, that piston rings and
low weight and use of low-cost, non-strategic skirts are unnecessary. These advantages are
materials (Miller and Grimes, 1982). Other gas made possible by the negligible coefficient of
turbine engine applications using CC compos- thermal expansion of this kind of CC (0.54 x
ites include exhaust nozzle flaps and seals, 1O-6 / o F). (Carbon-carbon composites can
augrnenters, combustors and acoustic panels. have a range of thermal expansion coefficients,
350 Carbon-carbon composites

REFERENCES
Beatty, RT. and Kipplinger, D.V., 1970, Gas pulse
impregnation of graphite with carbon. Nuclear
Application and Technology, 8(6):488-495.
Bokros, J.c., 1969, Deposition, Structure and
Properties of Pyrolytic Carbon. Chemistry and
Physics of Carbon-A Series of Advances, (ed.
Philip L. Walker, Jr.) pp. 1-118. Marcel Dekker,
Inc.
Buch, JD., 1984, Graphite Crystals - A General
Model for Diverse Carbon Forms. Metal Matrix,
Carbon and Ceramic Matrix Composites, (ed. John
10 II I 2 D. Buckley) NASA CP-2357, pp. 119-135.
INCHE S Buckley, J.D., 1967, Statis, Subsonic and Supersonic
Oxidation of IT Graphite Composites, NASA TLN
D-4231.
Fig. 15.25 Carbon-carbon automotive piston. Cogburn, J.w., Fain, c.c., Edie, D.o. and Leigh,
H.D., 1987, Processing C-Shape Pitch-Based
Carbon Fibers. Metal Matrix, Carbon and Ceramic
Matrix Composites, (ed. John D. Buckley) NASA
depending on the processing techniques.} CC CP-2482, pp. 185-200.
material maintains its strength at elevated Cook, J.L., F. Lambdin and P.E. Trent, 1970,
temperatures allowing the piston to operate at Discontinuous Carbon/Carbon Composite
Fabrication. Carbon Composite Technology - With
higher temperatures and pressures than those Special Emphasis on Carbon/Carbon Systems. Proc
of a comparable metal piston. The high emit- 10th Ann. Symp. New Mexico Section of ASME
tance and low thermal conductivity of the CC and University of New Mexico, pp. 143-171.
piston should improve the thermal efficiency Curry, D.M., Scott, H.C. and Webster, CN., 1979,
of the engine because less heat energy is lost to Material Characteristics of Space Shuttle
the piston and cooling system. The elimination Reinforced Carbon-Carbon. Paper read at the
of rings reduces friction, thus improving 24th National SAMPE Symposium, 1-9 May,
1979, at San Francisco, CA.
mechanical efficiency. Diefendorf, R.J., 1987, Carbon/Graphite Fibers.
Besides being lighter than conventional pis- Engineered Materials Handbook 1: 49-53.
tons, the CC piston can produce cascading Dienes, GJ., 1952, Mechanism for Self-Diffusion in
effects that could reduce the weight of other Graphite. Applied Physics 23(11): 1194-1200.
reciprocating components such as the crank- Edie, D.D., Fox, N.K., Barnett, B.C. and Fain, c.c.,
shaft, connecting rods, flywheels and 1986, Melt-Spun Non-Circular Carbon Fibers.
balances, thus improving specific engine per- Carbon 24(4): 477-482.
Gray, P.E. and Engle, G.B., 1985, Wettability of
formance (Taylor, 1985). Carbon/Carbon Composites and Carbon Fibers
by Glass Sealants Used in Oxidation Inhibition.
ACKNOWLEDGEMENTS Metal Matrix, Carbon and Ceramic Matrix
Composites, (ed. John D. Buckley) NASA
The author acknowledges Mr. D.M. Curry of CP-2406, pp. 149-162.
NASA Johnson Space Center and H.C Scott Johnson, A.c. and Finley, J.w., 1985,
and CN. Webster of the Vought Corporation Carbon/Carbon Composites for Advanced
Spacecraft. Metal Matrix, Carbon and Ceramic
for the data, as referenced, on which a portion
Matrix Composites, (ed. John D. Buckley) NASA
of the present paper is based. CP-2406, pp. 175-190.
Acknowledgement is also given to Mrs. H.A. Kanter, M.A., 1957, Diffusion of Carbon Atoms in
Coombs for her valuable contribution in Natural Graphite Crystals. Physics Review 107
assisting in the formatting of this paper. (3):655-663.
 References 351

Klein, J., 1986, Carbon-Carbon Composites. Ransone, P.O. and Maahs, H.G., 1985, Effect of
Advanced Materials and Processes 130 (5):64-68. Processing on Microstructure and Mechanical
Kotlensky, WV. and Pappis, J., 1969, Mechanical Properties of 3-D Carbon-Carbon. Metal Matrix,
Properties of CVD Infiltrated Composites. Proc. Carbon and Ceramic Matrix Composites, (ed. John
95th Biennial Conf.Carbon Defense Ceramic D. Buckley) NASA CP-2406, pp. 289-303.
Information Center, Compilers, pp. 76-80. Ransone, P.O. and Ohlhorst, C.W, 1984, Interlaminar
Lambdin, E and Cook, J.L., 1971, Fabrication of Shear and Out-of-Plane Tensile Properties of
Carbon-Carbon Composites Electrostatic Fiber Thin 3-D Carbon-Carbon. Metal Matrix, Carbon
Deposition (Flocking). Y-1786 (Contract No. and Ceramic Matrix Composites, (ed. John D.
W-7405-eng-26), Y-12 Plant, Union Carbide Buckley) NASA CP-2357, pp. 137-148.
Corp. Rummler, D.R and Sawyer, J.W, 1984, Properties
Lambdin, E, Cook, J.L. and Marrow, G.B., 1969, and Potential of Advanced Carbon-Carbon for
Fiber-Reinforced Graphite Composite Fabrication Space Structures. Metal Matrix, Carbon and
and Evaluation. Doc. Y-1684, TID-4500 (Contract Ceramic Matrix Composites, (ed. John D. Buckley)
W-7405-eng-26), Nuclear Division, Union NASACP-2357, pp.149-170.
Carbide Corp. Sawyer, J.W and Moses, P.L., 1985, Effect of Holes
Maahs, H.G. and Ransone, P.O., 1985, Mechanical and Impact Damage on Tensile Strength of Two-
Property Evaluation of 2-D Carbon-Carbon Dimensional Carbon-Carbon Composites.
Panels Fabricated From a Specialty-Weave Metal Matrix, Carbon and Ceramic Matrix
Fabric. Metal Matrix, Carbon and Ceramic Matrix Composites, (ed. John D. Buckley) NASA
Composites, (ed. John D. Buckley) NASA CP-2406, pp. 245-260.
CP-2406, pp. 261-276. Stoller, H.M., Butler, B.L., Theis, J.D. and Lieberman,
Miller, T.J. and Grimes, HH., 1982, Research on M.L., 1974, Carbon Fiber Reinforced-Carbon
Ultra-High-Temperature Materials-Monolithic Matrix Composites. Composites: A State of the Art,
Ceramics, Ceramic Matrix Composites and (eds. JW. Weeton and E. Scala) Metallurgical
Carbon/Carbon Composites. Advanced Society of the American Institute of Mining,
Materials Technology, (eds. Charles P. Metallurgical and Petroleum Engineers, Inc., pp.
Blankenship and Louis A. Teichman) NASA 69-136.
CP-2251, pp. 275-291. Strife, J.R and Sheehan, J.E., 1988, Ceramic Coatings
Ohlhorst, C.W and Ransone, P.O., 1985, Effects of for Carbon-Carbon Composites. Ceramic Bulletin
Thermal Cycling on Thermal Expansion and 67(2): 369-374.
Mechanical Properties of Advanced Taylor, A.H., 1985, Carbon-Carbon Pistons for
Carbon-Carbon Composites. Metal Matrix, Internal Combustion Engines. NASA Tech
Carbon and Ceramic Matrix Composites, (ed. John Briefs 9 (4):156-157.
D. Buckley) NASA CP-2406, pp. 289-303. Theis, J.D., Jr., Taylor, A.J., Rayner, RM. and Frye,
Pierson, H.O., 1968, Development and Properties of E.R., 1970, Filament Wound Carbon/Carbon
Pyrolytic Carbon Felt Composites. Advanced Heatshie/d SC-l1FW-Y12-7, A Process History.
Techniques for Material Investigation and SC-DR-7D-425, Sandia Labs.
Fabrication 14, National Symposium and Webb, R.D., 1985, Oxidation-Resistant
Exhibit, Society of Aerospace Material and Carbon-Carbon Materials. Metal Matrix, Carbon
Process Engineers, Paper II-4B-2. and Ceramic Matrix Composites, (ed. John D.
Buckley) NASA CP-2406, pp. 149-162.
HAND LAY-UP AND BAG MOLDING 16
D .R. Sidwell

16.1 INTRODUCTION Table 16.1 Advantages and disadvantages of hand

lay-up structures
1bis chapter presents practical fabrication and
tooling methods that have been successful for Advantages
high performance applications. Increasing per-
Design flexibility
formance reqillrements limit traditional hand Large and complex items can be produced
('wet') lay-up methods. Today's prepreg resin Minimum equipment investment is necessary
systems allow for longer working life and low Tooling cost is low
cure temperatures. Improvements have made The start-up lead time and cost are minimal
the current prepreg systems the choice for high Design changes are easily effected
performance composite structures. Molded-in inserts and structural reinforcements
Composites are a combination of a high are possible
Sandwich constructions are possible
performance resin matrix and various fiber Proto-typing and pre-production method for
combinations, which have brought about a molding processes
change in the engineer's traditional approach Semi-skilled workers are needed, and worker
and allow matrix alloying for specific struc- training is minimal
tural applications. Improved materials, Disadvantages
analysis and manufacturing methods have let
traditional composite structures reach new The process is labor-intensive
limits of achievement. Table 16.1 presents Only one tooled (molded) surface is obtained
Quality is related to the skill of the operator
some of the advantages and disadvantages of It is a low-volume process
hand lay-up of composite structures. Longer cure times required
Composite structures have been manufac- The waste factor can be high
tured from fiberglass and wood to some of the
most expensive 800 GPa pitch fibers for space
applications. The engineer's imagination is the a structural part upon which the external paint
only limitation in the fabrication of composite finish is applied. The reverse of this procedure
structures. Large, complex co-cured structures occurs with composite materials fabrication,
are successfully manufactured with all engi- where the formless materials harden and take
neering and manufacturing disciplines the shape of the container into which they are
working as one. applied. The makeup of the resulting laminate
In the fabrication of wood or metal prod- is an elementary engineering material. The
ucts, flat sheet stock is joined together to form physical properties can be changed by varying
the resin and fiber ratio, the type and direction
of reinforcement and the type of resin matrix.
Handbook of Composites. Edited by S.T. Peters. Published Today bag molded (vacuum and pressure)
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 composites provide higher performance and
 History 353

are competitive with alternate types of con- tional and woven product forms (fiberglass,
structions that result from optimization of graphite and aramid) were developed for com-
process controls, design refinements and mercial applications; aerospace structures were
improved materials. Composite designs and the prime users of 100% graphite materials.
bag molding processes are interrelated by pro- The fabrication of a 1979 Ford LTD with a
duction goals, the desire to control finished weight of 1138 kg (2504Ib) proved
manufacturing costs and by service or mission that graphite could be used as reinforcement
requirements. The responsibilities shared by for just about any automotive component
fabricators and designers include: including bumpers, hoods, frames, doors and
drive line. The major drawback to automotive
• reproducibility and knowledge of process-
use of these composites was the processing
ing characteristics of candidate materials;
cost. During the 1990s, processing times were
• knowledge of requirements for facilities
reduced to acceptable levels (less than 5 min)
and tooling including provisions for assem-
for automotive components, using resin trans-
bly, inspection and quality control;
fer and resin injection molding. Improved
• reduction of total part count;
mechanical properties, lower costs and
• avoiding design as replacement for sheet
improved resin systems allowed for greater
metal;
inroads for fiber-reinforced plastics in the
• selecting the best material for the applica-
automotive industry.
tion.
The use of ultra high modulus pitch-based
carbon fiber in modern communication satel-
16.2 HISTORY
lites has led to refinement in the art of hand
lay-up and vacuum bag molding. The primary
Composite materials first arrived in the New structural element of some communications
World with the early Spanish explorers who satellites is a central cylinder. A typical cylin-
produced adobe bricks for the building trades der is about 180 kg (80Ib) and supporting a
from a mix of straw and adobe clay. Modern 13000 kg (6000 lb) payload at launch. A typical
day users developed the chopper gun and cylinder consists of 520 GPa (75 msi) pitch
fiberglass roving that use low cost inexperi- fiber skin laminates and aluminum honey-
enced labor; many fabrications utilize 'wet' comb core, the structure is oven co-cured
lay-up for making bath tub enclosures and utilizing a nylon vacuum bag. The require-
truck hoods. Early use of carbon fiber compos- ments for ever lighter structures created the
ites was primarily for secondary structure need for ultrathin unidirectional prepregs.
aerospace applications. In the early 1970s, the Solar array panels for a modern communi-
first woven graphite fabric was developed for cations spacecraft required the use of
aerospace structural applications. 0.0004 cm (0.001 in) thick unidirectional
Complex and highly detailed hybrid (unidi- 520 GPa (75 msi) graphite prepreg. Solar array
rectional and woven carbon materials) panels, made with three-ply skins bonded to
structures were developed for an interconti- an ultralight honeycomb core, are manufac-
nental ballistic missile (ICBM) program. It did tured on a regular production basis, utilizing
not take long to integrate the woven graphite vacuum bag molding process and oven cure.
prepregs into mainstream production, reducing An example of what can be involved in a
labor costs over 70% from that of the previous high visibility composite materials application
unidirectional tape structures. During the was the 1992 America's Cup Challenge, the
1980s, composite hybrids were developed first application of graphite composite materi-
using the best of all material forms available. als for the America's Cup racing yachts. The
The hybrid structures utilizing both unidirec- example utilized here is the 'Spirit of
354 Hand lay-up and bag molding

Australia' program entered by the Darling ial from the supplier, to the freight forwarder,
Harbor Yacht Club, Sydney, NSW, Australia. airline, customs, importer clearing house and
(See section 16.7.2.) Design and analysis were then delivered to the manufacturing facility.
performed by Ian Murray and Associates, This effort can go astray; the shipment can be
yacht designers; the design parameters were off-loaded at an intermediate stop or off-
established by the America's Cup Rules loaded at the incorrect destination. Sea
Committee. The program objective, deter- shipping requires 30 days to get material from
mined in large part by available funding, was the manufacturer to a sea forwarder, provid-
to manufacture one America's Cup racing ing the container is off-loaded as scheduled.
yacht. A program of this magnitude requires a All of these considerations combined with
substantial amount of material. The 'Spirit of availability of funds and tight fabrication
Australia' was a relatively small program schedule playa primary role in design and
compared with some of the other syndicates. fabrication decisions.
Twenty-seven America's Cup yachts were fab-
ricated, representing eight countries, with a
16.3 MATERIALS
total investment of over $600 000 000. The
material combinations utilized in the hull con- Fiber reinforced composite materials are now
struction of the various entries included being used to satisfy the high strength and/or
Spectra, aramid, graphite and wood. The over- stiffness-to-weight requirements of aerospace
all goal was to achieve the specified minimum structures. Fibers of extremely high strength
laminate density and still retain the required are incorporated into a relatively weak matrix
load carrying capabilities. that is used to bind the fibers; the combination
Various hybrids of Spectra/graphite, results in a light, extremely strong structural
aramid/graphite and graphite/wood were material. Experience and theoretical studies
utilized to achieve the design results. The pro- have shown that fibers will always be stronger
gram for 'Spirit of Australia' required much than bulk materials. Some of the properties
dedication from supporters and suppliers to that can be improved by combining the con-
make it to the elimination trials with competi- stituents into a composite are strength,
tors. There were significant other concerns in stiffness, wear resistance, corrosion resistance,
addition to the normal decision factors in fatigue life, temperature resistance and electri-
selecting composite materials for a program, cal and thermal conductivity.
such as mechanical properties, tack, resin flow The relationship between resin viscosity and
and gel. These other concerns can have a large the cure cycle can be used to obtain maximum
influence on the program outcome and require performance in a composite structure. Figure
planning the materials' delivery schedule, 16.1 shows the relationship between resin vis-
testing, packaging, export license, custom for- cosity and a stepped autoclave cure cycle. The
malities and shipping method. If the program curves shown are for a typical 121°C (250°F)
is within a few hours flight time, the prepreg is curing epoxy resin system. The viscosity will
usually packed with dry ice and shipped on vary with a change in heat-up rate and tem-
the next available aircraft. On international perature. This information relating the effect of
programs, the weight of just the dry ice can rate and temperature is available from the
equal the material cost and delivery takes 3--4 material suppliers. Using this information as a
days. Scheduling becomes the critical path in starting point, a cure cycle can be defined for a
overall planning. In the case of 'Spirit of specific composite structure. The cure cycle
Australia', America's Cup challenge 1993, air production capability becomes increasingly
shipments were a minimum of three days. The important as the complexity of the composite
three days were a best effort to get the mater- structure increases. To help in planning a
 Materials 355

Apply 689 KPa

(100 PSI)
when Part
Reaches 78 DC
Heat at 2 DC :t:3 ce, Hold 15 Heat at 2 DC per Hold 120 Minutes at Cool Down to 75
M· ut s
It-----+I_'n_e_ _-+I ______-+I _ _ _ _ _--1I_____-+I_
Per Minute Minute 120°C DC
Cure Cycle

125

100

-u;
~
~ ~
f 75 :::>
!
~ 10
~
>I: E
{E.
·us
CD 50
II:

15
.1
o 200
Time (Minutes)
Resin Flow I I ..
Initial Resin Removal of Maximum Resin Resin Crosslinking and Cool Down
Flow Entrapped Air Resin Flow Gel
and Resin
Bleed

Fig. 16.1 Typical autoclave stepped cure cycle.

composite structure, the material supplier can spray-up. In contact molding, after a gel coat-
provide information on specific applications, ing is applied to the mold (if required),
processing parameters, material properties, fiberglass chopped mat, cloth or woven roving
test data and fabrication methodology for the is placed into the mold, saturated with resin
material. Because most materials have compet- and brushed or rolled to compact the material
itive products available, one must consider not and remove entrapped air. This process can
only the price and delivery available but also also utilize 'wet' prepregs. The wet composite
local technical support in selecting a supplier. is then vacuum bagged to compact the mater-
There are two basic types of composite ial and remove entrapped air. Spray-up
matrix fiber combinations: chopped-fiber reinforced composites consist of relatively
reinforced composites and continuous-fiber short fibers mixed with a matrix or binder
reinforced composites. The selection of the material resulting in a composite with gener-
proper matrix material for use in a particular ally lower physical properties.
structure is influenced by many factors Continuous fibers in a matrix material result
including weight limitations, environmental in a composite that exhibits higher strength and
resistance and types of loading. stiffness than the chopped-fiber counterparts.
'Wet' lay-up procedures are divided into Fabrication of parts that use continuous-fiber
two main types, contact or open mold and reinforced composites can be achieved by
356 Hand lay-up and bag molding

several techniques such as hand or machine 1. Thermoplastics. Thermoplastics may be

lay-up, filament winding and pultrusion. shaped, remelted and reshaped, much like
Hand lay-up composites are a special case candle wax. Examples of thermoplastics are
of continuous fiber reinforced composites. nylon, polysulfone, polyethylene, polyvinyl
Layers of unidirectional or woven bidirec- chloride and polyarylsulfone.
tional composites are combined to result in a 2. Thermosetting. Thermosets, once shaped and
material exhibiting desirable properties in one set, cannot be remelted and reset (like a
or more directions. Each layer is oriented to cooked egg it cannot be uncooked). When
achieve the maximum utilization of its proper- heated, a thermosetting plastic becomes
ties. Layers of different materials (different increasingly infusible. It undergoes an irre-
fibers in different directions) can be combined versible chemical change called
to further enhance the overall performance of polymerization (a process in which simple
the laminated composite material. Some of the molecules combine to form more complex
initial considerations that must be considered ones) during curing of the material.
before undertaking the manufacture of a com- Examples of thermosets are polyesters,
posite structure are: epoxies, phenolics, cyanates and poly-
imides.
• performance goal to be achieved;
• can the structure be designed to reduce part In advanced composite materials manufac-
count? ture, there are controls to provide a workable
• do not duplicate a sheet metal design; resin system. The main control points include
• manufacturability - in-house or outside volatile content to minimize voids and delam-
supplier; inations, tack that controls the handling of the
• overall cost compared with performance; prepregs, flow to allow removal of excess
• use only the amount of expensive materials resin and entrapped air and gel that deter-
needed. mines the point at which the resin starts cross
linking.
Many structures can be designed and manu-
factured with fiberglass laminates using
small amounts of graphite, aramid or Spectra, Volatile content controls
hybridized to obtain the additional stiffness
or reduced density. The high performance of Quality control test results do not distinguish
graphite (both PAN and Pitch) is attractive inert from reactive volatiles but are used to
for fishing rods, tennis rackets, golf clubs and establish volatile content values. They provide
racing applications. Special incentives for qualitative evaluations of prepreg advance-
performance can make cost a secondary ment, degree of volatilization of solvents and
objective. amount of degradation due to aging. Volatile
contents measured prior to lay-up are com-
pared to values when the prepreg was first
16.3.1 MATRIX MATERIALS received and may provide indications of
Plastics are used as the matrix material in excessive aging.
most fiber reinforced composites. These
materials are generally of high molecular
Tack controls
weight and intricate molecular complexity.
Their form and properties can be transformed Tack is the adhesion characteristic that is con-
by the techniques of applying heat and/or trolled to facilitate lay-up operations. It is
pressure. In general, plastics can be divided affected by resin and inert volatile content,
into two categories: prepreg advancement and humidity and
 Materials 357

temperature of the lay-up room. Sometimes, individual plant fabrication conditions. Air
tack is increased by increasing resin and conditioning minimizes local temperature and
volatile content, less prepreg advancement, or humidity sensitivities; reducing the tempera-
a slight increase in lay-up room temperature. ture makes heavy tack manageable. Judicious
At other times, tack can be modified by the use of hot air guns can improve dry tack.
prepreg manufacturer by changes in resin for-
mulations or new additives blended into the
Flow controls
resin. Alterations to the formulations should
not be accepted without prior requalifications Plow measurements indicate resin capabilities
by the user. Often, properties are severely to fuse successive plies in a laminate and bleed
affected for a cured laminate; therefore, per- out the void producing gas reaction by-prod-
formance and durability must be verified. ucts. Flow is also an indicator of prepreg age
Prepregs with excessively heavy tack gener- or advancement. It is often desirable to opti-
ally cannot be handled without grossly mize resin content and tack to attain adequate
disrupting resin distribution and fiber orienta- flow but, in some cases, flow is controlled by
tion or causing a roping (fiber bundling) of the thickening additives in a resin.
composite laminate. Excessive tack makes it
difficult to achieve reproducibility in compos-
Gel times
ite structures since undetermined amounts of
resin are always removed when the release Gel time is an indicator of the degree of cure of
film or backing is separated from the prepreg. resin systems. The amount of staging or
In general, all the disadvantages of wet lay-up advancement limits gel time for prepregs.
systems are inherent in the excessively tacky Most prepregs are formulated to attain a use-
prepregs. fullife (out time) of 8 days or more at standard
Prepregs with no tack are either excessively atmospheric conditions. Out time can be pro-
advanced or have exceeded their normal stor- longed by storage at -18°C (OOP), but each time
age life. Such materials cannot attain adequate the prepreg is brought to equilibrium at room
cured properties and should be discarded. temperature, the useful life is shortened. Gel
Exceptions are silicones and some polyimides time measurements can be used as quality
that can only be prepared with no tack. Lay- control verifications. Criteria based on those
ups with these materials are limited to those results determine whether to initiate more
occasions where lower mechanical properties costly property testing or to dispose of an
can be tolerated in exchange for improved overage prepreg.
heat resistance or electrical properties.
Thermoplastic prepregs typically do not
Storage
exhibit much tack.
Most prepregs are staged to attain more Most advanced composite hand lay-up mate-
manageable tack. The tack qualities should be rials that are utilized today are
adequate for adhering the prepreg to prepared time/temperature critical. Precautions must
molding surfaces or preceding plies for a lay- be taken in handling, storage and processing.
up and still light enough to release the backing During the initial processing, where resin is
film without loss of resin. Provided that cured impregnated onto the reinforcement fiber, the
laminates will not be adversely affected, tack degrading of the materials starts the moment
qualities can be specified to require the the resin is applied to the reinforcement.
prepreg to remain adhered to the backing until However, cases exist where material has been
a predetermined force is applied to peel it off. stored at -18°C (OOP) for as long as 5 years and
Tack requirements can be modified to suit was capable of molding a commercially
358 Hand lay-up and bag molding

acceptable laminate. A new generation resin Fiberglass fiber

system, the cyanates, tends to be more stable • high impact strengths;
than epoxies and outlife has been increased • good overall performance;
from 10-14 days to 30-45 days. Storage at • low costs;
-18°C (OOP) has also increased from 6 months • available in many forms;
to 1 year. • higher density.
In planning a project, one should consider
the amount of material needed within a given Aramid fiber
period of out time to complete the lay-up and • high impact strengths;
cure of a composite component. If one can fab- • high specific modulus;
ricate enough composite parts in a 10 day • poor compression strengths;
period to utilize a 22 kg (50 lb) roll of mater- • high costs;
ial, then that is the size that one should • difficult to handle.
prOcure. If not, reduce the size of the roll Spectra fiber
accordingly. By doing this, the material roll • very low density;
has to be removed from the freezer only once • difficult to handle;
and the recorded out time can be accurately • bond interface poor;
maintained. • very high impact resistance;
• high costs.
16.3.2 REINFORCEMENTS
Carbon fiber reinforced composites
In continuously reinforced composites, the
fibers carry nearly all of the load applied to the Carboni graphite fiber is an attractive material
system, with the matrix transferring the load because it has the best balance of material
into the fibers. In chopped fiber reinforced characteristics to compete with metals. Carbon
plastics, the load is distributed between the fiber is the end product of a series of manufac-
fibers and matrix, depending upon the nature turing processes that begin with the basic PAN
of the constituents, percentages of each and the precursor that is carbonized and graphitized.
orientations of the fibers. Selection of the It is then collimated, woven or, chopped to
proper fiber for a specific application depends become tape, cloth, or molding compound
upon the type and magnitude of loading, oper- when prepreged with an appropriate matrix
ating temperature and cost limitations. The material. Carbon composites offer many
following are attributes of several applicable advantages over conventional materials.
fibers: Major advantages are:
• high specific strength and stiffness;
PAN-based carbon fiber • can be woven into cloth;
• high strengths; • dimensional stability at temperature;
• improved strain to failure; • tight product control;
• workable; • ease of manufacturing;
• available in many forms. • high fatigue resistance;
• extremely complex structures can be fabri-
Pitch-based carbon fiber cated.
• high to ultra high modulus;
• low strain to failure; While offering significant advantages, carbon
• high cost; fiber does present the following problems rel-
• can be difficult to handle. ative to conventional materials that must be
 Materials 359

considered when designing for specific appli- Some are better than others; some have appli-
cations: cation only to specific needs of a particular
structure. Some of these are presented to show
• low strain to failure compared to most metals;
the basic selection process that is required in
• relatively low impact resistance;
the initial design phase and the manufactura-
• potential for galvanic corrosion;
bility of each product form into a successful
• potential matrix bonding problems;
composite structure.
• low interlaminar shear;
All fibers can be obtained in many different
• high cost; forms: unidirectional tapes, unidirectional
• chemical resistance.
tows or yams, woven cloth, chopped and in
hybrid combinations. The combinations are
Non-woven and collimated tape unlimited in hybridizing fibers together. The
During lay-up with unidirectional tape, there following list shows some of the hybrid rela-
are several techniques that will help achieve a tionships and Table 16.2 shows some of the
cost effective structure. As soon as unidirec- properties of composite material forms.
tional tape has been selected, start preliminary Carbonlfiberglass
planning as to ply orientations and develop • high impact strengths;
basic patterns for the initial fabrication. The • high modulus;
results of this preplanning will be that wastage • high performance;
will be minimized and better quality can be • cost effective.
built into the finished part.
Using unidirectional tape on contoured sur- Spectra/carbon
faces presents a number of lay-up problems. • improved bond interface;
As the tape is placed on the mold surface, the • little improvement in overall properties;
orientation can change dramatically during its • low density.
application. In order to minimize angle Aramidlfiberglass
changes, narrow pieces of unidirectional tape • reduced overall costs;
can be used. Scrap material can be used for • limited applications.
local reinforcements and to fill gaps between
pieces of unidirectional tape. Gap filling pieces Aramid/graphite
up to 3.2 mm (1/8 in) wide are generally • limited applications;
acceptable. • high costs;
• highly resistant to cracking;
• better impact resistance.
Woven fabric
Joints with woven materials can be more of a 16.3.3 EXPENDABLE MATERIALS
challenge than those with unidirectional tape.
On a highly stressed composite structure, the Supporting materials can contribute to a pro-
selvage edge must be removed to minimize gram's success as much as the composite
the discontinuities in the lay-up joints. Joints structures. If there are problems with the vac-
in succeeding plies must not be adjacent but uum bag, release film, or sealant tape, the final
staggered by a minimum of 51 mm (2 in). product quality can be compromised although
the overall unit cost of most of these materials
is low. Specifications to control support mate-
Hybrid materials
rials have been slow to develop with the
Many combinations of composite materials industry. There have been problems with
have been developed over the past 15 years. expendables such as release coating of porous
360 Hand lay-up and bag molding

Table 16.2 Limits of composite material forms

Material Fiber Tensile" Compression c Tensile"

form modulus properties properties modulus
CPa MPa MPa CPa
(msi) (ksi) (ksi) (msi)
Fabric
Graphite 8HS 228 586-724 522-689 62-83
(33) (85-105) (80-100) (9-12)
Graphite 5HS' 228 758-896 758-827 76
(33) (110-130) (110-120) (11)
Graphite 5HSb 537 927-1114 240 188
(75) (132-158) (34) (26.7)
E-glass 8HS 72.4 462-483 469-510 22.0-26.2
(10.5) (67-70) (72-74) (3.2-3.8)
Kevlar® 49 8HS 130.9 510-579 138-179 29.6-33.8
(19.0) (74-84) (20-26) (4.3-4.9)
Unidirectional tape
Graphite 228 1586-1792 1517-1724 124-138
(33) (230-260) (220-250) (18-20)
Graphite' 228 1930-2137 1379-1723 138-152
(33) (280-310) (200-250) (18-20)
Graphite 345 1069-1172 896-1034 200-228
(50) (155-1700) (130-150) (29-33)
Graphiteb 520 1760 420 300
(75) (256) (61) (44)
Graphite 720 1760 330 420
(105) (256) (48) (61)
Graphite 780 1760 330 450
(114) (256) (48) (65)
E-glass 72.4 1034-1171 483-620 41-55
(10.5) (150-170) (70-90) (6-8)
5-2 glass 86.8 1241-1379 552-690 48-62
(12.6) (180-200) (80-100) (7-9)
Kevlar® 49 130.9 882 64.9
(19) (128) (9.4)

• High strength >3447 MPa (>500 ksi) PAN based graphite fiber
b High strength >3630 GPa (>500 ksi) pitch based graphite fiber
C Composite laminate properties

Teflon fabric where there was insufficient coat- encounters a problem utilizing support mate-
ing material to provide an effective release, rials, contact the supplier who can provide
resulting in a bond between the Teflon fabric technical assistance on its products. During
and the composite structure and when using a the lay-up of composite structures, care must
perforated release film, the perforations were be taken to insure that all areas are covered
tom rather than pin pricked, allowing exces- with a release or separator film. The bleeder
sive resin bleed. and! or breather will bond nicely to composite
Specifications would help limit the incon- laminates if there is no separator.
sistency of some of these products. If one
 Tooling 361

16.4 TOOLING for steel, include superior durability, ample tol-

erance for elevated temperature service and
Tooling includes materials, equipment, or
good thermal conductivity.
forms onto which (or into which) the product
Ceramics have favorable characteristics for
is made, assembled, or cast. Tooling issues are
molds. They have the lowest coefficient of
a result of a number of interacting require-
thermal expansion and their thermal conduc-
ments that are considered when selecting the
tivity approaches that of some hardened tool
most cost effective tooling. These require-
steels. However, ceramics are brittle at ambi-
ments are as follows:
ent temperatures and they must be protected
• ability to achieve a uniform heat up rate, from processing and handling hazards. One
taking into account the mass of the tool; way to get protection is to enclose ceramic
• allowing sufficient movement of the lami- inserts in a steel case.
nate while achieving pressure and Dual steel molds are candidate tools for
compaction in all regions; reproducing high quality composites.
• accommodating resin flow; However, these tools are costly and produc-
• facilitating or allowing for removal of cured tion quantities are often not sufficient to
part; amortize tooling costs at competitive prices for
• realistic tolerance for tool and laminate; the production items so less costly alternatives
• finishing requirements - coarse finish can are desirable.
create lock-on problems; Aluminum molds are less costly. Although
• adequate area for applying sealant tape for thermal conduction is better for aluminum
vacuum bagging and test coupons as than for steel, the tools are less durable and the
required; thermal expansion is large. Shallow or flat
• provisions for vacuum fittings and han- mold plates are usually limited to cures below
dling of tool. 177°C (350°F). Other metal tools include
sprayed or electroformed molds reinforced
Tooling is less expensive for vacuum bag and
with cast backings. Alternate types of tooling
autoclave/ oven molding than for matched die
can include composite molds usually based on
molding methods. Molds and molding plates
high temperature resistant cast or laminated
are required to withstand curing conditions
epoxy resins.
without distorting or degrading and to toler-
Master forms for laying-up composite tools
ate handling during the fabrication processes.
can also be fabricated using any of the materi-
They are not necessarily resistant to unbal-
als described; a mock-up model of the item
anced pressures. The higher costs of composite
may be used, or plaster masters can be pre-
tooling can be amortized by taking advantage
pared from models. The quality of the plaster
of the improved capabilities to mold complex
masters depends on the strain compatibility
constructions. Composites that may ordinarily
between the plaster and its reinforcements and
require secondary bonding are often more eco-
on the condition of the hardened surface.
nomically co-cured.
Composite tools can be laid-up using fiber
orientations that most closely match the
expansion of the items to be produced.
16.4.1 MOLD PREPARATIONS
Fiberglass and graphite fibers are the principal
Coefficients of thermal expansion for conven- reinforcements. Woven fiber reinforcements
tional tooling materials and composites are are the most economical to use. Mold mainte-
given in Chapter 25. For metals, the coefficient nance is best relegated to specialized
for steel compares most closely with the coeffi- personnel while preparations for the bag
cients for the composites. Other characteristics molding processes are assigned to production
362 Hand lay-up and bag molding

personnel. A successful practice is to provide PTFE (polytetrafluoroethylene) is stable and

production personnel with soft tools and sol- is often contained in mold releases for service in
vents that do not degrade the molding excess of 260°C (500°F). The mold releases con-
surfaces. If the production tools and solvents tain suspensions of micropulverized PTFE in a
are inadequate for removing debris and clean- volatile dispersant. Depositions on mold sur-
ing, the molds are taken out of service for faces do not form continuous films, but the
maintenance, repair, or replacement. After PTFE particles provide excellent dry lubrication
they are returned to service, they are solvent- for the release of the cured composite.
wiped clean and mold release agents are Furthermore, the residual particles on the cured
applied. composite surfaces are easily removed with a
solvent-wipe. Since a variety of commercial
mold releases that contain fluorocarbon (or
16.4.2 RELEASE AGENTS
equivalent) are on the market, it is essential that
Release agents for bag molding composites manufacturers' recommendations on uses and
include carnauba paste wax, aerosol dis- limitations be scrupulously followed.
pensed compositions that contain carnauba, Silicone oils and greases are to be avoided
fluorocarbon resins, or silicone resins, plastic since they are the most persistent contami-
films and metal foils. On most occasions, the nants of molded composite surfaces. They
wax or resin mold releases do not contaminate release secondary bonds and coatings from
the composite surfaces excessively nor prevent composite surfaces as effectively as they
subsequent secondary bonding or coating release the cured composites from the molds.
operations. Prior to bonding, the composite Silicone oils and greases migrate and defeat
surfaces are cleaned with solvent and lightly most removal attempts. They contaminate the
sanded to remove resin gloss. The user should solvent wetted cloths and sand papers so that
be aware of national or regional limitations on instead of removing the silicones, they spread
solvent usage due to toxicity or ODS (ozone them. Contaminated surfaces may be salvaged
depleting substances) concerns. Peel plies can for painting by sandblasting with virgin grit.
be used to protect clean surfaces for primary Table 16.3 summarizes the precautions nec-
adhesive bonds. Plastic films, metal foils and essary for successful hand lay-up and vacuum
sprayed metal coatings also serve as release bagging operations
agents when they are integrally laminated to
the co-cured composites.
16.4.3 MOLD DESIGN
Both the polished wax surfaces and the
sprayed wax coatings are excellent mold Mold design is a function of cost and projected
releases for composites cured below 121°C life and/or use. A production mold should be
(250°F). However, the wax degrades and dis- made carefully from the best materials. Such a
colors the composites at higher molding mold will be designed by an experienced
temperatures. Commercial fluorocarbon mold designer who will incorporate the necessary
releases are used for higher cure temperatures. thicknesses, materials, structural reinforcement
FEP (fluoroethylene propylene) mold and hardware required for the intended use.
releases form a continuous film on mold sur-
faces. Although the condition of the release
16.4.4 PATIERNS (PLUGS)
film is easy to maintain below 177°C (350°F),
the coating degrades at higher temperatures A pattern (plug) is a temporary form in the
above this. Fluorine which is noxious, corro- exact shape, contour and finish of that to be
sive and highly toxic can be released from the molded. (If the outer shape of the items
polymer above 177°C (350°F). desired the inside contour is used.) Patterns
 Tooling 363

Table 16.3 Hand lay-up and vacuum bagging precautions

Mold release application Breather

• Selection of correct mold release • Able to maintain vacuum path
• 'Seasoning' new tool to insure coverage a • Care not to puncture nylon film
minimum of three coast
• Compatibility of resin system and mold release Bagging film
• Repeated applications can cause excessive • Higher the percent elongation the more
buildup forgiving
• Pliability
Sealant tape • Defect free
• Ease of use, release from backing • Select for temperature performance
• Double or single strip application
• Removal after cure Thermocouples
• Cheapest when suitable • Through bag/sealant tape, potential vacuum
leaks
Flash breaker tape • Placed on outside of bag - reliability
• Check compatibility
• Useful for holding prepreg during lay-up and Vacuum fittings
heat debulks • Caution not to allow resin to fill
• Source for potential leaks
Separator film • Removal from bagging film before disposal
• Compatability and will release as intended • Integral with tooling
• Quality control on perforations Vacuum lines
• Drapability for complex shapes
• Ensure fittings do not leak
• Elongation (%) (high elongation)
• Hose has not been crushed
Bleeder • Not pinched off
• Not all bleeders the same, select for application • Not filled with resin
• Ensure no compaction during cure
• Potential seal off during cure
• Excessive bleed can saturate bleeder

are made of many materials: wood, plaster, the inside contour of the hull is constructed
plaster / metal and other combinations. Almost from stanchions and stringers and covered
any· material can be considered as a pattern with strips of wood, with a laminate of ± 45°,
material if it holds its shape. 90° layers. The tooling plug is finished and
It is assumed that when only one large com- covered with mold release.
posite structure is required, such as a 23 m
(75 ft) America's cup yacht hull, the cost of
16.4.5 INTEGRATING INSPECTION AND
making a pattern and a mold in order to make
MACHINING
the hull may not be justified. However, the
construction of a pattern that becomes a male A machine tool is for machining. If it isn't mak-
plug can be cost effective for high performance ing chips, it's wasting time, so keep non-cutting
composite structures. To avoid excessive cost time to an absolute minimum. That's the stan-
with this tooling approach, one must remem- dard philosophy most shops try to live by. In
ber that it is the total cost of the end product fact, many shops are investing in pallet shuttles,
not just the cost of the pattern (plug) that must quick-change fixtures, tooling systems, rapid
be considered. A limited use plug for a large transverse fixtures, programming and schedul-
marine hull 14-23 m (45-75 ft) in length would ing systems to keep spindles tuming and
be made as follows: a simple wood frame to cutting tools working at optimum capacity.
364 Hand lay-up and bag molding

Inspecting the workpieces right on the five- parts. If the workpiece will not fit into the hard
axis mills, has the equipment functioning as gauge, it has out-of-tolerance features and will
both machine tool and coordinate measuring not fit mating parts. The workpiece is rejected.
machine, so that the production of parts and A soft gauge can be used to make similar
producing inspection data become equally either-it-fits-or-it-doesn't comparisons. Instead
important. Such a radically different plan of placing two physical objects together, two
means that design, numerical control (NC) CAD models are laid one over the other on the
programming, machining and inspection can- graphics screen. The software version of the
not be separate functions. Just as each machine checking fixture is the soft gauge. The soft-
tool would have to serve more than one role, ware version of the workpiece to be inspected
one computerized database would have to is a geometric model constructed from inspec-
share the same information with designers, tion data. Out-of-tolerance conditions will be
programmers, operators and inspectors. This just as conspicuous in this comparison, but
combination allows a shop to machine, inspect analysiS is far more complete and much faster.
and analyze any surface without removing the Moreover, a soft gauge is created directly from
work piece from the five-axis machine. This the original design data. Because it is created
system helps produce higher quality tooling on a computer screen instead of in a tool room,
with significant gains in productivity. a soft gauge can be constructed quickly and
A large machine bed will accommodate modified easily. It spares the high cost of
unusually long workpieces and also leave building and validating a hard gauge.
room for smaller workpieces to be clamped on
one end while another workpiece is being
Closed-loop machining
machined at the other end (see Fig. 16.2).
Closed-loop machining begins with electronic
data representing part geometry from the cus-
Soft gauge
tomer. This data describes the outer surface of
A soft gauge can be compared and contrasted the customer's end product. The CAD system
with a 'hard' gauge such as a conventional then creates a 3D model of its surface. Once
checking fixture used for inspection and qual- this surface has been established, all manufac-
ity control. If a workpiece drops into the hard turing operations will be derived from and
gauge, it is acceptable and will fit with mating related to it. NC tool paths will be generated

Fig. 16.2 Closed-loop machining - mounting vari- Fig. 16.3 Closed-loop machining - touch probe
ous tools. (Courtesy of Coast Composites, Inc.) inspection tool. (Courtesy of Coast Composites, Inc.)
 Tooling 365

from it. Using dynamic display of the tool the composite materials. For complex shapes
path, programmers can visually verify the NC with integral stiffeners, each block of silicone
program, check clearances and make sure rubber is wrapped on all but one side, the side
gouges are avoided. in which the tooling rubber is removed (see
The inspection path will be generated from Figs 16.4(c) and (d». In tooling a thermal
the same surface geometry. By referencing the expansion molding, it is best to avoid using
soft gauge, the inspection path will be sure to rubber on both sides of a laminate as illus-
include checks of all critical features. The path trated in Fig. 16.4(e) unless straight edges are
of the probe can be visually verified in the not critical. Vacuum bag assist (see Fig. 16.4(f)
same way as the NC program. provides an alternate method.
After executing the NC program, the work- The linear thermal coefficient of most sili-
piece can be inspected immediately using the cone rubbers that have been measured fall into
touch probe in the spindle (see Fig. 16.3). This the range of 1-2.1 x 10-5• This range is consis-
inspection can be considered in-process, tent over a 23-246°C (75--480°F) temperature
because the workpiece is still fixed on the range. The rubbers are said to have a linear
machine tool and can be remachined without expansion of approximately 17 times that of
being moved or refixtured. This approach is carbon steel which is why they are used to
called closed-loop machining. mold composites by thermal expansion mold-
Results of this inspection routine are auto- ing techniques. Precautions in mixing some
matically used to create a 3D model of the silicone rubber compounds are required if full
features checked. By comparing this model to potential is to be achieved. During prolonged
the soft gauge, any out-of-tolerance conditions storage, the catalyst tends to separate and set-
can be identified. It will also show where addi- tle to the bottom of the container. Mixing the
tional machining passes will be required. catalyst prior to adding it to the base rubber
Final inspection can performed on the will allow correct mixtures and long tool life.
machine tool. These results are compared to
the soft gauge again to verify that the con-
toured surface of the graphite tooling will
produce the intended part. Using this machin-
ing approach for inspection on the machine
tool reduces inspection time by 80%. The
biggest savings come from eliminating work-
piece moves and additional setups and from
streamlined programming of the inspection
routines. By integrating inspection and
machining, overall manufacturing cycle time (a) CompreSSion molding
can be reduced by 30%.

16.4.6 THERMAL EXPANSION MOLDING

Thermal expansion molding techniques are

utilized for special applications of small com-
plex composite structures and composite
tubing with critical outside surfaces. Figures
16.4(a) and (b) illustrate the methods allowing (b) Oven cure critical outer surfaces
the expansion of the silicone rubber to provide Fig. 16.4 Thermal expansion molding. (Continued
the required pressure for the compacting of on next page.)
366 Hand lay-up and bag molding

Incorrect mixtures will be light or dark in color

and materials such as Silastic J® will start to
crumble within a few thermal cycles.
The use of thermal expansion rubber can be
hazardous if not planned well. The tooling
rubber can exert up to 6.9 MPa (1000 psi)
during its first few thermal cycles. The tool-
ing rubber requires the minimum of five full
cure heat cycles free standing after the initial
cure in order to stabilize the expansion charac-
(c) Enclo ed molding teristics. The tooling is capable of producing
50-75 composite parts before having to be
replaced.
One other problem with thermal expansion
rubber is with its removal from the composite
structure. Sharp pointed objects will have a
very lasting effect on tool life; once the rubber
is damaged, it will continue to tear, needing
replacement much sooner than usually
required. Silicone rubber is very slow to cool
down and extra time must be allowed because
the rubber is impossible to remove from the
(d) draft molding composite part until it has shrunk back to its
original size. Putting the thermal expansion
rubber tooling into a freezer can accelerate the
manufacturing cycle. However, since some of
the tools are heavy, due to steel outer encase-
ments, a 12-15 h cooling down period should
be planned into the manufacturing cycle.
Experimentation is suggested with this mold-
(e) Oven/press cure ing method since extreme pressures can be
generated and undesirable results may occur
if the molding method is picked arbitrarily.

16.5 BAG MOLDING PROCESS

Molding methods include vacuum bag, pres-
sure bag, oven and autoclave molding. Bags,
(f) Vacuum bag assi t
the thin and flexible membranes or silicone
rubber shapes, separate the laid-up construc-
Silicone rubber tions from atmospheric pressure during
composite cures. The bagged lay-ups in auto-
D Steel molds claves are usually vented to pressures lower
Compo ite laminate
than those applied to the bag. Consolidations
and densifications of the lay-ups are achieved
by the resulting pressure differentials across
Fig. 16.4 Continued. the bag contents. Consolidations are achieved
 Bag molding process 367

when the separate plies of prepreg in the lay- up assemblies. When the presses are closed,
ups and other adherents are bonded together. the sealed chambers are pressurized and
Densifications result in reduction of voids and heated to attain molding conditions similar to
removal of excess resin. Other results desired those of an autoclave. Unlike the specialized
of bag molding methods during cure include pressure bag molds, the modified presses are
prevention of blistering in the composites, used to cure many different composite con-
increased controls of pressure and heat appli- structions.
cation and control of the fiber/resin ratio. Autoclave and pressure bag molding condi-
Consolidations and densifications of vac- tions to 177°C (3S0 0P) and 1379 kPa (200 psi)
uum bag moldings can be achieved by are routinely attained. Newer, customized
atmospheric pressure alone as the bagged lay- autoclaves attain cure conditions that exceed
ups are evacuated throughout the cure cycles. 260°C (SOOOP) and 3447 kPa (500 psi). Pire haz-
The pressure-bagged and autoclaved-cured ards are greatly increased at elevated
composites are pressurized by hot gases. Vents temperatures and pressures. Pressure vessel
to the atmosphere or vacuum provide for the fires are minimized by uses of fire retardant
escape of the volatilized reaction by-products processing materials and inert pressurizing
and the entrapped air from the curing com- gas. Pire prevention measures include uses of
posites. If the pressures within the bag are not silicone rubber, nylon or Tedlar bags. Before
reduced from those applied to the bag, the cure cycles are initiated, the pressure vessels
membrane remains inert and there is no com- are purged of all enclosed air. After the ther-
paction. Of the three methods, vacuum bag mal cure is completed, the pressure vessels
molding is least limited as to the size of con- and their contents are cooled to 68°C (1S00P)
structions that can be processed. On a few before the pressure is relieved and the auto-
occasions, 'wet' lay-up vacuum bag molded clave is opened.
composites are room-temperature cured. Most
are thermally cured to produce improved
16.5.1 EXPENDABLE VACUUM BAGGING
properties. Thermal cures are best attained in
TECHNIQUES
air circulating ovens / autoclaves, but can also
be achieved in infrared heated and passive Bleed-out systems are devised to maintain
type convection ovens. reduced pressures within the bags' contents.
Pressure bag molding methods are efficient The bagged lay-up includes the bleed out sys-
for producing both deeply contoured struc- tem designed for the composite part. Bagged
tures and shallow composites. Sonar domes, lay-ups can be bled in two ways: vertically or
radomes and antenna housings are examples edge bled. The classical differences between
of deeply contoured composites. Architectural the two can be seen by comparing Pigs 16.5
panels, door panels and aircraft fairings are and 16.6. Many of today's resin systems are
examples of shallow composites. Heavy mostly 'net resin' and do not require any resin
molds are built to reproduce deeply con- bleed during cure. This allows for better con-
toured structures. Each specialized mold is trol of the resin content of the composite
constructed to withstand the elevated temper- structures. If a resin bleed sequence is pre-
atures and increased pressures required for ferred, the following sequence can be used as
the cures. a guide.
Shallow items may often be bag molded in
• The surface of the mold is prepared with
modified compression presses. The lower
the release agent.
press platens contain vents and vacuum lines.
• The composite plies are applied and rubbed
The upper press platens are made hollow to
out to remove entrapped air.
enclose the mold plates together with the laid
368 Hand lay-up and bag molding

Fig. 16.5 Vacuum bag edge bleeder - schematic.

NOTE HEAVY PROTECTION AGAINST BAG PERFORATIONS DUE TO INCREASED

AUTOCLAVE PRESSURES

Fig. 16.6 Vacuum bag vertical bleeder - schematic.

 Bag molding process 369

• A perforated release film is applied over the In a typical vacuum bag lay-up, there are
composite laminate and extended approxi- several methods available; some use double
mately 3.2 mm (1.25 in) beyond all edges. sealant tape side/side, some single and some
• A predetermined number of bleeder plies one on top of the other. The best system is the
are applied over the release film and one that works. The side/side method is used
extended to the perimeter of the lay-up. to provide some insurance during cure that
• A perforated release film is applied over the the bag will not shrink, pulling an edge off,
bleeders and extended 3.2 em (1.25 in) from causing loss of vacuum. The over/under
edge. method is used to provide ease of placement
• One or two layers of a non-woven breather of ears to allow some movement of the vac-
is placed over the lay-up and extended over uum bag. No matter which method is chosen,
the release film. it is important to remember that vacuum bags
• Sealant tape is applied around the perime- tend to pull down more than expected and
ter of the bleeder. can puncture, if bagging is over a sharp
• The vacuum bag is positioned and sealed. object. During the application of a vacuum
• The contents are evacuated and smoothed bag, 'ears' are required to facilitate the uni-
and the bag is checked and sealed against form application of vacuum to the composite
leaks. laminate.
• The bagged lay-up is ready to be cured. Vacuum bag bridging is one of the leading
causes of resin rich and excessive voids in cor-
In any bagging sequence, the types of release
ners of composite laminates. Figure 16.7
film, bleeder, breather and bagging materials
illustrates this common problem. One method
used vary from company to company and from
of eliminating bridging of the vacuum bag is
supplier to supplier. Each supplier has typical
presented in Fig. 16.8 by means of 'ears' in the
data sheets on expendable materials to achieve
bag. Another method to help reduce resin rich
the most efficient use of the materials.

VACUUM BAG

BRIDGING

1\SEALANT
TAPE

LAMINATE BEING FORMED

MOLD

Fig. 16.7 Vacuum bag bridging (Morena, J., Advanced Composite Moldmaking; New York, Van Nostrand
Reinhold,1988).
370 Hand lay-up and bag molding

(SEALANT TAPE

VACUUM BAG
4-INCH HIGH PLEAT OR FOLD

1\ SEALANT
TAPE

LAMINATE BEING FORMED

MOLD

Fig. 16.8 Elimination of vacuum bag bridging (Morena, J., Advanced Composite Moldmaking; New York, Van
Nostrand Reinhold, 1988).

Fig. 16.9 Large vacuum bagged structures. (Courtesy of Richmond Aircraft Products.)
 Bag molding process 371

and excessive voids in comers is the place- 16.5.2 REUSABLE VACUUM BAGGING
ment of an intensifier over the area, usually TECHNIQUES
placed between the separator film and There are material and recurring labor cost
breather. The intensifier can be molded rubber disadvantages to the use of expendable vac-
in the radius desired or some sealant tape to uum bags of plastic films for fabricating
fill the comer. 'Ears' may be required in sev- production composites. Expendable bags, laid
eral sections of a complex part. Experience will up of plastic films and associated sealants, also
determine the height of the ear for a specific incur recurring costs. Expendable bags can be
application; 10 cm (4 in) is a good starting laid up only once because of degradation dur-
point. Some will be smaller and some will be ing handling and the thermal cures.
larger depending upon the complexity of the Use of silicone rubber reusable bags can
component being vacuum bagged. reduce fabrication costs and defective com-
There is essentially no limitation on the size posite parts because of resulting work
of thermoset composite structures. The use of simplification and more positive control of the
the thermoset vacuum-bagged composites bag molding cure conditions. Figure 16.11
(autoclave, oven or integrally heated) will con- illustrates an example of a component being
tinue to provide excellent composite manufactured utilizing a reusable silicone
structures for many years to come. Figures
16.9 and 16.10 illustrate some more complex
uses for expendable vacuum bags.

Fig. 16.11 Disposable vacuum bagging (top);

Fig. 16.10 Complex vacuum bagging. (Courtesy of reusable vacuum bagging (bottom). (Courtesy of
Richmond Aircraft Products.) The Danner Corporation.)
372 Hand lay-up and bag molding

vacuum bag and an expendable vacuum bag. are combined when the product is made. The
There are often difficulties in having the facili- composite designer must consider how the
ties to handle large reusable vacuum bags load bearing fibers are placed and ensure that
weighing several hundred pounds. A cost eval- they stay in the proper position during the fab-
uation must include all aspects of the program. rication.
There are some very large aircraft components
utilizing expendable vacuum bag materials
16.6.1 DESIGN PROCESS
very successfully. The more complex the com-
posite structure becomes, the more effective With some large composite structures, poten-
the reusable vacuum bagging system is. tial problem areas can best be identified using
scale models. Working problem areas on an
individual basis, a major factor in the success
16.6 DESIGN
of a program is the amount of planning that
The fundamental information needed for any can take advantage of the work force experi-
design includes the stresses applied under ence. Successful composite structures are not
storage and use and the strength of the mater- fabricated by one person; they require team
ial used. Assume that the size, shape, quantity work from all disciplines. It is considerably
and rate of production have dictated the use of harder to make a smaller composite structure,
open mold techniques. Then the final thick- than make the full scale article. Procedures
ness, orientation and quantity of reinforcing developed for the scaled article can, however,
fibers are dependent upon the stresses that be easily translated into a full scale structure.
must be resisted, how often and for how long.
It is essential that the designer find out
16.6.2 FIRST ARTICLE FABRICATION
what strength can be built into the laminate.
This sets composite structures apart from The first article, also known as tool proof arti-
other types of materials, since the material is cle, can be used to provide information not
made· during fabrication of the product; the only on the tool to manufacture to the correct
percentages and orientation of the reinforce- tolerances but also to produce an acceptable
ment and the types of resins determine the composite structure meeting the design
properties of the final laminate. requirements. In addition, it can be utilized as
During the initial phase of the development a proof of the documentation of quality con-
of a composite structure, there is a need for trol inspection requirements, manufacturing
design, manufacturing engineering, tooling, procedure's verification and allow design
materials and quality control to provide inputs engineering to review overall requirements.
so that the selection processes can be estab-
lished. By coordinating early in a program,
16.6.3 DESIGN DETAILS
one can focus on the real problems of design
and manufacturing. The preplanning phase
Parts with severe contour and thickness
will allow for a program to develop at a more
variations
rapid pace. Confirming materials, manufac-
turing methods, tooling concepts and design It may seem easy to incorporate variations in
requirements early can avoid the extra time contour and thickness into the design of a new
and expense to attempt to make the composite product but in open mold products such vari-
fabrication process successful. ations must be made with caution. The
The design of composite structures, while molding operation requires laying the mater-
essentially similar to conventional design, ial on the mold to follow mold contour. H the
does have the added dimension that materials angles are sharp (900 without radius), the lay-up
 Design 373

will not follow the mold surface and will Changes in thickness
develop voids and resin rich areas in the lami-
To change thickness in open mold construc-
nate in the vicinity of the angle. For instance
tion is to add or remove plies of material. An
in inside right angle comers without radius'
abrupt change means that the plies must be
(Fig. 16.12(a» the laminate will not pack int~
carefully laid up in a precise pattern. An
the comer. When there are sharp outside cor-
abrupt change in thickness (Fig. 16.13(a»
~ers, the laminate (Fig. 16.12(b» will not wrap
results in a stress concentration and should be
tightly over the comer. The solution to such
avoided as delamination is sure to occur at
problems is to design with a generous radius,
such a point. The solution to this problem is
preferably 4.75-12.75 mm (0.187-0.500 in)
not to have abrupt changes but to gradually
inside and out. The laminate will then follow
change by stepping back or 'shingling' the lay-
the contour. Abrupt changes in direction are
up (Fig. 16.13(b».
high stress areas and tend to delaminate and
crack. They should be avoided and moderate
self reinforcing curvatures used. Openings
The best opening is a round hole; the worst, an
opening with sharp, non-rounded comers.

L
The solution to stresses in an opening is to use
large radii in the comers, to build up thickness
graduall~ at the sharp comers, or to design a
molded ill flange around the opening (Fig.
16.13(c».

(a)

(a)

(b)

II

_1-
-==i-
(b) (c)

Fig. 16.12 Corner lay-up techniques (a) radius cor- Fig. 16.13 Changes in ply lay-up (a) abrupt
ner; (b) no radius corner. changes; (b) stepped piles; (c) hole reinforcement.
374 Hand lay-up and bag molding

Joints and bonding However, these structures are not generally

highly stressed structures. .
Although common practice is to use joints
Large amounts of waste can be expensIve
intended for other methods of fastening, struc-
for any program. As a rule, in programs where
tural adhesives require joints of a special
is little or no preplanning, the waste factor can
design. It cannot be stressed too strongly that
be 20-35%. This may be acceptable for low
the practice of using ordinary joints that have
cost materials such as fiberglass but for carbon
been slightly altered can lead to disastrous
prepreg materials costing over US$5900 /kg
results. The type of joint used depends on the
for 827 GPa (120 msi) pitch fiber, the cost for
basic characteristics of the adhesive since
the waste can break an otherwise successful
structural bonds act over an entire area and
composite program. A well-planned. appro~ch
not at a single point as rivet fasteners do. A
to the cutting, kitting, lay-up and mspection
joint should therefore be designed to ~e
requirements can reduce the waste factor to
concentrations of stress. There are four baSIC
10-15%. Large structures tend to have less
types of stress encountered in structural bond-
waste than small components. The carbon
ing: tensile, shear, cleavage and peel.
epoxy central cylinder for a modem comm.u-
nication satellite has less than 3% matenal
16.6.4 DESIGN CONSIDERATIONS waste, the majority being for localized rein-
forcements. The America's Cup racing yacht
The initial lay-up starts with the preparation
was fabricated with less than 2% material
of the prepreg ply kits. Individual ply kits will
waste.
reduce the overall labor requirements during
the ensuing lay-up. Hybridization o~ materi~s
can be achieved on a ply-by-ply basIS. Off aXIS 16.6.5 GENERAL DESIGN PRACTICE
(± 45°) ply orientations can be prepared utiliz-
• Attention to ply orientation on strength
ing a woven graphite cloth. Interleaf layers of
controlled laminates can prevent matrix
titanium or fiberglass can distribute high load
and stiffness degradation. The 0° ply orien-
inputs into the laminate. .. tation is used to carry the longitudinal
With large composite structures, I.e. rac~g
loading, the 90° ply orientation the trans-
yachts, the prepreg materials are usually dis-
verse loading and the ± 45° ply for shear
pensed directly onto the lay-up. One must
loading.
allow for 'fresh' prepreg, with maximum tack,
• In order to minimize in-plane shear, place
to be applied to the laminate with minimum
the +45° and -45° plies together; the in-
out time of the material during the fabrication
plane shear is carried by the tension and
phase. .. compression in the 45° plies.
One of the first reactions to usmg hand lay-
• To minimize warpage and interlaminar
up fabrication methods for a ~omp?si~~ shear within a laminate, maintain the sym-
structure is: "this will be too labor mtensive .
metry about the center line of the laminate.
This can be misconstrued to mean that all
• Stress concentrations can be minimized by
composite structures done by the hand lay-up
designing tapered or stepped laminate
are expensive. Complex, integral, stiffened
thickness changes.
composite structures may not only be cost
• The placement of specific ply orientations
effective but may not be able to be fabricated
can influence the buckling strength and
by other methods of manufacture.
damage tolerance. The ou~er ply ori~ta­
Compression molding, resin transfer mold-
tions influence the lammate bendmg
ing and sheet molding processes. are cost
characteristics more than plies placed at the
effective for specific types of matenal forms.
neutral axis.
 Applications 375

16.7 APPLICATIONS manufacturing/ design reviews, the decision

to fabricate the central cylinder as a one piece
16.7.1 AEROSPACE
monococque structure was made. One quarter
Two aerospace examples that were manufac- scale models were made as initial feasibility
tured utilizing the hand lay-up and vacuum trials and the finished cylinder was 108 cm
bagging process were an ICBM equipment (42 in) in diameter and 231 em (91 in) high.
section structure (300 parts) and a central With that initial success, a full size manufac-
cylinder (6 parts) for a modem communica- turing development cylinder required 1700
tion satellite. These two primary composite man-hours to fabricate. By production unit
structures illustrate that hand lay-up and vac- five, the fabrication time was reduced to 500
uum bag procedures can be effectively utilized man-hours per cylinder. The final design
on limited production programs. For greater incorporated compression molded graphite
production demand or with a less complex inserts for hard mounts and was co-cured as a
structure other methods (machine lay-up or one piece structure.
filament winding) could have been used. The program set new standards of coopera-
However, neither alternative process could tion between engineering analysis/design,
achieve the results required with materials manufacturing and quality control groups. A
selected. manufacturing plan including all required
inspection points and a detailed fabrication
sequence was prepared. The major problem
ICBM equipment structure
during the fabrication was the out time avail-
The ICBM equipment structure was initially able with the graphite epoxy resin system.
designed to utilize unidirectional With only 10 working days available, the kit
graphite/ epoxy materials because unidirec- preparation, lay-up, compacting cycles and
tional materials were available during the final cure was on a tight time schedule.
design phase. The development of woven This program used various operational pro-
graphite was undertaken in order to reduce cedures to achieve success. Tank inserts were
the manufacturing costs of the program. Initial compression molded from chopped
prototypes (unidirectional tape) required over carboni epoxy material prior to the start of the
2000 man-hours labor to complete. The first laminate schedule on the cylinder. It was
prototype utilizing woven graphite required determined early in the development cycle
only 900 man-hours, better than 50% reduc- that the unidirectional tape material handled
tion in manufacturing costs. Eventually better if it was not precut into kits, but pre-
production labor requirements were reduced pared just prior to application to the tool. A
to less than 200 man-hours. combination of heat debulks and pre-bleeding
was utilized to maintain the desired resin con-
tent, until the supplier could prepare an
Satellite central cylinder
acceptable net resin unidirectional tape. With
The graphite/epoxy central cylinder was a limited facilities to autoclave a large cylinder
unique structure from its inception. The pro- with in-house capability, other methods of
ject was undertaken not just because manufacture were employed. After the com-
composites would be lighter but that the pleted hand lay-up was done, a vacuum
metallic (beryllium) design required a longer debulk was applied to ensure all air was
manufacturing time that was unacceptable. removed and the lay-up was compacted. The
The materials of choice were a 520 GPa cylinder was then wrapped with perforated
(75 msi) pitch graphite fiber, epoxy resin and shrink tape (to allow for resin/air bleed dur-
aluminum honeycomb. During the initial ing cure), then breather cloth was applied
376 Hand lay-up and bag molding

prior to final nylon vacuum bag. Studies were Challenge. A typical America's Cup yacht uti-
conducted on utilizing a silicone rubber vac- lized over 13 006 m2 (140000 ft2) of ply surface
uum bag, but overall program costs and area of unidirectional graphite tape in the hull
difficulties in handling a large (heavy) bag, and deck structure and 9290 m2 (100 000 ft2) of
pushed the utilization of nylon bagging film ply surface area for a one piece graphite mast
for the final vacuum bag. The part was then structure.
oven cured. Due to the overall size and past experience
with the boat builder, a male wood plug was
fabricated with integral heated wires imbed-
16.7.2 MARINE APPLICATIONS
ded to reduce the heat sink effect during the
The fabrication of an America's Cup racing final co-cure of a complete hull laminate. As
yacht (Fig. 16.14) presented another set of the maximum temperature allowed was 90°C
requirements. With limited cure temperature, (183°F), wood was a good choice.
resin matrix and control density requirements, The hull laminate was then applied, start-
the challenge was to be able to hand lay-up ing with the inner skin (as on an America's
and cure a large composite structure to meet Cup racing yacht) then film adhesive and hon-
design requirements. Marine applications had eycomb core was applied. The inner laminate
reached new heights when the new rules went and honeycomb was vacuum bagged and par-
into effect for the 1992 America's Cup tially cured. The honeycomb core was
smoothed and all joints were filled prior to the
application of the outer skin laminate. The
outer laminate and any local reinforcements
were applied to the core and inner laminate
(Fig. 16.15). The completed hull laminate was
then vacuum bagged and oven cured.
The outside of the hull was essentially com-
plete but had a rough surface. The roughness
was greatly dependent upon the care of work-
ers during the lay-up of the outer laminate
skin. With reasonable precautions, prepreg
material can be placed in such a way that there
are no overlaps and all gaps have been filled
with additional fiber. This effort alone can save
hundreds of man-hours during the final fin-
ishing.
This effort saved over 1200 man-hours com-
pared with a 'wet' prepreg that did not
produce an acceptable outer surface finish.
The oven cure achieved maximum mechanical
properties required by the design. Laminate
design of the hull allowed for the use of matrix
butt joints for each strip of material. The over-
all hull structure met all design goals and the
800 kg (1760 lb) weight objectives.
The finishing procedure, essentially the
same as one used today in an auto-body shop,
Fig. 16.14 'Spirit of Australia'. consisted of filling and grinding the surface in
 Applications 377

cost. A discussion with design/analysis and

manufacturing concluded that it was possible
to fabricate a one piece mast within the facili-
ties available and technical expertise within
the syndicate. The design of the tooling was
aimed at providing a capability to cure the car-
bon/epoxy resin within the America's Cup
rules. The requirements specified the cure
temperature 120°C (250 0 P), cure pressure (3
atmospheres), laminate density/modulus of
graphite fiber and overall mast profile. All
these objectives were realized within 100 days,
from start to completion of the first one piece,
35 m (115 ft) in length, America's Cup com-
posite mast. Performance exceeded all
expecta tions.
Additional masts starting with mast num-
ber 3 were manufactured with the final weight
objective of less than 450 kg (990 lb). Total fab-
rication time for each mast, from the start of
material kit preparation to completion of the
cure, required a maximum of 21 days. The
selected carboni epoxy resin system was
workable for this limit. The mast, spinnaker
pole and rudder stock utilized variations of
pressure bag molding with integrally heated
Fig. 16.15 'Spirit of Australia' laminate laydown fiberglass tooling.
(top) 45° ply laminate; (bottom) 0° ply laminate.
These two tooling approaches allowed for
some of the largest one-piece structures to date
two or three stages and culminating with a to be fabricated in the commercial marine mar-
final coat of good grade epoxy or urethane ket. Logistics can become a major role player
marine paint. This step is often postponed in the planning required in that not only the
until the boat is completely assembled and support materials are on hand but that power
ready for fitting out. Then, the hull was is available; materials handling is taken into
inverted and placed in a fitted saddle; the male account and there is sufficient crew available
plug was removed. Additions of bulkheads, to complete the lay-up within the material out
flooring and interior completed the work. time limits.
The manufacture of the carboni epoxy mast As seen in the above examples, the one
for the 'Spirit of Australia' presented some major factor for all the programs was the
unique challenges. The major program chal- material out time. With the development of
lenge was how to obtain a mast that would new resin matrixes, the design of complex
provide the required performance at affordable structures can be achieved.
MATCHED METAL COMPRESSION 17
MOLDING OF POLYMER COMPOSITES
Enamul Haque and Burr (Bud) L. Leach

17.1 INTRODUCTION Reinforcement may take the form of fibers

particles, laminate, flakes and fillers:
In today's highly competitive global economy,
Depending upon the type and orientation of
the need for materials with the right properties
the reinforcement and the manufacturing tech-
to meet the demands of design, environment,
n~logy r~quired to produce them, composites
durability and economics is growing.
With vanous properties and cost can be fabri-
Composite materials, with their high strength
cated. Polymer composites are composites in
and stiffness-to-weight ratios, have many
whi~ the binder or matrix is a polymeric
advantages and are a desirable engineering
m~te~ and the .reinforcement is usually a
material.
thin fibrous matenal. Polymer composites can
. There is no universal definition of compos-
have either a thermosetting or thermoplastic
Ites. In general, a composite material is a
matrix. In this chapter we will discuss ther-
heterogeneous material system consisting of
mosetting matrix based composites.
two or more physically distinct materials. In a
composite material system, the individual Reinforcing fibers may also be of various
~ds, with glass (E-type), carbon, or organic
materials exhibit their unique properties and
fibers (e.g. aramid) being the most common.
the composite as a whole shows properties that
Glass fibers are the most widely used type of
are different from its constituents. In addition
reinforcement since they offer good strength
to the constituents' unique properties, the
and moderately high temperature resistance
properties of composites are also dependent on
(~bout 260°C) at a cost effective price. Glass
the form and structural arrangements of the
fibers also come in various forms. They can be
constituents and the interaction between the
con~uous filaments, cut or chopped strands,
constituents.
rovmg and yarns, or in the form of cloth, mats
Broadly speaking, composites consist of
or tapes. This allows glass fibers to be used in
two components, a binder or matrix and a
a variety of applications such as lay-up, fila-
reinforcement. The matrix functions as the
ment winding, matched die-molding, etc.
body constituent, serving to bind the rein-
In this chapter we discuss the matched metal
~orcement together and giving the composite
compression molding of thermoset based poly-
Its bulk form. The reinforcements are the struc-
mer composites. Matched metal compression
tural.constituents, providing high strength to
molding is a molding process in which the cure
the mternal structure of the composite.
is obtained while the material is restricted
between two mold surfaces and the loading
and closing of the mold causes the material to
Handbook of Composites. Edited by S.T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 0412540207 conform to the desired configuration. This
 Background 379

process enables large scale production of large 17.2.1 BULK MOLDING COMPOUND (BMC)
surface area parts with contour problems and
tight tolerances. Matched metal compression BMC has been defined as 'a fiber reinforced
molding employs a 'mold' or match dies. The thermoset molding compound not requiring
male mold is matched to the female mold so advancement of cure, drying of volatile, or other
that when the dies are closed, a controlled space processing after mixing to make it ready for use
results. A preform charge is placed on the core at the molding press'2. BMC can be molded
and the cavity is pressed against it, applying without reaction byproducts under only enough
direct pressure on the material. The pressure in pressure to flow and compact the material. BMC
this type of molding varies from 1.38 to 6.895 is usually manufactured by combining all the
MPa (200 to 1000 psi) and curing temperatures ingredients in an intensive mixing process.
from 125°C to 160°C (260°F to 320°F). Recent advances in BMC technology dictate that
both the dry ingredients and wet ingredients be
batch mixed separately and then combined
17.2 BACKGROUND
together in an intensive mixer. The BMC is usu-
Advanced polymer composites are now being ally in a fibrous putty form when it comes out of
applied extensively for all types of applications the mixer and resembles 'sauerkraut'. It is usu-
in the industrial and automotive markets. Table ally compacted and extruded into bars or 'logs'
17.11 shows the usage of composites in various of simple cross section.
markets during 1991-1993. This section deals The earliest BMCs were probably made
primarily with thermoset polymer composites about 1950, employing a process of impregnat-
used in matched metal compression molding. ing roving strands with blend of resin, filler, etc.
The two most popular reinforced molding com- and chopping them to a length in the wet stage.
pounds used in the plastics industry are Premix Since wetting glass fibers with a resin contain-
or BMC and SMC (also referred to in modified ing much filler is difficult and slow, these
versions as HMC and XMC). Low Pressure premixes had a high glass content. The first
Molding Compounds (LPMC), ZMC and TMC high volume commercial BMC was made with
are also becoming popular. sisal fibers and used in molding automobile

Table 17.1 US Composites shipments: 1991-1993"

Millions of pounds
Markets 1991 1992 1991-1992 1993 1992-1993
% change (projected) % change
Aircraft/ aerospace/military 38.7 32.3 -16.5 26.0 -1.95
Appliance/business equipment 135.2 143.2 +5.9 146.7 +2.4
Construction 420.0 483.0 +15.0 522.0 +8.1
Consumer products 148.7 162.2 +9.1 164.1 +1.2
Corrosion-resistant equipment 355.0 332.3 -6.4 336.8 +1.4
Electrical! electronic 231.1 260.0 +12.5 273.0 +5.0
Marine 275.0 304.4 +10.7 317.2 +4.2
Transportation 682.2 750.0 +9.9 810.0 +8.0
Other 73.8 83.4 +13.0 88.0 +5.5
Total 2359.7 2550.8 +8.1 2683.8 +5.2

" Includes reinforced thermoset and thermoplastic resin composites, reinforcements and fillers. Source: SPI Composites
Institute
380 Matched metal compression molding of polymer composites

heater housings. Improvement in the binder 17.2.3 THICK MOLDING COMPOUND (TMC)
chemistry of glass fibers, development of a
TMC was developed by Takeda Chemical
chemical thickening system and thermoplastic
Industries, Ltd (Osaka, Japan). TMC is suited
low profile additives help BMC to attain
to compression, injection and transfer mold-
strength, chemical resistance and to overcome
ing and is usually processed on the same
surface irregularities. Consequently, BMC was
equipment as SMC and BMC materials. TMC
accepted for use in the electrical, chemical and
composites are usually produced up to 51 mm
appliance industries.
(2 in) thick and glass fiber length can vary
Today, BMCs are accepted as high perfor-
from 6.4 to 50.8 mm (0.25 to 2 in). In TMC,
mance engineering thermoset molding
continuous impregnation and high sheet
compounds and used extensively in the elec-
weight result in complete wet-out of resins,
trical, automotive and consumer goods
fillers and reinforcing fibers. Better wet-out
industries. BMC is increasingly injection
results in improved mechanical properties
molded to take advantage of the automation
and reduced porosity. TMC is usually used in
and reproducibility afforded by the process,
business machine housings, appliance com-
although it is also both transfer molded and
ponents and other consumer related
compression molded.
industries.

17.2.2 Z MOLDING COMPOUND (ZMC) 17.2.4 SHEET MOLDING COMPOUND (SMC)

ZMC was developed in 1979 in France to SMC is a type of fiber reinforced plastic which
improve BMC performance. BMC suffers primarily consists of a thermosetting resin,
from glass fiber degradation during injection glass fiber reinforcement and filler. Additional
molding and ZMC was developed to keep ingredients such as low-profile additives, cure
shear forces as low as possible during mold- initiators, thickeners and mold release agents
ingS. A special type of injection molding are used to enhance the performance or pro-
machine developed by Billion in France com- cessing of the material4•
bines the advantages of both a screw machine The development of SMC started in the
and a plunger machine. The ZMC injection early 1950s after the finding that the viscosity
machine uses a screw to homogenize and pre- of unsaturated polyester resins increases with
cisely measure the shot. The injection is made the addition of Group IIA metallic oxides,
like a plunger by the displacement of the hydroxides, or carbonates5• The first pub-
screw and inner barrel inside the main barrel. lished report on SMC was presented at the
In a ZMC, the different components are Cleveland Section of the Society of Plastics
mixed in conventional mixers like BMC. The Engineers meeting. The report involved work
compound viscosity is usually low and done in Germany using fiberglass mat
adapted to injection machine characteristics. impregnated with a resin mixture containing
The design of the mold plays a key role in the magnesium oxide6• At the same time a num-
processing phase and ZMC parts cannot be ber of US patents7,s were published on the use
successfully made unless part design and of Group II metal oxides, hydroxides, or car-
mold design are combined upfront. bonates for use on adhesives. The early
Compared to SMC, ZMC parts have lower applications of SMC materials were in electri-
mechanical properties, but higher perfor- cal and industrial goods. During the next two
mance when compared to conventional decades, growth in commercial usage of SMC
injection molded BMC. followed the evolution of continuously
improving equipment, low profile additive,
 Background 381

catalyst, etc. The automobile industry star~ed 17.2.5 LOW PRESSURE MOLDING COMPOUND
using SMC in the early 1970s for producmg (LPMC)
exterior body components, such as hoods or
Low Pressure Molding Compound (LPMC) is
grille opening panels. With the introduction an SMC type material which can be molded at
of high strength SMCs in the mid-1970s, usage 1.38-2.07 MPa (200 to 300 psi) instead of 5.52
of SMC increased to structural components. -6.90 MPa (800-1000 psi) required for stan-
SMC is currently used extensively in trans- dard SMC
portation, construction (door panels),
LPMC is made by replacing the chemical
appliances (washing machine door, refrigera-
thickening mechanism of alkaline earth
tor housing), furniture (chair, tabletop) and
oxides (Group II) with a physical thickening
business machines (computer housings). The mechanism utilizing a crystic polyester. The
transportation industry has the highest level material is heated to melt the crystic and then
of consumption of SMC For instance, in the the other ingredients are added, mixed
North American market alone, the annual rate together and run through a modified SMC
of consumption exceeds 100 million kg9 •
machine maintaining the elevated tempera-
Details of SMC manufacturing are available ture.
in the literatures. SMC offers many advantages
The thickening occurs as the material cools
which include variety, part consolidation, to ambient temperature and the compound is
lightweight and dimensional stability.
ready to mold at that time. Cooling rolls speed
With the evolution of flexible backbone the cooling process and thus the material can
polyester resin systems and dev~lopment. of be molded right off the SMC machine without
special additives, flexible SMC IS becommg waiting for the 24-48 h thickening of standard
very popular and is now competing with
SMC LPMC allows the molder to use lower
thermoplastics for vertical body applications. tonnage presses to mold larger parts and use
Special applications SMC is also becoming
less steel in building the tools as they do not
popular. With the addition of hollow micros-
have to deal with high pressures and corre-
phere glass bubbles in a. standard SM~ sponding forces. The shelf life of LPMC is
formulation, lower denSIty (1.3-1.4) IS
much longer than SMC and the physical prop-
obtained for weight reduction. High strength
erties are comparable.
molding compound (HMC) is a SMC contain-
ing 65% chopped glass fiber instead of the
usual 25-35%. HMC uses little or no filler and 17.2.6 CONTINUOUS IMPREGNATED
can be compounded on a standard SMC COMPOUND (CIC)
machine. Directionally reinforced molding
compound (XMC) is a directionally oriented In 1986, continuous impregnated compound
moldable resin-glass fiber sheet containing (CIC) was developed in Germany. This is sim-
65-75% continuous reinforcement. XMC is ilar to TMC Like TMC, the impregnation is
also usually compounded on standard fila- made between two rolls but the compound is
ment winding equipment and has strength removed by doctor blades and carried by a
five times greater than SMC Unidirectional screw or plunger to boxes or drums. CIC is
molding compound (UMC) is a system of usually injection molded, but can also be injec-
chopped and continuous fibers produced on tion/ compression molded. Properties are
a modified SMC machine. An advantage of comparable to BMC but processing is easier
UMC is that different varieties of fibers can than BMC Modified CIC is also known as
be used. KMC (Kneaded Molding Compound).
382 Matched metal compression molding of polymer composites

17.3 FORMULATIONS butadiene copolymers. Extensive details of LPA

mechanism are published in the literature9,18,19.
Polymer composites have the unique and dis-
tinct advantage in that their properties can be
tail?re? to meet different applications by 17.3.3 INITIATORS AND INIDBITORS
deslgrung the formulations. The major compo-
~tiators are used to initiate the curing reac-
nents of polymer composites used in matched
metal compression molding are resin, low pro- tion at elevated temperatures. Composites
file additive, fiber, filler, initiator, inhibitor, are polymerized or crosslinked by a free rad-
internal mold release agent and other addi- ical mechanism in which the double bond of
the polyester chain reacts with the vinyl
tives (e.g. viscosity reducer, toughness
~o~omer (u.sually styrene). This copolymer-
enhancer, etc.).
lzation reaction provides a three-dimensional
network that converts the viscous liquid resin
17.3.1 RESIN paste into a hard thermoset solid. Initiators
d~compose at elevated temperature and pro-
Unsaturated polyesters and vinyl esters are the
principal resins used in polymer composites for vlde a source of free radicals to initiate the
compression molding. Epoxies are also used for copolymerization reaction. Peroxyesters and
peroxyketals are the most common classes of
specialty products which require longer cure
cycles and higher strength. Phenolics are being peroxides.
used for formulating composites, especially Inhibitors are added in small quantities to
p~olong shelf life, modify cure rate and mag-
SMC, in applications which require lower flam-
ru~de of exotherm to prevent cracking of
mability, reduced smoke generation and higher
thermal stabilitylo. Details of resin chemistry thick molded sections. Inhibitors are also
are available in the literaturell- 14• Styrene is used to improve resin stability. Two general
classes of inhibitors are commonly used, sub-
commonly used for cross-linking of both poly-
stituted phenolic derivatives and the
ester and vinyl ester resins. Low styrene
qu~ternary ammonium salts (e.g. hydro-
polyesterlS is becoming popular due to strin-
qumone, p-benzoquinone, etc). An excellent
gent EPA requirements on styrene vapors. New
resin technology is also being considered for review on initiator and inhibitor chemistry is
available elsewhere 9.
compression molding. They include hybrids of
unsaturated polyester and urethane16, acryles-
terol resin with polyisocyanate17, etc. 17.3.4 FILLERS
Fillers are used to improve physical properties,
17.3.2 LOW PROFILE ADDITIVES reduce volumetric shrinkage of the resin and to
reduce costs. Fillers are typically divided into
Low profile additives are thermoplastics that
functional and non-functional categories.
are added to the formulation in 2-5% (by
Examples of functional fillers include alumina
we~ght) of the final product or 10-20% (by
trihydrate for flame retardancy, hollow glass
welght) of the organic portion of the formula-
bubbles for lower weights, mica and wollas-
tion to control the shrinkage of the cured
composites. Typical thermoplastics include tonite for reinforcement. Non-functional fillers
are used for cost reduction and are mineral
polyvinyl acetates, poly methyl methacrylate
based. Ground limestone (CaC03) is the most
and copolymers with other acrylate, vinyl chlo-
ride-vinyl acetate copolymers, polyurethane, common type of filler. An excellent review is
available on filler use in composites9,20.
polystyrene, polycaprolactone, cellulose acetate
butyrate, saturated polyester and styrene-
 Molding 383

17.3.5 FIBERS 17.3.8 OTHER ADDITIVES

Glass fiber reinforcement is used to achieve Pigments are added to produce color in the
necessary dimensional stability and mechani- molded part. Common pigments include cad-
cal properties. E-glass is the most common mium salts, carbon black, titanium dioxide,
fiber reinforcement for composites. iron oxides, organic dyes and pigments, etc.
Depending on the binder chemistry and Various types of viscosity reducers are used to
amount, glass fibers are classified as hard or lower the viscosity of the paste to increase
soft type. Other types of fibers include carbon, filler loading and glass wet-out. Other addi-
aramid (Kevlar), 5-2 glass, etc. Glass loading tives include various elastomeric additives
normally averages 30% by weight in compres- (e.g. Hycar, Kraton, etc.) to increase toughness.
sion molded composites, but can vary from Tables 17.3 to 17.6 show typical formula-
18--65% by weight. Table 17.221 shows typical tions for BMC, 5MC, ZMC and LPMC.
fibers used in polymer composites.
17.4 MOLDING
17.3.6 INTERNAL MOLD RELEASE Matched metal compression molding is one of
Internal mold release agents are added to facil- the oldest manufacturing techniques in the
itate part ejection from the mold. Major types plastics/ composites industry. The recent
of release agents include metallic stearates, development of high strength, fast cure,
fatty acids, fatty acid amides and esters and
Table 17.3 BMC formulation
hydrocarbon waxes. Zinc stearate and calcium
stearate are the most widely used internal PHR
mold release agents in 5MC and BMC.
Polyester resin 60.0
Low profile additive 40.0
Styrene 5.0
17.3.7 THICKENERS
Initiator 1.5
The addition of Group II oxides and hydrox- Inhibitor Trace amount
ides to carboxy-terminated unsaturated Mold release 4.0
Pigment 0.25
polyester / vinyl ester resin increases its viscos-
Thickening agent 1.0
ity. Magnesium oxide, magnesium hydroxide, Filler 50-200
calcium oxide, calcium hydroxide or combina- Compound
tions of those materials are the most popular Glass fiber 10-25%
thickeners. Paste 75-90%

Table 17.2 Glass fibers used in composites21

Fiber Specific Tensile Tensile Tensile Coefficient

gravity strength modulus failure of thermal expansion
(CPa) (CPa) strain (%) (x W-~;oC)

E-glass 2.54 3.45 72.4 4.8 5

S2-glass 2.48 4.30 86.9 5.0 2.9
Carbon 1.76-2.15 1.5-5.6 220-690 0.3-1.2 -{).1 to -1.2 (longitudinal)
(graphite) 7-12 (radial)
Kevlar49 1.45 3.62 131 2.8 -2 (longitudinal)
59 (radial)
384 Matched metal compression molding of polymer composites

Table 17.4 SMC formulation physical and mechanical properties can be

obtained in compression molded parts. Figure
PHR 17.1 shows a schematic of a compression
Polyester resin 55.0 molding process.
Low profile additive 40.0 This section addresses the compression
Styrene 5.0 molding of composite parts using SMC. BMC
Initiator 1.5 molding is similar except for the charge prepa-
Inhibitor 250 ppm ration step. The compression molding process
Mold release 4.0
Pigment 1.0
can be divided into four distinct steps.
Thickening agent 2.0
Filler 150-250
Compound Heat and pressure
Glass fiber (25.4 rom) 25-30%
Paste 70-75%
Cavity
Table 17.5 LPMC formulation
Charge
PHR
Polyester resin 65.0
Crystic 15.0
Styrene 5.0
Initiator 1.2
Mold release 5.0
Pigment 1.2
Filler 220
Compound
Glass fiber 25-30%
Paste 70-75% Heat and pressure

Table 17.6 ZMC formulation Fig. 17.1 Schematic of a compression molding

process.
PHR
Polyester resin 65.0
Low profile additive 40.0
Styrene 5.0
17.4.1 CHARGE PREPARATION AND
Initiator 1.5
Inhibitor 100 ppm PLACEMENT
Mold release 4.0 When the SMC has reached its desired molding
Filler 220
viscosity, pieces of SMC are cut to pre-specified
Compound
Glass fiber 15-25% size after removing the carrier films. The SMC
Paste 75-85% is cut using slitters, pizza type or guillotine type
cutters. Several pieces of SMC plied together
form the 'charge'. The charge pattern/ply
SMC/BMC and advancement in press tech- dimensions are chosen so as to cover 20-80% of
nology is making the compression molding the mold surface area. The charge pattern and
process very popular for mass production of placement on the mold determines the quality
composite parts. In comparison with the injec- of the molded parts, since it influences the
tion molding process, in general, better length of flow in the mold, fiber orientation,
 Properties 385

flowline and other surface defects. In order to entire surface. After IMC injection, the press is
reduce cycle time, sometimes the charge is pre- closed and the curing operation is repeated at
heated to a temperature below gel point using or above the SMC molding pressure.
infra-red or dielectric heaters. Sometimes the IMC is injected at high pressure
without mold opening and closing prior to
17.4.2 MOLD CLOSING AND FILLING complete cure of the SMC charge.

After proper placement of the charge in the

core of the mold, the cavity is quickly closed to 17.4.4 PART EJECTION AND POST-CURE
contact the top surface of the charge. The cav-
At the end of the molding cycle, the mold is
ity is then closed at a slower rate, usually 4-12
opened and the part is ejected from the core
mml s. In most cases the mold is heated to (for
with the use of integral ejector pins and
example) 150°C, which causes the charge vis-
allowed to cool to ambient temperature. Hot
cosity to be reduced. With increasing mold
parts are handled carefully and are usually
pressure as the mold is closed, the charge
placed on a support racks to cool to ambient
flows towards the cavity extremities, forcing
temperature. As the part cools outside the
air out of the cavity. The mold closing speed is
mold, it continues to cure and shrink which
very important as it induces gelation of the top
creates residual stresses due to differential
charge surface if the closing speed is slow or it
cooling at various sections in the part. After
causes trapped air if closing speed is fast. The
the part is placed on support rack, it is
filling stage is usually completed in 0.5-20 S5.
deflashed while still hot and stored in racks for
Vacuum molding is increasingly being used
secondary operations like punching, drilling,
during charge flow to reduce surface porosity
bonding, etc.
and air entrapment in the part. Vacuum level
The compression molding process is com-
is usually in the range of 7-9 x 1()4 Pa (21-27 in
plex and there are several important variables
Hg). The molding pressure based on projected
that influence molding. Compression molding
part area ranges from 1 to 10 MPa (100 to
may also produce a variety of surface and
1200 psi). Higher molding pressure causes
internal defects which can be eliminated by
sink marks, while lower pressure cause scum-
proper material selection, part design and
ming of the mold and porosity.
molding technique. Details of the molding
variables and the source and remedies of
17.4.3 CURING major molding defects are available in the lit-
After filling, the charge remains in the hot erature5,9.
mold for the crosslinking reaction to be com-
pleted. The curing time is usually between 25 s
17.5 PROPERTIES
to 3 min, but depends on several factors,
including resin-initiator-inhibitor reactivity, The properties of a polymer composite can be
part thickness and mold temperature. tailored, within limitations, to meet different
Sometimes in class A or appearance grade applications by designing its formulation. This
parts, in-mold coating (IMC) is used to unique characteristic of polymer composites
enhance the surface of a molded part. The makes definition of detailed properties diffi-
most common method of IMC injection cult. The properties are usually used for
requires opening the mold by a small amount information and guidelines for preliminary
(0.2-0.5 mm) after the curing cycle. IMC is part design, material selection and to under-
usually a coating of polyester or stand the effect of formulation variables on
polyester-urethane hybrid which covers the mechanical properties.
386 Matched metal compression molding of polymer composites

17.5.1 STATIC PROPERTIES 17.5.3 OTHER PROPERTIES

Table 17.7 shows the static and impact proper- Several other tests are now being performed to
ties of SMC, BMC, ZMC and LPMC. In correlate properties with operation conditions.
general, tensile and flexural properties are The dynamic mechanical analyzer (DMA) is
routinely measured and are presented here. used to measure complex and storage modulus
Compressive and shear properties are mea- at various temperatures and frequency ranges.
sured only for use in special applications. The The effect of environmental conditions on var-
static properties of SMC and BMC are highly ious properties is tested to simulate end-use
dependent on the fiber content, length, type environment. Creep and stress relaxation tests
and orientation. Tensile strength increases sig- are also done on SMC/BMC for use in struc-
nificantly with increasing fiber content; tural applications. Electrical properties are also
however, the tensile modulus is affected only important, permitting BMC to be used in elec-
moderately. Increasing the length of chopped trical applications. Arc resistance is important
fiber increases the tensile strength, but has no and dielectric strength, dielectric constant, dis-
effect on the modulus. Glass fiber type (E-glass sipation factor, etc. are also measured.
or 5-glass) has a significant effect on both the
tensile strength and modulus. The resin chem- 17.6 APPLICATIONS
istry also influences tensile properties at low
Reinforced composites materials offer the max-
fiber content. In general, flexural and com-
imum design versatility and capability of any
pressive properties follow the same trend as
material. With the excellent cost/performance
the tensile properties. Flexural strength is
characteristics of reinforced composites, the
always higher than tensile strength, though
variety and quantity of products being pro-
the modulus may be comparable.
duced with these materials grow annually
around the world. Matched metal molded rein-
17.5.2 FATIGUE PROPERTIES forced composites should be considered when
the finished product can be enhanced by one or
The fatigue properties of SMC and BMC are
more of the following characteristics.
usually based on tensile cyclic loading of
unnotched specimen. A typical 5-N diagram is Part consolidation Reinforced composites can
shown in Fig. 17.221. In general, the fatigue be molded in three dimensions in one opera-
strength increases with increasing fiber con- tion. Complex shapes that require multi-piece
tent and there is no fatigue limit, unlike low assembly using materials, such as wood or
carbon steel. Details of such testing are pub- steel, may be molded in one step with the use
lished elsewhere21,22. of ribs, bosses and varying wall thickness.

Table 17.7 Static and impact properties'

Property Tensile Tensile Flexural Flexural IZOD Specific Coefficient

strength modulus strength modulus impact gravity of thermal
(MPa) (CPa) (MPa) (CPa) (unnotched) expansion
(x lO--6!,C)

SMC 65-100 9.5-14 130-200 8-14 600--1200 1.3-2.0 8-14

BMC 30-70 8-12 50--150 9-17 100--700 1.7-2.1 15-20
ZMC 30-70 8.5-12.5 50--150 7-12 200--500 1.8-2.0 11-27
LPMC 65-100 9.5-14 120--200 8-14 600--1200 1.8-2.0 7-10

• Published industry data

 Applications 387

100
R =0.05

80

60

40

20

o~----~------~------~----~~----~------~----~
0.1 1 10
Number of cycles, N

Fig. 17.2 Typical fatigue S-N diagram for SMC (21) .6.: at -4Q°C; 0: at 23°C; 0: at 93°C. [Reproduced from
Composite Materials Technology: Processes and Properties (ed P.K. Mallick and S. Newman) by permission of
the publisher.}

Light weight Reinforced composites offer a from most organic chemicals and can be for-
greater strength-to-weight ratio than most mulated to resist acidic and basic solutions.
non-reinforced plastics and many metals.
Electrical resistance Reinforced composites
Dimensional stability Reinforced composites are very poor conductors of electricity. As
can maintain the critical tolerances required of such, they have a high dielectric strength for
the most demanding applications. Composites application in the electrical and electronic
meet the most stringent material stiffness, industries23.
dimensional tolerance, weight and cost criteria
Resistance to minor impact Reinforced com-
in many diverse applications.
posite components have an excellent memory
High strength Reinforced composites have characteristic. Instead of yielding or deforming
excellent strength-to-weight properties. By under minor impact as with steel, a reinforced
weight, reinforced composites surpass the ten- composite panel will deflect and spring back to
sile strength of iron, carbon and stainless its original surface form4 (Fig. 17.3).
steels. Many glass reinforced compounds
Surface quality Reinforced composites can
equal or exceed the flexural strength and
achieve a variety of surface textures, from very
impact resistance of metals23 •
smooth and glossy to a rough texture.
Corrosion resistance Reinforced composites Insignias and alphanumeric characters can be
do not rust or corrode, are resistant to attack molded as raised or indented characters.
388 Matched metal compression molding of polymer composites

and performance requirements of the product

or component. The designer must:
1. establish size and shape limitations based
on:
• basic end use function;
• aesthetics and marketing;
Steel • shipping limitations;
• weight requirements;
• strength and stiffness requirements;
• flexibility requirements or limits;
• process limitations.
2. establish the structural requirements based
on:

ZZZZZ
• various loads that will be impacted to the
part including weight, pressure and
dynamic loads;
Composite
• duration of the loads on the part;
Fig. 17.3 Minor impact. • temperature variations on the part and
surface;
• number of cycles of temperature change;
Molded-in color Color can be added to the
• liquid, moisture and vapor resistance
reinforced composite compound, often elimi-
requirements;
nating the need for a secondary painting
• relative significance of strength-to-
process.
weight ratios.
Recycling Most reinforced composites can be
3. establish the non-structural requirements
recycled either by regrinding or pyrolysis. based on:
Reground material can be used as filler or rein-
• corrosion, weathering, moisture and
forcing material. Pyrolysis reduces the temperature resistance;
composite into its basic components by heat-
• moisture and vapor penetration for con-
ing the material in the absence of oxygen. The
densation protection;
process yields gas, oil and solid by-products
• fire safety relative to combustibility;
that can be recycled back into composites, or
• flame-spread rate requirements;
used in building and agriculture materials 4.
• light transmission (transparency, translu-
Thousands of products are molded each year
cency and opaqueness);
utilizing reinforced composites: aerospace,
• surface textures, both aesthetic and func-
automotive parts, sports and recreational
tional;
equipment, boats and business machines to
• surface coatings for protection or aesthet-
name a few. This wide variety of applications is
ics;
indicative of the versatility, capability and cost
• thermal insulation;
effectiveness of reinforced composites.
• noise and sound control;
• dielectric requirements for electrical
17.7 DESIGN CONSIDERATIONS insulation24 •
Given the wide range of options provided by With the establishment of the functional and
reinforced composites, it is imperative that the performance requirements the product design
designer accurately establish the functional can be developed.
 Design considerations 389

There are some general design principles

which can assist in the development of struc-
turally efficient configurations for reinforced
composite components.

17.7.1 SHELL AND PLATE CONSTRUCTION

These are the most common configurations of

reinforced composite parts. A reinforced com-
posite component is constructed from layers of
reinforced composite materials molded into
the shape desired creating a geometric 'shell'.
It is good practice to design so that only forces
that place a part in tension or compression are Rectangu la r

applied to any component. Compound curve

shapes provide good transmission of uniform Fig. 17.5 Corrugated configuration.
loads into tensile and compressive forces
within a part. can degrade 'appearance' surfaces (Fig. 17.6).
Ribbed configurations are often used to Minimum draft angle of 10 for the first
achieve required strength and stiffness in 76.2 mm (3 in) of depth, 2 0 for 76.2-101.6 mm
structural components (Fig. 17.4). (3-4 in) of depth, 3 0 for 101.6-127 mm (4-6 in)
Corrugated or open ribbed configurations of depth and 10 for every additional 50.8 mm
are used to achieve needed structural depth (2 in) thereafter is recommended on all sur-
while efficiently using materials and fabrica- faces parallel to the mold movement. This
tion processes (Fig. 17.5). pertains to all part details, such as ribs, bosses,
elevation changes and holes. A draft angle of
lOon standing ribs and bosses will yield a
17.7.2 DRAFT
thickness change of 0.43 mm (0.017 in) per
Draft is a slight angle introduced relative to the inch per side24 •
direction of the opening and closing of the 'Zero draft' may be obtained by designing
mold. It is necessary to design the part so that the mold in such a way so the draft-free sur-
all side walls, both interior and exterior, have face lies at an angle to the mold direction. This
draft. This enables the part to be removed from will affect the positioning of bosses, ribs and
the mold without hanging or rubbing, which other details of the part.

Transverse
Rib

Molded Longitudinal Rib

Fig. 17.4 Rib configuration.

390 Matched metal compression molding of polymer composites

Draft angle
(10 recommended)

,/,.- Minimum
Recommended
Radius

Direction of Fig. 17.8 Minimum draft.

mold movement

Fig. 17.6 Draft.

17.7.3 RADIUS 17.7.4 NOMINAL THICKNESS

In mold making, the radius defines the curva- The nominal thickness is the overall design
ture established between two intersecting thickness of most of the part. It is desirable to
surfaces. The more generous the radius, the establish uniform thickness throughout a part,
better the flow of molding material for a to achieve minimum cure time, uniform cool-
stronger part (Fig. 17.7). ing and minimize warpage and shrinkage
A minimum radius of 1.59 mm (1/16 in) is (Fig. 17.9).
recommended for all radii for both interior Nominal thickness for reinforced compos-
and exterior plane intersections. Radii should ites is 2.54-4.57 mm (0.100-0.180 in).
be designed to maintain relatively uniform Recommended minimum thickness is 1.53 mm
part thickness (Fig. 17.8). (0.060 in). Recommended maximum thickness
Ribs and bosses opposite an appearance 25.4 mm (1.00 in).
surface should have the radii eliminated to By designing hollow ribs, bosses and eleva-
reduce the likelihood of warpage or 'sink' tion changes can achieve intricate part
(surface depression). geometry while maintaining nominal thick-
ness throughout the part.

17.7.5 EDGE STIFFENING

Edge stiffening is a design characteristic

applied to unsupported edges to prevent
warping or bowing. Edge turning is preferable
to edge thickening due to the possibility of
porosity at the edge of part caused by the lack
of molding pressure on the thick area (Fig.
Fig. 17.7 Outside radii.
17.10).
 Design considerations 391

Fig. 17.9 Nominal thickness. Uniform thickness promotes uniform flow and curing and minimizes the risk
of warpage, distortion and telegraphing at thickness changes through the surface.

17.7.6 RIBS for draft angles. They should be dimensioned

so that their thickness at the juncture of the rib
Linear projections 90° from the plane surface
with its plane surface is between 75 and 90% of
of a part are called ribs. The use of ribs will
nominal (Figs 17.11 and 17.12).
allow the part to meet strength and rigidity
requirements, preventing warpage and bow-
ing in large plane surfaces while reducing the
bulk and mass of a part.
Ribs should be designed to maintain the
nominal thickness and follow the guidelines

~ 0.5° draft

--.1 l~l.5mm
(0.06") min
Fig. 17.11 Rib geometry for class 'N surfaces .

................,............,.....,.~i
•
T

Fig. 17.10 Edge stiffening (a) Preferred edge flange

designs to increase panel stiffness; (b) Thickening Fig. 17.12 Rib geometry for non-appearance
the edge flange may increase cycle times. surfaces.
392 Matched metal compression molding of polymer composites

17.7.7 BOSSES 17.7.9 MOLDED-IN TIfREADS

Projections from a plane surface of a part, It is difficult to mold a thread into reinforced
called bosses, provide attachment and support composites and requires highly sophisticated
for related components. They may be solid, and costly molds and molding procedures.
hollow or have molded in inserts. They should Molded threads should be rounded rather than
also follow the guidelines for draft angles and sharp. Rounded threads will resist chipping
nominal thickness (Fig. 17.13). and cracking and will also facilitate flow of
molding material into all areas of the thread.
Molded threads are usually preferred over
inserts if the threaded hole diameter is over
12.7 mm (0.5 in), unless the thread is to be sub-
jected to continual fastening and unfastening24•

17.7.10 MOLDED SURFACES

Surfaces exhibited by the part as it comes from

the mold have not been subject to any post-
molding operation other than the removal of
Nomlnallhickness should be
malnlamed Ihroughoul pari
I flash. The surface of the mold will reproduce
itself as the surface of the part. Many different
textures can be produced on the surface of a
part. High gloss surfaces can be produced by
highly polished molds. Draft is critical when
parts are to be textured on a vertical wall. For
every 0.254 mm (0.001 in) of texture depth,
draft must be increased by 1°.
Raised or indented characters can be
molded into the part. The characters should be
rounded and smooth and positioned on the
surfaces parallel to the parting line of the
mold.

Fig. 17.13 Boss design. 17.8 TOOLING

A good set of matched metal chromed steel

17.7.8 INSERTS tools is required for the optimum conditions
when one is molding reinforced composites.
Inserts are objects (usually metal) which are
Anticipated production quantities expected
molded into a part to facilitate repetitive fas- from the mold or the product end use, or both,
tening and unfastening of associated parts and should dictate the choice of steel as shown in
can be provided with male or female threads. Table 17.85•
Inserts can provide bearing or bushing sur-
faces, electrical or other mechanical
connections. Inserts should have knurls, 17.8.1 MOLD STRESSES
grooves or shoulders to lock them in place and
It is important to consider the stresses created
should be located parallel to the direction of
by the flow of material at typical molding
mold travel.
 Tooling 393

pressures from 4.13 to 8.37 MPa (600 to 1200 forced composite parts should be designed so
psi). Due to unbalanced flow, narrow mold the height of any projecting mold section does
sections that project from the mold surface not exceed two times the width of its base.
could bend or break under such stresses. To Angular sections must not be less than 3004
ensure sufficient strength in the mold, rein- (Figs 17.14 and 17.15).

Table 17.8 Mold steel selectionS

A. Production planning volumes
Type of steel
Planning volumes Core Cavity
5000-20000 parts/y AISIa-1045 steel AISI-1045 steel
20 000-30 000 parts / y AISI-4140 forged steel AISI-4140 forged steel
prehardened to Rockwell C prehardened to Rockwell C
of 28-32 of 28-32
Over 30 000 parts/y AISI-4140 forged steel P-20 forged steel prehardened
prehardened to Rockwell C to Rockwell C of 28-32
of 28-32
100000 parts or less AISI-1045 steel AISI-1045 steel
100000-200000 parts AISI-4140 forged steel AISI-4140 forged steel
for mold life prehardened to Rockwell C prehardened to Rockwell C of
of 28-32 28-32
Over 200 000 parts AISI-4140 forged steel P-20 forged steel prehardened
during mold life prehardened Rockwell C to Rockwell C of 28-32
of 28-32
B. Product end use
Structural items where AISI-1045 AISI-1045 steel
surface appearance is
not critical, such as
reinforcing panels,
truck front ends, etc.,
where molded surface
quality is of secondary
importance.
High-quality surface AISI-4140 P-20 forged steel
appearance decorative
items, such as grille
opening panels, head
lamp surroundings,
quarter wheel opening
covers, etc. where a
high degree of polish
is required on the outer
part of cavity surface.

• American Iron and Steel Institute

394 Matched metal compression molding of polymer composites

Mold section
Weak mold
, , - toonarrow
/ ' " ection

Fig. 17.14 Projecting mold section; any projecting Fig. 17.15 Angular mold section: should not be
mold section should not exceed two times the <30°.
width of its base.

17.8.2 SHEARS
building and finishing allowances 4 Fig. 17.16).
All matched metal compression molds use Knife edge shears are to be avoided as they
telescoping shear edges around the perimeter create a thin mold section which can bend or
of the part. The shear of the mold halves is break under molding pressures (Fig. 17.17).
never in contact but bypass each other as the Shear edges should be flame hardened to a
mold closes, leaving a thin amount of flash Rockwell C of 55-60.
(0.154-0.254 mm, 0.006-0.01 in). The bypass-
ing feature allows the mold cavity to be fully
17.8.3 HEEL BLOCKS
filled regardless of small variations in charge
weight. Compression molds should have heel blocks
A minimum of 3° of draft is preferred for and wear plates suitable to withstand all lat-
return flanges. A minimum of 1.0 mm (0.04 in) eral forces at 12.41 MPa (1800 psi) molding
nominal flat clearance should be provided to pressure. The heel blocks should be an integral
keep the cavity and core halves of the mold part of the mold and flame hardened to a
from contacting each other. A surface normal to Rockwell C of 50-55. Bronze wear plates
die draw should be provided at the edge of the should be bolted opposite the heel blocks and
part. A nominal angular tolerance of ±20° and a have a minimum of 3.18 mm (1/8 in) chamfer
1.5 mm radius should be allowed for mold lead in to avoid shearing off the wear plate.
 Acknowledgements 395

r
,
1.0mm (0.04")
min flat

t
....... 0.15mm (0.006")
flash

Fig. 17.16 Shear edge design; part design requirements to maintain sufficient tool strength at the shear
edges.

17.8.5 MOLD STOPS

Mold stops should be provided to control the

mold's vertical travel. They should be flat and
oil-hardened steel to a Rockwell C of 55-60.
The minimum part thickness can be controlled
Fig. 17.17 Knife-edge shear should be avoided. It
by the mold stops.
can be eliminated by adding a minimum flange of
5 mm (0.2 in) to the part edge. 17.8.6 SURFACE POLISH
The mold surface should be polished to the
degree that is required on the surface of the
17.8.4 GUIDE PINS
part. The final 'stoning' or polishing on verti-
Compression molds should have leader or cal walls should be done in draw direction.
guide pins, the diameter of such to be a mini-
mum of 2% of the width plus length of the
ACKNOWLEDGEMENTS
mold. All guide pins should be the same
height and be a minimum of 6.35 mm (0.25 in) The authors would like to express their grati-
longer than the highest point of the core. They tude to Mr. M. Kilpinen and Mr. E. Kleese of
should be chamfered for lead in and guide pin GenCorp Automotive for their support and
retention should be a minimum of 1.5 times permission to write this manuscript. Thanks
the pin diameter. One guide pin should be off- are due to Ms. Marialyce Orr for editing and
set to prevent misalignment of core and cavity. proof-reading the manuscript.
396 Matched metal compression molding of polymer composites

REFERENCES 11. Boening, H.V. 1964. Unsaturated Polyesters:

Structure and Properties. Amsterdam: Elsevier.
1. Society of Plastics Industry Composites
12. Bruins, p.F. 1976. Unsaturated Resin Technology.
Institute. Semi-Annual Statistical Report, 24
New York: Gordon & Breach.
August, 1993.
13. Anderson, T.F. and Messik, V.B. 1981. Vinyl
2. Young, R., 1969. Thermoset Matched Die
Ester Resins. In Developments in Reinforced
Molding. In Handbook of Fiberglass and Plastics
Plastics I. London: Applied Science.
Composites, (ed. G. Lubin) pp. 391-448. New 14. Dow Chemical Company. 1975. DERAKANE
York: Van Nostrand Reinhold.
Vinyl Ester Resins for Corrosion Resistance.
3. Guillion, D. 1983. Section IE, 38th Annual
15. Walewski, L. and Stockton, S. 1985. Mod. Plast.
Conference of Composite Institute, The Society
62:78.
of the Plastics Industry, Inc.
16. Edwards, H.R 1987. Mod. Plast. 64:66.
4. SMC Design Manual. 1991. Catalog Number
17. Butwin, F.J. and Howes. w.e. 1986. 41st Annual
AF-180. (a) pp. 1; (b) pp. 9; (c) pp. 12; (d) pp. 25.
Conference of Composite Institute, The Society
SMC Automotive Alliance, Society of Plastics
of the Plastics Industry.
Industry, Inc.
18. Bartkus, E.J. and Kroekel, e.H. 1970. Appl.
5. Meyer, R 1987. (a) pp. 1-18; (b) pp. 179-204; (c)
Polymer. Symp. 15:113.
pp. 47-64. Handbook of Polyester Molding
19. Atkins et al. 1976. Section 2E. 31st Annual
Compounds and Molding Technology. New York: Conference of Composite Institute, The Society
Chapman & Hall.
of the Plastics Industry, Inc.
6. Kalluar, M. 1990. Section 4F, 45th Annual
20. Monte, S.J. 1978. In Handbook of Fillers and
Conference of Composite Institute, The Society
Reinforcements for Plastics. New York: Van
of Plastics Industry, Inc.
Nostrand Reinhold.
7. Weaver, w.1. 1951. US Patent 2549732.
21. Mallick, P.K. 1990. In Composite Materials
Production of Polymerized Unsaturated Resin
Technology: Processes and Properties, (ed. Mallick,
Materials of Superior Water Resistance.
P.K. and Newman, S.) pp. 25-32. New York:
8. Frilette, V. 1951. US Patent 2568331.
Oxford University Press.
Copolymerization of Styrene and Unsaturated
22. Heimbuch, RA. and Sanders, B.A. 1978. In
Resins.
Composite Materials in the Automobile Industry.
9. Kia, H. 1993. (a) pp. 1-3; (b) pp. 29-48; (c) pp.
American Society of Mechanical Engineers :
49-78; (d) pp. 95-114. In Sheet Moldi~g 111-137.
Compound: Science and Technology, (ed. Kla 23. Owens-Corning Corp. 1984. Publication 5-PL-
Hamid). Cincinnati: Hanser/Gardner 12910. (a) pp. 2; (b) pp. 3.
Publications, Inc.
24. Owens-Corning Corp. 1986. Publication 5-PL-
10. Gupta, M.K., Hoch, D.W. and Keegan, J.F. 1987.
13920. (a) pp. 5; (b) pp. 16; (c) pp. 22.
Mod. Plast. 64:70.
TEXTILE PREFORMING 18
Frank K. Ko and George W. Du

18.1 INTRODUCTION resulting composite can be tailored. These

fiber placement methods create textile pre-
Textile preforming is a fiber placement
forms which possess a wide spectrum of pore
method utilizing textile processes prior to the
geometries and pore distribution; a broad
formation of composite structures. Textile pre-
range of structural integrity and fiber volume
forms are the structural backbone of a
fraction; and fiber orientation distribution as
composite analogous to the structural steel
well as a wide selection of formed shape and
framework in a building. Starting with linear
net shape capability.
assemblies of fibers in continuous and/or dis-
A~ illustr~ted in Fig. 18.1, textile preforming
crete form, these micro-fibrous structures can
provldes a link between raw material systems
be organized into one-, two- or three-dimen-
and the composite product. Depending upon
~ional structures by means of twisting,
the textile preforming method used, the range
mterlacing, intertwining or interlooping. By
of fiber orientation and fiber volume fraction
proper selection of the geometry of the fibrous
of the preform will vary, subsequently affect-
structures and architecture and the method of
ing matrix infiltration and consolidation as
placement or geometric arrangement of the
fibers, the structural performance of the well as the translation efficiency of fiber prop-
erties to composite product.

RTM
Vacuum Impregnation
Pultruslon

Squeeze Casting
Vacuum Infiltration
Composite
Design and Product
material system CVI Assembly
selection Sol Gel
Directed Oxidation

CVD
Liquid Impregnation

Fig. 18.1 The role of preforms in composite processing.

Handbook of Composites. Edited by S.T. Peters. Published

in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7
398 Textile preforming

When combined with high performance Svedova, 1990). For more specific references
fibers, matrices and properly tailored one can examine the book by Goswami et al.
fiber/matrix interfaces, fiber architecture (1977) on yarns (linear fibrous assemblies);
promises to expand the design options for the Lord et al. (1973) on weaving; Spencer (1983)
manufacturing of tough and reliable structural and Raz (1987) on knitting and Krcma (1971)
composites. With an integrated network of on nonwovens.
structural cells in two- and three-dimensional
arrangements, textile structures not only pro-
18.1.1 1HE ROLE OF TEXTILE PREFORMS
vide a mechanism for structural toughening of
composites but also facilitate the processing of The final goal of manufacturing a composite
composites into net or near net shape struc- structure is to meet design requirements
tural parts. including performance and cost. How success-
Considering the important role which tex- fully the goal can be met depends on the
tile preforms play in the chain of composite effective use of the reinforcement material and
manufacturing processes, there is a worldwide the cost of manufacturing a quality product.
revival of interest in the technology and sci- Preform fiber architecture plays a key role
ence of the processing of textile preforms for in composite manufacturing by facilitating
composites. The serious interest in the subject processing steps including forming and resin
of textile composites can best be illustrated in infiltration. The properties of the composite
the two recent publications on textile compos- will also vary depending on the fiber architec-
ites, one in English (Chou and Ko, 1989) and ture. Fiber orientation (O) and volume fraction
the other in Russian (Tampolski'i et al., 1987). (Vf ) are key engineering parameters for textile
These two books have been translated respec- composite formability, permeability and per-
tively into Russian and English. Special formance. Accordingly, after reviewing the
journal issues have been devoted to the subject experimental evidence of the dynamic interac-
of textile composites as well (for example, tion of process-structure-performance, this
Journal of The Textile Institute, 1990, No.4, article will examine textile preforms through a
Textile Institute). This article intends to pro- unit cell based analysis relating the preform
vide a quick reference to textile preforming geometric parameters of Vf and (J and their
from the point of view of composite process- relationship to textile preform performance.
ing science. To facilitate discussion, textile
preforms are classified into linear, planar and
Formability
three-dimensional fibrous assemblies. After a
brief introduction of the processing technol- The manufacturing of composites often
ogy of textile preforms, discussion will focus requires transformation of the fiber reinforce-
on the processing kinematics of the preforms ments into various structural shapes through
and the geometric parameters which control net shape fabrication or formed shape process-
two of the most important parameters in com- ing. While 3-D textile preforming is more
posite design: fiber volume fraction Wf ) and suitable for the creation of net structural
fiber orientation (0). The role of fiber architec- shapes, 2-D textile preforms are usually
ture in the forming, resin infiltration and the formed into shapes by molding or stitching.
translation of fiber properties to the composite One of the earlier studies of fabric formability
is also discussed along with experimental evi- in composite manufacturing (Potter, 1979)
dence. For readers interested in further showed that total available deformation could
pursuing the technological aspect of the sub- be imposed on the fabric uniformly and that
ject, there are several outstanding general the modes of deformation are important para-
references on industrial textiles (Kaswell, 1963; meters for fabric formability. Potter
 Introduction 399

demonstrated that weft knitted fabrics are sig-

nificantly more conformable than biaxially
woven fabrics because deformation of the
knitted fabric in the axial, transverse and bias
direction are 50%, 50% and 26% respectively,
compared to 0%, 0% and 45% for the woven
fabric.
The same point can also be illustrated in a
comparison of the shear resistance of biaxially
and triaxially woven glass fabrics (Scardino
Shear st rain •.10
and Ko, 1981). Comparing the strain behavior
of plain woven and triaxial basic fabric of sim-
ilar area density (281.5 compared with
284.8 g/m2), under biaxial loading, it was
found that the shear deformation of the triax-
ial fabrics is considerably more uniform than
that of the biaxial fabrics (Fig. 18.2). As a result
of this comparison, it was found that triaxial
fabric is more adaptable to 3-D draw molding
than biaxial fabrics made from the same yam.
In quantifying the formability of fabrics,
Dow (1985) suggested that yam slippage and
low yam jamming angles are required for
fabric conformability. Accordingly, in fabric 100
formability modeling, fiber volume fraction ORIENTATION e (degree.)
distribution, fiber orientation and fiber inter- (b)
lacing intensity as well as the limit of
geometric deformation (all of which are gov-
erned by the architecture created by specific
textile preforming techniques) must be con-
-
.@
5
800
She.r at rain •.10

lit
sidered.
lit
...
I/)

a
Permeability '"
Textile fluid flow permeability is an indication i'"
...:z:c
K
of how easily and uniformly a matrix can be 200
infiltrated into the fibrous assembly. McCarthy ",

et al. (1991) concluded that the permeability of 00 100

textile preforms is mainly affected by fiber vol-
ume fraction and fiber orientation (Fig. 18.3). (c)
ORENTATION e (d'9r •• )

In the same figure, it can be noted that, for the

same fiber volume fraction, ordered structures Fig. 18.2 (a) Triaxial and biaxial plain-weave fabric
such as 0/900 woven fabrics have higher per- specimens partially deformed (50% of maximum
force) in ball-burst test at same displacement of ball
meability than disordered structures such as into fabric plane; (b) effect of test direction on the
discrete chopped fibers. shear stiffness of biaxial fabrics at various normal
The dependence of permeability on fiber stresses; (c) effect of test direction on the shear stiff-
volume fraction was also observed by Laos et ness of triaxial fabrics at various normal stresses.
400 Textile preforming

--
600
4 layers meability of the preforms, especially for pre-

- forms with high fiber volume fraction. The

0/90 clDth
500

>.
raOODmmat
Koczeny-Carman equation was found to be
..,::: 400 adequate to provide a quantitative relation-
~
.~ 300
ship between permeability and preform
.=..., porosity.
~.:; 200
~
'00
Properties

o , , Fiber architecture plays an important role in

0.2 0.3 0.' 0.5 0.6 the translation of fiber properties to the com-
Fib"r Volume Fraction posites as well as controlling the level of
Fig. 18.3 Permeability against fiber volume fraction
matrix infiltration. The dynamic interaction of
of textile preform. material (fiber), fiber architecture and compos-
ite properties are best illustrated by Figs. 18.5
0.1 r------------'------, and 18.6. In Fig. 18.5 (Dow, 1985), the in-plane
tensile and shear properties of carboni epoxy
0.05 composites having a 60% fiber volume fraction
are compared for unidirectional angle ply
structure, 2-D woven 3-D braided structures.
0.02
...
0/ i::.
0.01 I •• '

O( ....
•Q 6:,..,
9'0 ... ~
N-
:!:!
0.005
O/ ::C
:. o/Oi
~ 0
d.
-
0.002 0,/0
:0 ~.

---- I----
...... ....T. . . . . . .

0.001 i::.4 ,...,......,.."

.....
If.
utM' A ....

.....-
~/.
0.0005

i::./
0 ...
, -.-
"--IMoT'" ASIa
............",
0..9002 [J
~ ....
0.0001 L......L.. . . . . o.J.................J................l..~::r:=c:::=r:::::J
0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.6C
Porosity, 0

Fig. 18.4 Normalized through-the-thickness per-

meability against porosity for multiaxial warp knit
fabric preforms.

al. (1991), as shown in Fig. 18.4. In their study

Fig. 18.5 Range of values of in-plane Young's mod-
of carbon multiaxial warp knit preforms, it ulus, E and shear modulus G with T-300 carbon
was found that the introduction of through fibers U; 5200 epoxy at a vohuri'e fraction reinforce-
thickness fibers significantly increased the per- mentof 0.6.
 Introduction 401

18.1.2 CLASSIFICATION OF TEXTILE

••
400
oShott fibers
<i AUnidirBCIoonal PREFORMS
a..
e 300 • A
.Woven lamInates
• Three-climenslonal
.:::
c;,
There is a large family of textile preforming
~ 200
A
A
~ methods suitable for composite manufactur-
Ui
2 ing (Ko, 1989). The key criteria for the selection
<U A
3x 100 of textile preforms for structural composites
0
III
u: are (a) the capability for in-plane multiaxial
0
0.0 0.1 0.2 0.3 0.4 0.5
reinforcement, (b) through thickness reinforce-
Void fraction
ment and (c) the capability for formed shape
and/ or net shape manufacturing. Depending
Fig. 18.6 Effect of fiber architecture on the flexural on the processing and end use requirements,
strength of SiC/SiC composites. some or all of these features are required.
On the basis of structural integrity and fiber
It can be seen that with each reinforcement linearity and continuity, fiber architecture can
group, fiber orientation has a significant influ- be classified into four categories: discrete; con-
ence on the overall performance of the tinuous; planar interlaced (2-D); and fully
composite. Within the envelope of in-plane integrated (3-D) structures. In Table 18.1 the
properties of unidirectional reinforcement lies nature of the various levels of fiber architec-
a broad range of properties for 2-D and 3-D ture is summarized (Scardino, 1989).
textile composites. A discrete fiber system such as a whisker or
For very brittle composites such as the fiber mat has no material continuity; the orien-
SiC/SiC composites produced by the chemical tation of the fibers is difficult to control
vapor infiltration (CVI) process, the effect of precisely, although some aligned discrete fiber
fiber architecture on the level of infiltration is systems have recently been introduced. The
quite evident. Comparing the flexural strength structural integrity of the fibrous preform is
of short fiber mat, unidirectional tape, woven derived mainly from inter-fiber friction. The
laminates and 3-D braided reinforced ceramic strength translation efficiency, or the fraction
matrix composites, one can see that a 3-D fiber of fiber strength translated to the non-aligned
architecture produces structures superior to fibrous assembly of the reinforcement system,
others because of the three-dimensional inter- is quite low.
connected network of SiC fibers. The The second category of fiber architecture is
unidirectional composite has the highest vari- the continuous filament, or unidirectional (0°)
ations of void factor and unidirectional system. This architecture has the highest level
strength. It is clear in Fig. 18.6 that the perfor- of fiber continuity and linearity and conse-
mance of composites depends not only on the quently has the highest level of property
volume fraction but also the fiber orientation translation efficiency and is very suitable for
of the preform (Ko et al., 1988a). filament wound and angle ply tape lay-up

Table 18.1 Fiber architecture for composites

Level Reinforcement Textile Fiber length Fiber Fiber

system construction orientation entanglement
I Discrete Chopped fiber Discontinuous Uncontrolled None
II Linear Filament yam Continuous Linear None
III Laminar Simple fabric Continuous Planar Planar
IV Integrated Advanced fabric Continuous 3-D 3-D
402 Textile preforming

structures. The drawback of this fiber architec- While weaving, braiding and knitting can
ture is its intra- and interlaminar weakness produce planar or 3-D structures, nonwoven
due to the lack of in-plane and out-of-plane fabrics can be a 2-D planar system with ran-
yam interlacing. dom or organized fiber orientation, as well as
A third category of fiber reinforcement is the orthogonal 3-D system. The 2-D and 3-D
the planar interlaced and interlooped system. fabrics are distinguished by yam orientation
Although the intra-laminar failure problem distribution and the number of yam diameters
associated with the continuous filament sys- in the thickness direction. A 2-D fabric consists
tem is addressed with this fiber architecture, of two to three yam diameters in the thickness
the interlaminar strength is limited by the direction with fibers oriented in the x-y plane.
matrix strength due to the lack of through A 3-D fabric, consisting of three or more yams
thickness fiber reinforcement. in the thickness direction, is a fibrous network
The fully integrated system forms the wherein yams pass from surface to surface of
fourth category of fiber architecture wherein the fabric m all three directions.
the fibers are oriented in various in-plane and
out-of-plane directions. With the continuous
filament yam, a three dimensional network of 18.2 YARN PREPARATION FOR TEXTILE
yam bundles is formed in an mtegral manner. PREFORMING
The most attractive feature of the integrated
18.2.1 CLASSIFICATION OF YARNS
structure is the additional reinforcement in the
through-thickness direction which makes the In addition to preforming methods and para-
composite virtually delamination-free. meters, the physical properties of textile
Another mteresting aspect of many of the fully preforms, such as thickness, fabric tightness,
integrated structures such as 3-D woven, knits fiber orientation, etc., are also affected by the
and braids is their ability to assume complex characteristics of yarns. Linear fibrous assem-
structural shapes. blies can be composed of filaments of discrete
Another way of classifying textile preforms (staple yarns) or continuous lengths (filament
is based on the fabric formation techniques: yarns). Staple yams are held together by an
through fiber entanglement or yam twisting, appropriate level of twist, whereas filament
interlacing, interlooping, intertwining or mul- yarns mayor may not have twist. Larger fila-
tiaxial placement. While most textile preforms ment bundles that have little or no twist are
are converted from fiber to yam or yam to fab- called rovings; yarns are usually smaller than
ric structures, some preforms, such as fiber rovings and some level of twist is generally
felts, are converted directly from fiber to fab- added. The majority of high performance
ric. In Table 18.2, the four basic yam-te-fabric yarns are continuous filament yams having
formation techniques are compared. single or multiple strands.
Table 18.2 A comparison of fabric formation techniques

Preforming technology Yarn introduction direction Formation technique

Weaving Two (0°/90°) (warp and fill) Interlacing (by selective insertion of
90° yarns into 0° yam system)
Nonwoven Three or more (orthogonal) Mutual fiber placement
Knitting One (0° or 90°) Interlooping (by drawing loops of
(warp or fill) yarns over previous loops)
Braiding One (machine direction) Intertwining (position displacement)
 Yarn preparation for textile preforming 403

18.2.2 YARN GEOMETRY volume fraction can be shown to equal (Du,

Popper and Chou, 1991)
The geometric parameters which describe a lin-
ear fiber assembly include: shape of bundle 3Nr {Nr -1) + 1
cross-section, number of fibers in the bundle, ~open = 1 - (2Nr -1)2 (18.1)
dimension and shape of fiber cross-section,
bundle twist level, degree of fiber migration in where Nr is the number of rings and its rela-
the radial direction and fraction of interfiber tionship to the number of fibers, Nf, is given by
packing. Usually, the fiber bundles are assumed
to be circular in cross-section, but in reality, the Nr = ~ + ~[! + ~(2Nf -1)] (18.2)
fibers or filaments can be packed in to various
shapes. Most of the engineering fibers, such as For large numbers of fibers the fiber volume
glass and carbon, have a circular or near-circu- fraction approaches 0.75.
lar cross-section with a constant diameter. For In squared packed bundles, the fibers are
some ceramic fibers, such as Nextel, the fibers arranged in a square array. For any number of
have an ellipse shape and varying dimensions. circular fibers, if the outer layer is completely
The filament bundles used for composites have filled, the fiber volume fraction can be shown
a small amount of twist, usually less than 4 tpi, to equal the area ratio of a circle to an enclos-
the bundle surface twisting angle is small. ing square:
The geometry of interfiber packing in fiber
1t
bundles has been studied by a number of ~square = 4 = 0.785 (18.3)
researchers primarily for textile applications
(Heade et al., 1969). Three basic idealized Similar to square packing, the fiber volume
forms of circular fiber packing were identified: fraction of close packed bundle is equal to the
open-packing, in which the fibers are arranged area ratio of a circle to an enclosing hexagon:
in concentric layers (Fig. 18.7(a»; square-pack-
1t
ing, in which the fibers are enclosed by a ~dose = 2~3 = 0.907 (18.4)
square (Fig. 18.7(b»; and close-packing, in
which the fibers are arranged in a hexagonal The level of bundle fiber volume fraction pre-
pattern (Fig. 18.7(c». dicted by the above models assumed circular,
In open-packed bundles the fiber volume square and hexagonal fiber bundles for open,
fraction, defined as fiber to bundle area ratio, square and close packing, respectively.
has been computed as a function of the num- However, they apply equally well to other
ber of fibers. If the outer ring is completely shapes if the number of fibers is sufficiently
filled and the fibers are circular, the fiber large.

. .. "..
,;. .................
,...-........-......-
~
•• •• •• + •
~

" ...
• ••
~
~ .

lo-" 4
, ~
..' •
,
.( ":
•
•
~
~
• .
•
~"
.,
. ....
....
" • •••• ~

• · •
• ... • · +...
y '
-4
r.
I ...
f )0;

· ·
~

(a) (b) (c)

Fig. 18.7 Idealized fiber packings: (a) open packing; (b) square packing; (c) close packing .
404 Textile preforming

(surface helix angle) and the twist level, as

shown in Fig. 18.3. Clearly, for a given twist
inserted to the fiber bundle, as fiber orienta-
tion angle increases, yam diameter increases
(as indicated by equation (18.6» whereas the
fiber volume fraction decreases as can be seen
in Fig. 18.9, which is useful in determining the
twist level of fiber bundles. For example, to
obtain a fiber volume fraction (or fiber packing
fraction) 0.8 and a fiber orientation angle 10°,
twist level of 3 tpi should be used for the 12K,
Fig. 18.8 Geometry of twisted yarn. 7 pm fiber diameter carbon yams.

1.0
For twisted fiber bundles, the fibers are no 0.9
longer aligned along the bundle axis. Instead, -> 0.8
the fibers assume a helix configuration within 0.7
the bundle, as shown in Fig. 18.8. The fibers in 0.6
different radial layers of the bundle have a dif- 0.5
0.4
ferent helix angle:
0.3
(}i = tan-1 [n(Di - d)T] (18.5) 0.2
0.1
where d is the fiber diameter, T is twist level 0.0
and Di is the diameter of the fiber layer. 0 10 20 30 40 50 60 70 80 90
Apparently, fibers at the outer layer have a e (0)
maximum orientation (helix) angle:
Fig. 18.9 Relationship of fiber volume fraction to
() = tan-1 [n(D - d)T] (18.6) fiber orientation at various twist levels.
where D is the diameter of the bundle.
According to Hearle (1969), yam diameter is
related to the number of filaments (n) in the
yam and the packing fraction of the fibers (K) 18.3 WEAVING
in the following relationship:
18.3.1 PROCESSING TECHNOLOGY

D = ~(n~2) (18.7) Weaving, which is the interlacing of two sets

of yams usually at a right angle to each other,
Viewing a yam bundle as an individual pre- requires holding one set of yarns in parallel
form, its fiber volume fraction is actually equal rows and passing another set over and under
to its fiber packing fraction, i.e. Vf = K. the first set. The set of lengthwise yams is
Combining equations (18.6) and (18.7), we called warp and the set of crosswise yams is
have the fill. The simplest two-harness loom is
sketched in Fig. 18.10 and may be used for
v = n (1 + tan () )-2 (18.8) making biaxial plain weave fabrics shown in
f ndT Fig. 18.11(a). For more complex weaves such
For a 12K carbon yam with 7 pm fiber diame- as twill and satin as illustrated in Figs 18.11(b).
ter, one can establish the relationship between and (c), looms with more than two harnesses
its fiber volume fraction, fiber orientation are required.
 Weaving 405

harnesses are lifted. Fabrics with as many as 17

layers have been successfully woven with this
method.

Shunle Reed

Fig. 18.10 A simple two-harness weaving loom Fig. 18.12 Structural geometry of triaxially woven
(Smith and Block, 1982). fabrics (basic weave).

In a novel departure from standard weaving

methods, triaxial weaving interlaces two warp
yams and one filling yam at 60° angles as
shown in Fig. 18.12 (Dow et aI., 1970). The tri-
axial fabric has excellent dimensional stability
and is currently used almost exclusively for . "
industrial products.
Three-dimensional woven fabrics are pro-
duced principally by the multiple warp
weaving method, long used for the manufac-
turing of double cloth and triple cloths for
bags, webbings and carpets. A typical setup of
a multiwarp weaving loom is shown in Fig.
18.13. The number of layers of yams in the fab-
rics is governed by a special shedding
mechanism which controls the height that the Fig. 18.13 Setup of a multiwarp weaving loom.

v, v
•••
-
4
•'I •• •
•
-•
~ .
'I
; I '
,, .,
I

CA, (8) ce)

Fig. 18.11 Structural geometry ofbiaxially woven fabrics: (a) plain weave; (b) twill weave; (c) satin weave.
406 Textile preforming

18.3.2 STRUCTURAL GEOMETRY Triaxial weave has 90±60° yarns oriented in

one plane, resulting in a high level of in-plane
Biaxial weaves consist of 0° and 90° yarns shear resistance (Dow et al., 1970). High levels
interlaced in various repeating patterns or of isotropy and dimensional stability can be
topological unit cells. The three basic weave achieved with triaxial weave at low fiber vol-
geometries from which many other patterns ume fraction. Figure 18.12 shows a schematic
evolve are the plane, satin and twill weave. A diagram of triaxial weave geometry.
schematic diagram for the various views of Using the multiwarp weaving method, var-
these three basic weaves are shown in Fig: ious fiber architectures can be produced
18.11. These three fabrics are distinguished by including solid orthogonal panels, variable
their frequency of yarn interlacing and the lin- thickness solid panels and core structures sim-
earity of the yam segments. The plane weave ulating a box beam, or truss-like structure as
has the highest frequency of yam interlacing illustrated in Fig. 18.14. Furthermore, by
whereas the satin weav,e has the least number proper manipulation of the warp yarns, as
of yam interlacing, with the twill weave some- exemplified by the angle interlock structure,
where in between. Accordingly, the plane the through-thickness yarns can be organized
weave has a higher level of structural integrity into a diagonal pattern, as shown in Fig.
and greater ductility due to the crimp geome- 18.14(e). One limitation of the multiwarp
try produced by yam interlacing. On the other weaving method is the difficulty of introduc-
hand, the satin weave has the highest level of ing yarns in the bias direction as in the triaxial
fiber to fabric strength and modulus transla- weaving or circular weaving process.
tion efficiency due to the low level of yarn However, this is now being addressed by a
interlacing and yam linearity. The low level of modification of triaxial weaving techniques
yarn integration in satin weave also allows (Dow, 1989).
freedom of yam mobility which contributes to As illustrated in Fig. 18.15 (Pastore and Cai,
higher fiber packing density and consequently 1990a), there are four basic components to a
higher level of fiber volume fraction . generalized three-dimensional woven fabric

(a)
.. (el

Fig. 18.14 Structural (b)

geometry of various
3-D woven fabrics: (a)
solid orthogonal paneli
(b) variable thickness
solid paneli (c) rectan-
gular core structurei
(d) triangular core
structurei (e) angle
interlock. (el
 Weaving 407

geometry: warp, web, fill and surface weave When surface weave yams are employed in
yams. Warp yams are the system of yams the fabric, there are two yams for every warp
which run in the machine direction and have plane of the fabric. This system of yams con-
no crimp. These are also called 'stuffer' yams tributes the least to the mechanical properties
or 'longitudinals'. Because of their very low of the composite.
crimp, these yams provide the primary
strength and stiffness in the longitudinal (x)
18.3.3 DESIGN METHODOLOGY
direction of the material. Web yams run in the
machine direction and provide the interlacing Figure 18.16 gives the unit cell geometry for
necessary for fabric integrity. These yams con- plain biaxial weave, as proposed by Dow and
tain crimp in the through thickness direction, Ramnath (1987). In their analysis, Dow and
providing the z-directional properties of the Ramnath assumed circular yam cross-section,
system. These yams are sometimes called the same yam diameter and pitch length for
'weavers'. The 'weave angle' of the web yams both fill and warp yams. The expression of the
(0) refers to the angle of orientation of the web fiber volume fraction was derived:
yam with respect to the warp direction. Fill

("4n)
yarns are perpendicular to machine direction
and interlace with the web yams. These yams 2i+40
are sometimes called 'picks'. These yams also V, = K (~'(~) (18.9)
possess crimp in the through thickness direc-
tion, but this crimp is negligible compared to
that of the warp yams for these fabric systems. where K is the fiber packing fraction, d is the
These yams provide the transverse (y) direc- yam diameter, L is the pitch length, T is the
tional properties of the composite system. fabric thickness, 1 is the dimension shown in
Surface weave yams run in the machine direc- Fig. 18.16. The yam inclination angle to the
tion and form what is essentially a fabric plane, (), is given by:
two-dimensional weave on the surface of the
fabric. Surface weave yams are incorporated
into the structure when the web yams are
insufficient to provide a smooth surface on the
face and back of the cloth. These yarns experi-
ence crimp in the through thickness direction.

Surface weaver

filling ,========_==_===_
or I!!
yarn
Pick
Warp yarn ~lIIiIiI~~~!II!i~;;;~~
Fig. 18.15 Schematic illustration
of generalized 3-D woven fabric
projected to the x-z (fabric
length-thickness) plane.
408 Textile preforming

lengths for fill and warp yarns can be analyzed.

Based on the structural geometry shown in
Fig. 18.15, the orientations and volumetric dis-
tributions of all yarns in the 3-D weave can be
calculated also using the unit cell method.
Detailed analysis is given by Ko and Du (1992).

1.0
0 .9
0 .8
0 .7
-> 0 .6
0.5
0.4
0 .3
0.2
Fig. 18.16 Unit cell geometry of plain weave.
0 .1
0 .0
The fabric thickness is very close to two yam 0 10 20 30 40 50 60 70 80 90
diameter, i.e. e (0)
(18.11)
Fig. 18.17 Relationship of fiber volume fraction to
and approximately: fiber orientation for plain weave.
1 1 (18.12) 18.4 KNITIING
d = tanO
18.4.1 PROCESSING TECHNOLOGY
Equation (18.9) is then simplified to
1 Knitting is the interlocking of one or more
v: (1t) ~+26 (18.13) yarns through a series of loops (also called
stitches). The lengthwise columns of stitches
f = 4 K (~J corresponding to the warp in woven fabrics
are called wales; the crosswise rows of stitches
Figure 18.17 plots the fiber volume fraction corresponding to the filling are known as
against the yam inclination angle. It can be courses. Knitted structures can be classified by
seen that as the inclination angle increases, the basic loop formation mechanism into weft
pitch length becomes longer which results in a knits and warp knits. In weft knitting, as
lower fiber volume fraction. The woven fabric shown in Fig. 18.18(a), yam feeding and loop
has the tightest structure at the inclination formation occur at each needle in succession
angle of 60° (when L/ d =3 in equation (18.10». along the wale direction and all the courses of
In this calculation, the fiber packing fraction 1C loops are composed of single strands of yam.
is assumed to be 0.8. In warp knitting, there is a simultaneous yarn-
The above analysis is given only for the sim- feeding and loop-forming action occurring at
plest of woven structures. Different weave every needle and all the wales of loops are
patterns, non-circular yarn cross-sectional composed of single strands of yam as illus-
shape, different yam dimensions and pitch trated in Fig. 18.18(b).
 Knitting 409

Direction of knitting
,,/'
"'':'" /

(b)

Direction of knitting

Fig. 18.18 Yam feeding and loop formation: (a) weft knitting; (b) warp knitting (Spencer, 1983).
Stitch (loop) formation is similar in both weft technology can be found in Spencer (1983) and
and warp knitting. The formation of the Raz (1987).
stitches in a single wale is illustrated in Fig. Knitted 3-D fabrics are produced either by
18.19. In Step I, the needle rises through loop weft or warp knitting. An example of a weft
A from its lowest position; in Step 2, yam slips knit is the near net shape structure knitted
under the tip of the needle and onto the stem; under computer control by the Pressure Foof®
in Step 3, ascending hook catches the new yam process (Williams, 1978). In a collapsed form
at the top of its rise and begins to descend; in this preform has been used for carbon-carbon
Step 4, the new yam slips under the tip and aircraft brakes. While weft knitted structures
into the hook; in Step 5, the needle moves have applications in limited areas, multiaxial
down until the tip slides under loop A and the warp knit (MWK) 3-D structures are more
hook pulls the new loop through. After the promising and have undergone a great deal
completion of five steps, loop B is formed and more development in recent years. Schematic
the process is repeated. In a knitting operation, of a MWK LIBA system is given in Fig. 18.20,
each of the needles is controlled by a cam to in which up to six layers of insertion yarns
rise and fall in synchronization with the other plus one layer of non-woven can be stitched
needles. Detailed description of the knitting together.

2 3

Fig. 18.19 Stitch for-

mation in knitting
machines (Smith and
Block, 1982).
410 Textile preforming

(a)
~""
"

Angles
,
,,'~
adJustabl~.) (b) ~"" ""r--
Angles adjustable .)

Fi~.18.~O Multiaxial warp knit with four layers (0°, 90° and ±6) of inserted yarns and (a) chain stitch or (b)
tricot stitch.

18.4.2 STRUCTURAL GEOMETRY ture for the incorporation of 0° and/or 90°

Knitted fabrics are traditionally identified insertion yarns.
with socks, underwear and sweaters. In the The MWK fabric system consists of warp (0°),
search for methods to reduce composite man- weft (90°) and bias (±8) yarns held together by
a chain or tricot stitch through the thickness of
uf~cturing costs, textile preforms including
knItted structures are receiving increased the fabric, as illustrated in Fig. 18.2l.
Theoretically, the MWK can be made to as
interest in the composite industry. While con-
formability and productivity are obvious many layers of multiaxial yarns as needed, but
attributes for knitted preforms, the availability current commercially available machines only
of a broad range of micro- and macrostructural allow four layers (the Mayer system) of 0°,90°,
+0 and -fJ insertion yarns, or six layers (the
geometries has only recently been recognized.
~IBA.system) of 2(90°),0°,2(+0) and 2(-fJ)
The non-linearity of knitting loops, severe
bending of yarns during the knitting process msertion yarns to be stitched together. All lay-
and limited fiber packing density resulting in ers of insertion yarns are placed in perfect
the formation of resin pockets within a knit- order each on top of the other in the knitting
ting loop prevent knits from being considered process. Each layer shows the uniformity of the
uncrimped parallel yarns. The insertion yams
for structural applications.
usually possess a much higher linear density
The development of technology for the
than the stitch yams and are therefore the
directional insertion of linear yarns in weft
major load bearing component of the fabric.
and warp knits greatly enhances opportunities
for knitted preforms for conformable struc-
tural composites by combining the
18.4.3 DESIGN METHODOLOGY
conformable foundation knit structure with
directional reinforcement. As shown in Fig. Similar to the 2-D woven fabrics, the unit cells
18.20, sewing threads (high twist yarns) or for the knitted structures are also different
very fine yarns are used to form a base struc- depending on the knit constructions such a~
 Knitting 411

Fig. 18.21 Multiaxial warp knit LIBA system.

stitch patterns and laid-in insertions. To illus- illustrated in Fig. 18.23(b). For untwisted fiber
trate the use of the unit cell method for bundles under compression applied during
relating fiber volume fraction, yarn orientation preforming or composite processing, they
and processing variables, a plain weft knit as have a ribbon-like cross-section similar to a
shown in Fig. 18.18(a) is selected as an exam- race-track with a width-to-thickness aspect
ple. The unit cell geometry identified for the ratio of f > 1 as illustrated in Fig. 18.23(a). For
plain weft knit is shown in Fig. 18.22, having a composite applications, untwisted fiber bun-
dimension of x (course width), y (half wale dles are usually used in knitting, which have
width) and z (fabric thickness). an aspect ratio f slightly larger than 1 at the
off-machine state. To increase the fiber volume
fraction for knitted structures, very high pres-
sure will be required to reduce the knit
thickness. Under the compression status, the
yarn aspect ratio f can increase to as high as 12
for untwisted bundles, provided that there are
no restrictions applied to yarn edges.

Fig. 18.22 Unit cell geometry of plain knitted struc-

ture. (a) f =wIt> 1 (b) f =1

In traditional textile fabric manufacturing, Fig. 18.23 Idealized yam cross-sections: (a) race-
highly twisted fiber bundles are used. These track cross-section with width-to-thickness aspect
materials can maintain a circular shape with a ratio / > 1; (b) race-track shape becomes circular
width-to-thickness aspect ratio of f = 1, as when/= 1.
412 Textile preforming

In this analysis, the knit thickness is assumed relative course width (x/w), relative half wale
to be approximately equal to two yam thick- width (y/w) and yam aspect ratio (f) under
ness (t) for computational purposes, i.e. compression is depicted in Fig. 18.24, using the
geometric model developed. In the calculation,
Z '" 2t (18.14)
we use the fiber packing fraction k =0.8, which
The yam orientation angle (0), which is the is within the range for tightly packed yam bun-
angle made by the fabric axis (in x direction) dles according to experimental observation.
and the yam path projected to the fabric sur- Also, to show the processing window of fiber
face plane (x - y), is given by: volume fraction in highest region, one can
assume yam jamming in the course (x) direc-
tion, i.e. x/w = 3 according to equation (18.18).
(18.15) As can be seen from Fig. 18.24, the fiber vol-
ume fraction Vf decreases with the increase in
relative half wale width in the range of y /w =
The fiber volume fraction (Vf)' which is defined 2-10. When y/w is beyond 10, the fiber vol-
as the ratio of volume of total fibers to the over- ume fraction slightly increases and soon
all composite volume, can be derived as: approaches a constant with the increase in rel-
ative half wale width. The wale width cannot
be smaller than 4 yam widths, or y/ w ~ 2 as
~+2 given by equation (18.19). Knitted yams have
1
~ = a K - - ---iw~_==
x J.-
2_ 2w an aspect ratio i'" 1 at free-stress status (as
cosO 1+tan2 - made off-machine) and the fiber volume frac-
ww fx
tion for the knitted preform has a minimum
value. Figure 18.24 shows that, for the plain

~1+:'7iI
weft knit at its tightest possible structure (x/w
= 3, y/w = 2 and k = 0.8), its maximum fiber

+
(18.16) volume fraction is only about 0.274.
To increase the fiber volume fraction, a com-
pression in the fabric thickness direction is
necessary. The effect of the compression is the
where k is the fiber packing fraction within increase in yam aspect ratio (i.e. yarns within
yam bundles and a is the shape correction fac- the knitted structure become wider in x-y plane
tor defined as:

(18.17) ooo r--r-------------------r~~

The limiting geometry of the knitted structure

due to yam jamming is governed by:
"E 030
'0
:-
"
x ~ 3 (18.18)
W .8u: 020
I K 1 (clrcuW y.". wIthouI compreuIon)
0.10

.JL ~ 2 (18.19) 000

1 10 100 1000
W Y/w

The processing window of fiber volume frac-

tion for knitted structures within the possible Fig. 18.24 Processing window of fiber volume frac-
ranges of key processing parameters, such as tion for the plain knitted structure.
 Braiding 413

but thinner in z direction). As a result, the yam tal. A schematic of a horizontal braider is
coverage over the fabric increases, whereas the shown in Fig. 18.25. Although braiding is simi-
volume of the preform decreases due to the lar to filament winding in many ways, the
decrease in fabric thickness. These two factors, major difference between braiding and fila-
the increased yam coverage and decreased pre- ment winding is that braids are interlaced
form volume, raise the fiber volume fraction to structures having as many as 144 or more inter-
a much higher level. As shown in Fig. 18.24, at lacing per braiding cycle (or pick).
a maximum aspect ratio f = 12, the fiber vol- Three-dimensional braiding technology is
ume fraction can be as high as 0.475. an extension of 2-D braiding technology in
A series of studies on the technology, struc- which the fabric is constructed by the inter-
ture and properties of the MWK preforms and twining or orthogonal interlacing of yams to
composites have been reported by Ko and his form an integral structure through position
co-workers (1980, 1982, 1985, 1986, 1988b). In a displacement. A unique feature of 3-D braids
recent study, a unit cell based geometric model is their ability to provide through the thickness
of the four-layer MWK structure as shown in reinforcement of composites as well as their
Fig. 18.21 was developed by Du and Ko (1992). ready adaptability to the fabrication of a wide
Based on the experimental observations, the range of complex shapes ranging from solid
unit cell geometry of the MWK fabric is iden- rods to I-beams to thick-walled rocket nozzles.
tified and a geometric model is developed Three-dimensional braids have been pro-
relating the fiber volume fraction and fiber ori- duced on traditional Maypole machines for
entation in terms of structural and processing ropes and packings in solid, circular or square
parameters. cross-sections. The yam carrier movement is
activated in a restricted fashion by hom gears.
A 3-D cylindrical braiding machine of this form
18.5 BRAIDING was recently introduced by Albany with some
modification that the yam carriers do not move
18.5.1 PROCESSING TECHNOLOGY
through all the layers (Brookstein, 1991). 3-D
Braiding is an old textile technology, tradition- braiding processes without using the hom
ally used for the manufacture of a wide variety gears, including Track and Column (Brown et
of linear products ranging from cables, electri- aI., 1988) and 2-Step (Popper and McConnell,
cal insulators and shoelaces to surgical sutures. 1987), have been developed since the late 1960s
Recognizing the high level of conformability in the search for multidirectionally reinforced
and the damage resistance capability of composites for aerospace applications.
braided structures, the composites industry A generalized schematic of a 3-D braiding
had found structural applications for braided process is shown in Fig. 18.26. Axial yams, if
composites ranging from rocket launchers to present in a particular braid, are fed directly
automotive parts to aircraft structures.
Two-dimensional braided structures are
intertwined fibrous structures capable of form-
ing structures with 00 and .j:fJ fiber orientation.
Although 2-D braids can be fabricated in tape
form, the majority of braided structures are
fabricated with a tubular geometry. Thickness
is built up by overbraiding previously braided
layers similar to a ply lay-up process. Braiding
can take place vertically or horizontally, but a Fig. 18.25 Schematic of tubular braider with gantry
majority of the composite braiders are horizon- system.
414 Textile preforming

Track and column braiding is the most pop-

ular process in the manufacturing of 3-D
braided preforms. The mechanism of these
braiding methods differs from the traditional
hom gear method only in the way the carriers
are displaced to create the final braid geome-
try. Figure 18.27(a) shows a basic loom setup in
a rectangular configuration. The carriers are
arranged in tracks and columns to form the
required shape and additional carriers are
added to the outside of the array in alternating
locations. Four steps of motion are imposed to
the tracks and columns during a complete
braiding cycle, resulting in the alternate x and
y displacement of yam carriers, as shown in
Fig. 18.27(b-e). The formation of shapes, such
Fig. 18.26 Schematic of a generalized 3-D braider. as T-beam and I-beam, is accomplished by
proper positioning of the carriers and the join-
into the structure from packages located below ing of various rectangular groups through
the track plate. Braiding yams are fed from selected carrier movements. The track and col-
bobbins mounted on carriers that move on the umn braiding machine can also be used to
track plate. The pattern produced by the create 2-step braids and other similar 3-D
motion of the braiders relative to each other structures by simply adding a certain number
and the axial yams establish the type of braid of axial yams and removing most of the braid-
being formed, as well as the microstructure. ing yams (Du and Ko, 1993a).

IL
y
I~
'-'" B
B ~
A :~ ~ I~

~
Track direction X

, , , , (a)
, , , , (b)

!
...
i
I

( 1
i
..,;
Ah ...
...
L.;
A ~B ~~ tL
( L.: ~ L.: foo- L.: r-- B
! ' ....
-'- -'- (e)-'- .-'-
I
! i 1 1 i
(e) (d)

Fig. 18.27 Formation of a rectangular 3-D track and column braid, using 4 tracks, 8 columns and 1 x 1
braiding pattern. (a) Initial loom setup; (b) Step 1: tracks move hOrizontally; odd tracks move to left and
even tracks move to right; (c) Step 2: columns move vertically; odd columns move down and even columns
move up; (d) Step 3: tracks move horizontally; odd tracks move to right and even tracks move to left; (e)
Step 4: columns move vertically; odd columns move up and even columns move down.
 Braiding 415

18.5.2 STRUCTURAL GEOMETRY from 5° in almost parallel yam braid to

approximately 85° in a hoop yarn braid,
As with woven fabric, braids can be formed
depending on the mandrel dimension, the
with different yam interlacing patterns by
machine speed ratio and the convergence
simply changing relative position of carriers
length (Du et al., 1990).
on the track ring. If one bias yam continuously
The 2-D braid can be defined as a fabric
passes over one yam and then under one yam
which consists of only two layers of bias yarns
of the opposing group, the pattern is desig-
interlaced with each other. In 3-D braided
nated as 1/ 1 braid, or diamond braid as
structures, at least three layers of bias yarns go
generally recognized. Other simple interlacing
through the thickness in a zig-zag manner
patterns in common use include 2/2, 3/3, 2/1
along the diagonal direction. Similar to the 2-D
and 3/1 braids. Figure 18.28 shows the pattern
structure, longitudinal yams can be incorpo-
of 2/2 braid with axial insertion. Among all
rated in the 3-D braid for the enhancement of
these patterns, the 2/2 braid is the most popu-
stiffness and strength in the length direction.
lar and has been referred to as regular,
Regardless of the difference in the carrier pro-
standard, plain or flat braid. The path of axial
pelling mechanism, there are basically two
yams is independent of braid interlacing pat-
~es of 3-D braiding looms: rectangular and
terns, they are always over one group of bias
cIrcular. The former is usually used to fabricate
yams, but under the opposite group.
solid structures such as panel, 1- and T-beam
The formation of shape and fiber architec-
etc. and the latter for making thick wall tubular
ture are illustrated in Fig. 18.29 which depicts
structures. Figure 18.30(a) shows a schematic
the process of braiding over an axisymmetric
of a 3-D braided slab.
shape of revolution. Braiding angle can range

Fig. 18.28 Yam structure in 2-D braid: braiding

yams at ±e to braid axis, optional axial yams at 0°
to braid axis.

Fig. 18.30 3-D braided solid slab (and its cross-sec-

Fig. 18.29 Braid formation over a shaped mandrel. tion as seen on SEM).
416 Textile preforming

18.5.3 DESIGN METHODOLOGY cross-section of yarn bundles in the fabric

(yarn level) and orientation and distribution of
The unit cell geometry of 2-step braids has fibers in the 3-D network (fabric level). The
been reported by Du et al. (1991). Based on unit-cell technique is commonly used to estab-
experimental observations, diamond and rib- lish the geometric relation. In most of 2-D
bon shapes for the axial and braider yarns, fabrics a unit cell geometry is readily identi-
respectively, are assumed in their analysis. The fied, but in complex 3-D fabrics it can be very
unit cell was defined; pitch length and per- difficult to define.
centage of braider yarns were identified as key The fiber volume fraction of a 3-D fabric
process parameters which control the braid depends on the level to which yams pack
microstructure and the jamming criterion for against each other in the structure and the
the 2-step braid was given. level to which fibers pack against each other in
The traditional approach used in modeling a yam, as illustrated in Fig. 18.7. In addition to
3-D braided composites is to artificially define the level of packing fraction, the fibers also
a unit cell geometry for a 3-D braided struc- establish the yarn cross-sectional shape, i.e.
ture without providing any relationship yam packing in fabrics. This shape plays a
between processing variables and geometric very significant role in determining how many
parameters. All fibers in the unit cell are fibers can be packed into a fabric. One good
assumed to incline in four different diagonal example is the yam packing in 2-step braided
directions, as well as along the longitudinal preforms (Du et al., 1991). Due to the use of
direction, if any. Fiber volume fraction is untwisted fiber bundles and high braiding
assumed to be either known or measured. The tensions, cross-section of axial yams in the 2-
approach used in geometric modeling of tex- step braid is deformed to prismatic shapes,
tile structures is to first determine the giving most the compact yam packing within
dimension, shape and fiber architecture of the the braided structure. For the track-and-col-
unit cell based on process and structural umn braids, the braiding tensions are lower
analysiS; using the unit cell geometry identi- compared to the 2-step braids and the cross-
fied, the relationship between processing sections of yams actually have a polygonal
variables and key geometric parameters can shape.
readily be established. The microgeometric model for the track-
The key geometric parameters of 3-D braids and-column braid has been investigated by
(which affect reinforcement capability and many researchers since the early 1980s
composite processability) include braider ori- (Pastore et al., 1990b; Li et al., 1990). The most
entation, total fiber volume fraction, volume recent one was given by Du and Ko (1993a),
fraction of inter-yarn void and axial fiber per- which does not only relate geometric parame-
centage of total fibers. Although there are only ters and processing variables but also provides
two simple process parameters adjustable to limiting braid geometry due to yarn jamming.
control the microstructure of 3-D braids (speed In their analysis, the yams are assumed as
ratio between braiding and take-up and linear rigid circular rods. This assumption is valid
density ratio of braider and axial yarns), the when braiding at high yam tensions. When
process-structure model of 3-D braid is com- low yarn tensions are used, yam crimp will be
plicated. introduced during braiding or during post-
Normally, yarn bundles consisting of preforming processing due to distortion. This
numerous continuous filaments are used for yam waviness (crimp) may increase the fiber
fabric preforms, thus, the fabric microstructure volume fraction of the braid with the sacrifice
has three levels: geometry of interfiber pack- of directional reinforcing efficiency.
ing in the yarn bundle (fiber level),
 Braiding 417

Figure 18.31 shows an idealized braid cross- where K is the fiber packing fraction (fiber-to-
section cut longitudinally at a 45° angle to the yam area ratio). Due to the bulky fiber and
braid surface. There are four groups of yarns nonlinear crimp nature, it is difficult to fabri-
inclined at angle e with the braid axis (z direc- cate the braid with tightest structure. In
tion) in different directions; the yarns in each practice, the yam orientation angle (braid
group are parallel to each other within a spe- angle) is determined from the yam diameter
cific plane. Two groups of yams are parallel to (d) and braid pitch length (hJ The fiber vol-
the x'-z plane; the other two are parallel to the ume fraction is controlled by the braiding
y'-z plane. The cutting plane is so selected that angle and the braid tightness factor. The gov-
it cuts through the diameter of a group of yarns. erning equations are given below:

e= sin-1 ~( (h/ ~2 + 4 ) (h z 2:: 2d) (18.21)

V = ~ < ~ 1C cos e (18.22)

f cos e - 2 1 + cos2e

X' where 'YJ is the fabric tightness factor, which is

within the range of 0 to rc/4. This tightness fac-
tor must be so selected that the required fiber
volume fraction is achieved and also that the
over-jamming condition is avoided.
A Figure 18.32 shows the Vr-O relationship
prior to and at the jamming condition, based
on the governing equations. The fiber packing
fraction, K, is assumed as 0.785. As can be seen,
there are three regions of fiber volume frac-
tion. The upper region cannot be achieved due
to the impossible fiber packing in a yam bun-
dle. Jamming occurs when the highest

Fig. 18.31 Braid cross-section cut longitudinally at 1.0

a 45° angle to the braid surface by the x'-y plane 0.9
ABeD. z is the braid length direction. O.S
0 .7
-> 0.6
05

The braid has the tightest structure when 04

each yam is in contact with all its neighboring

0 .2
yams, in other words, the yarns are jammed
against each other. At the jamming condition, 0 .1

the fiber volume fraction VI as a function of the

braid angle e was derived by Du and Ko
10 20 30 40 50 60 70 SO 90
e( )
(1993a):
Fig. 18.32 Relationship of fiber volume fraction to
V=~K cosO (18.20) braiding angle for various tightness factors. Fiber
f 2 1 + cos20 packing fraction K is assumed to be 0.785.
418 Textile preforming

braiding angle is reached for a given fabric

tightness factor 7J. The non-shaded region is
the working window for a variety of Vr-O com-
binations. Clearly, for a given fabric tightness,
the higher braiding angle gives higher fiber
volume fraction and for a fixed braiding angle,
the fiber volume fraction is greater at higher
tightness factors.
Theoretically, the 2-D braid can be consid-
ered as a single layer of 3-D braid. For prepreg
or tape braiding without much change in yarn
width and for the braiding of structures with
constant cross-sections, the V{..o relation is sim-
ple. For braiding of dry tows and structures
with variable cross-sections wherein a dynamic
interaction of the braiding machine and tow
geometry takes place, there is a need for a more
general representation of the kinematics of the Fig. 18.33 Schematic of the NovolteX® process.
braiding processes which allows for the tow
width to vary over a limiting geometry. Two
mathematical models have been developed, the thickness fiber reinforcement. For illustration
first is the kinematics model (Du et al., 1990) purposes, our analysis of fiber volume fraction
which provides the relationship between the distribution will be focused on orthogonal
braiding angle and the braiding process para- nonwoven 3-D fabrics.
meters and the other is the unit cell model (Du While woven 3-D fabrics have a long his-
and Ko, 1993b) which relates braiding angle to tory of development and is clearly a product
yarn geometry to predict fiber volume fractions of the textile industry, the class of orthogonal
Vf along both the braiding and axial directions. nonwoven 3-D fabrics is a product of the
twentieth century, developed in the aerospace
industry for specific composite applications.
18.6 NONWOVEN Pioneered by aerospace companies such as
General Electric and AVeO, the nonwoven 3-
18.6.1 PROCESSING TECHNOLOGY
D fabric technology was developed further by
Nonwoven structures are fiber to fabric assem- Fiber Materials Incorporated. Recent progress
blies produced by chemical, thermal or in the automation of the nonwoven 3-D fabric
mechanical bonding or a combination of the manufacturing process was made in France by
above. Starting with discrete fibers or continu- Aerospatiale (Pastenbaugh, 1988), in Japan by
ous filaments (mostly tows), the fibers are Fukuta of the Research Institute for Polymers
randomly distributed or preferentially ori- and Textiles (Fukuta et al., 1982 and 1984) and
ented by dynamic combing (carding) or more recently by Mohammed (1989).
hydrodynamic (waterjet) methods. Orthogonal nonwoven (ON) 3-D fabrics are
The Novoltex® structure developed by SEP fabricated by maintaining one stationary axis
as shown in Fig. 18.33 (Geoghegan, 1988) is an either by yam pre-deposition or a spacer rod
example of a mechanically bonded structure which is subsequently retracted and replaced
wherein multiple layers of oriented or random by an axial yarn. The placement of the planar
fiber webs are needled together to create an yam systems is carried out by inserting the
integrated structure which has through yarns orthogonal to the axial yam system in an
 Nonwoven 419

20'

Fig. 18.34 Orthogonal nonwoven by direct method.

alternating manner. In Fig. 18.34, the method 18.6.2 STRUCTURAL GEOMETRY

of direct formation of ON 3-D fabric is shown
Structural geometries resulting from the vari-
(Stover et aI., 1971). By proper arrangement of ous processing techniques are shown in Fig.
the axial yarns prior to planar yam placement, 18.35: (a) and (b) show the single bundle xyz
3-D fabrics of various shapes and densities can
fabrics in a rectangular and cylindrical shape;
be produced.
(c) demonstrates the multiple yam bundle
possibilities in the various directions.

y
(a) (b)
Fig. 18.35 Orthogonal
nonwoven fabrics.
420 Textile preforming

18.6.3 DESIGN METHODOLOGY against d/ dx ratios, assuming a fiber packing

A unit cell geometry for the orthogonal non- fraction of 0.8. For all three levels of dJ dx
ratios, the fiber volume fraction first decreases
woven 3-D fabric is shown in Fig. 18.36,
with the increase in dy / dx ratio, reaches a min-
assuming circular cross-section for yarns in all
imum and then increases. As can be seen in the
three directions. The fiber volume fraction for
figure, the maximum fiber volume fraction is
about 0.63 at either high or low dy / dx ratios,
whereas the minimum fiber volume fraction of
about 0.47 is achieved when both d/ dx and
dJ dx ratios are equal to 1.

18.7 SUMMARY AND CONCLUSIONS

In this chapter, we have first discussed the role
and importance of textile preforms in compos-
ite design processing and design, followed by
classifying them into linear (1-D), planar (2-D)
Fig. 18.36 Unit cell for orthogonal nonwoven struc-
ture. and three-dimensional (3-D) fibrous assem-
blies. The objective of this chapter is to
the 3-D ON structure can be shown to have the describe the design methodology of the fiber
following form: architecture for representative textile preform
structures currently used for composite rein-
forcements. After a brief introduction to the
formation technology of each preform, its fab-
+ dz) + d2y(dx + dz) + d2(d ric structure is shown and the geometry of a
d~d
xy z x + dy )
(18.23) unit cell is defined. The relationship between
(d x + dy)(dx + d)(d y + d) the engineering parameters (VfI 0) and the key
where dx' dy and dz are diameters of the yarns processing variables (such as preform pattern,
in x, y and z directions, respectively and K is tightness factor 1] and linear density ratio etc.)
the fiber packing fraction of the yarns. within the range of achievable geometry is
Figure 18.37 plots the fiber volume fraction established from the geometric model.
A summary of preform fabrication tech-
0.65 - , - - - - - - - - - - - - - , niques has been given in Table 18.2. Table 18.3
gives a summary of the engineering and pro-
0.60
cessing parameters. Ranges of fiber
orientation angle and fiber volume fraction for
each fabric preform commonly used for com-
0.55 posite reinforcements are also included in
Table 18.3. It should be noted that although the
0.50 achievable range of fiber volume fraction is
restricted by theoretical fabric geometric limits
0.45
Minimum fiber volume fraction
+-.-.~~~_~~~~~~~....I,
due to yam jamming, it is possible that a
1<f'
10-3 10.2 10. 1 10° 101 1 if cf1 104 higher fiber volume fraction can be achieved
in reality because of the compressible nature of
dy/dx
the preforms. A composite having a higher
Fig. 18.37 Process window of fiber volume fraction fiber volume fraction can be made simply by
for orthogonal nonwoven fabrics. squeezing the preform to a smaller mold
 Summary and conclusions 421

Table 18.3 Engineering and processing parameters for textile preforms

Preform Fiber orientation, 0 (0) VI Processing parameter bundle

Linear assembly o- yam surface helix angle Bundle tension, transverse
Roving 0=0 0.6 - 0.8 compression, fiber diameter,
Yam 0=5-10 0.7 - 0.9 number of fibers, twist level
Woven Of - yam orientation in fabric plane Fiber packing in yam, fabric
0c - yam crimp angle tightness factor, yam linear
2-D Biaxial Sf =0/90, 0=30-60
c - 0.5 density ratios, pitch count,
2-D Triaxial Of =0/90/±30-60, 0=30-60
c - 0.5 weaving pattern
3-D Woven Of = 0/90, 0=30-60
c - 0.6
Non-woven Ox - fiber/yam orientation along x axis (2-D non-woven) fiber
o - fiber / yam orientation along y axis packing in fabric, fiber
0: - fiber/yam orientation along z axis distribution (3-D orthogonal)
oxy - fiber distribution on fabric plane fiber packing in yam, yam cross
section, yam linear density
ratios
2-D Non-woven oxy ;;:: uniform distribution, 0.2 - 0.4
3-D Orthogonal Ox' 0y' Oz = a 0.4 - 0.6
Knit Os - stitch yam orientation Fiber packing in yam, fabric
0i - insertion yam orientation tightness factor, yarn linear
density ratios, pitch count,
2-D Weft knit 0=30-60
s 0.2 - 0.3 stitch pattern
3-DMWK Os = 30 - 60, 0i;;:: 0/90/±30-60 0.3 - 0.6
Braid °- braiding angle Fiber packing in yam, fabric
tightness factor, braid diameter,
2-D Braid 8=10-80 0.5 - 0.7 pitch length, braiding pattern,
3-D Braid 0=10-45 0.4 - 0.6 carrier number

during the process of matrix addition; how- Assuming a tightness factor YJ of 0.573, possi-
ever, a composite with a fiber volume fraction ble braiding angles range from 0 to 40°.
higher than theoretical maximum will have a Young's moduli and Poisson's ratios of fiber
certain degree of fiber crimp and its fiber ori- and matrix are given as Ef = 33.5 Msi, Em =
entation will also be distorted. 1.3 Msi, V f = 0.3 and Vrn = 0.11. The elastic con-
The geometric models of textile preforms stants of the carbon-carbon composite was
presented in this chapter provide a quantitative obtained from the Fabric Geometric Model
communication link between the preform (FGM) (Ko et al., 1987). Figure 18.38 shows the
manufacturer, composite processors and prod- composite stiffness in different directions
uct design engineers. By reducing fiber within the working window of fiber volume
architecture and textile preforming processes fraction and fiber orientation. As can be seen,
into engineering and processing parameters Young's modulus, EaxiaJ, in the axial direction
Vf , e and YJ, rational composite design proce- decreases and in-plane shear modulus, Gin_ lane'
dures and process control guides can be increases with the increase in braiding angfe e.
established. For example, the mechanistic Young's moduli in both hoop and radial direc-
design of a composite product can be demon- tions, Ehoop and EradiaJ' have the same value at
strated using a tubular 3-D braided zero braioing angle, but depart and both
carbon-carbon composite as an example. increase as the braiding angle become higher.
422 Textile preforming

16~~----------____- ' r -______~ 0.006 -r=-----:----:---:r-------:r---------,

14
Braid Axis
=-
~
~ 12 -=- 0.004
., ~
-§ 10 :cttl
"8 Q)
~ 8 E 0.002
Q)
a.
6 Umltingfiber
ardlitecture
4 0.000 t-~:;::;;;;;::::;::::::::::::::::_~~_~
o 3 6 9 12 15
Fiber diameter (J.lm)

o 5 10 15 20 25 30 35 40 45 50 55 60
Fig. 18.39 Effect of fiber volume fraction and diam-
Braid angle. 9(0) eter on preform permeability.
Fig. 18.38 Stiffness properties of 3-D braided car-
bon-carbon composite. Fiber packing and tightness can be quantified using the Kozeny-Carman
factor are assumed as /C = 0.785, TJ = 0.573; Young's equation, as shown in Fig. 18.39, noting that
moduli and Poisson's ratios of fiber and matrix are the permeability is in the unit of mass flow
given as Ef =33.5 Msi, Em =1.3 Msi, V f =0.3, Vrn =011.
rate per length of preform. In this example, the
shape factor tP is assumed to be 0.75 at a pres-
sure drop across the preform of 60 psi with a
The other example of the application of the fabric thickness of 0.5 in.
fiber architecture models to the composite pro-
cessing is to predict the permeability of fabric
preforms. As suggested by the well known REFERENCES
Kozeny-Carman equation, there are two Brookstein, D.S. 1990. Interlocked Fiber
major geometric parameters which greatly Architecture: Braided and Woven. Proc. 35th
affect permeability of fibrous materials, i.e. Intern. SAMPE Symposium, Society for the
porosity of fabric preforms e and characteristic Advancement of Material and Process
dimension of fibers tPD, where tP is the shape Engineering, Vol 35, pp. 746-756.
Brown, R.T., Patterson, G.A. and Carper, D.M. 1988.
factor and D is the fiber diameter. Other para- Performance of 3-D Braided Composite
meters which also affect the permeability are Structures. Proceedings of the Third Structural
flow properties, pressure drop and part thick- Textile Symposium, Drexel University,
ness. These parameters have been shown to be Philadelphia, PA.
independent of preform fiber architecture. Chou, T.W. and Ko, EK., eds. 1989. Textile Structural
From geometric analysis, one can construct Composites. New York: Elsevier.
the Vf-O relationship and determine their Dow, N.F. and Tranfield, G. 1970. Preliminary
Investigations of Feasibility of Weaving Triaxial
dependence on the process parameters. The Fabrics (Dow Weave). Textile Research Journal,
fabric porosity can easily be calculated from Vf pp. 986-998.
(e = 1 - Vf)' whereas the shape factor of fibers Dow, N.F. 1985. Woven Fabric Reinforced
tP is related to the fiber orientation () and the Composites for Automotive Applications.
flow direction in composite processing. From Technical Final Report, NSF Grant No. DMR-
our experiments, we have observed that tP '" 8212867, MSC TFR 1605/8102, December.
1.5 when most of the fibers are aligned paral- Dow, N.E and Ramnath, V. 1987. Analysis of Woven
Fabrics for Reinforced Composite Materials.
lel to the flow direction and tP '" 0.75 when the NASA Contract Report 178275.
fibers are perpendicular to the flow direction. Dow, RM. 1989. New Concept for Multiple
The effect of the fiber volume fraction and Directional Fabric Formation. Proc. 21st Intern.
fiber diameter on the permeability of air flow SAMPE Tech. Conf, September 25-28.
 References 423

Du, G.W, Popper, P. and Chou, T.W 1990. Process Applications. J. Industrial Fabrics 4(2).
Model of Circular Braiding for Complex- Ko, EK, Pastore, CM., Yang, J.M. and Chou, T.W
Shaped Preform Manufacturing. Proc. Symp. on 1986. Structure and Properties of
Processing of Polymers and Polymeric Composites, Multidirectional Warp Knit Fabric Reinforced
ASME Winter Annual Meeting, Dallas, Texas, Composites. In Composites '86: Recent Advances
Nov. 25-31. in Japan and the United States, eds. Kawata, K,
Du, G.W, Popper, P. and Chou, T.W 1991. Analysis Umekawa, S. and Kobayashi, A Proceedings,
of Textile Preforms for Multi-directional Japan-U.s. CCM-III, Tokyo, pp. 21-28.
Reinforcement of Composites. J. Mater. Sci. 26: Ko, EK, Pastore, CM., Lei, Charles and Whyte,
3438-3448. D.W 1987. A Fabric Geometry Model for 3-D
Du, G.W and Ko, F.K 1992. Analysis of Multiaxial Braid Reinforced Composites. Intern. SAMPE
Warp Knitted Preforms for Composite Metals Conference: Competitive Advancements in
Reinforcement. Proc. Textile Composites in Metals/ Metals Processing.
Building Construction 2nd Inter. Symp., Lyon, Ko, F.K 1988. Braiding, Engineering Materials
France, June 23-25. Handbook, Vol 1, Composites, ed. Reinhart, T.J.
Du, G.W. and Ko, EK 1993a. Unit Cell Geometry of Metal Park, OH: AMS International, pp.
3-D Braided Structures. J. Rein. Plas. Compo 12 519-528.
(2): pp. 752-768. Ko, EK, Whyte, D.W and Pastore, CM. 1988a.
Du, G.W and Ko, F.K. 1993b. Analysis And Design Control of Fiber Architecture for Tough Net-
Of 2-D Braided Preforms For Composite Shaped Structural Composites. MiCon '86:
Reinforcement. Proc. ICCM-9, Madrid, Spain, Optimization of Processing, Properties and Service
July 12-16. Performance Through Microstructural Control,
Fukuta, K, Aoki, E. and Nagatsuka, Y. 1984. 3-D ASTM STP 979, eds. Bramfitt, B.L., Benn, RC,
Fabrics for Structural Composites. 15th Textile Brinkman, CR. and Vander Voort, G.F.
Res. Symp., The Textile Machinery Society of Philadelphia: ASTM pp. 290-298.
Japan, Osaka, Japan. Ko, EK and Kutz, J. 1988b. Multiaxial Warp Knit for
Fukuta, K, Onooka, R, Aoki, E. and Nagatsuka, Y. Advanced Composites. Proc. 4th Ann. Conf Adv.
1982. Application of Latticed Structural Composites, ASM International, pp. 377-384.
Composite Materials with Three Dimensional Ko, F.K 1989. Preform Fiber Architecture for
Fabrics to Artificial Bones. Bull. Res. Inst. Polym. Ceramic Matrix Preforms. Ceramic Bulletin 68
Textiles. 131(2)2: 151. (2): 401-414.
Geoghegan, P.J. 1988. DuPont Ceramics for Ko, F.K and Du, G.W. 1992. Processing and
Structural Applications - the SEP Noveltex Structures of Textile Preforms for Composites.
Technology. 3rd Textile Structural Composites Proc. Science and Innovation in Polymer
Symp., Philadelphia, PA. June 1-2. Composites Processing, MIT, Cambridge, MA,
Goswami, B.G., Martindale, J.G. and Scardino, EL. July 16-17.
1977. Textile Yarns, Technology, Structure and Krcma, R 1971. Manual of Nonwovens. Manchester,
Applications. New York: John Wiley and Sons, UK: Textile Trade Press.
pp. 273-337. Li, W, Hammad, H. and EI-Shiekh, a. 1990.
Hearle, lWS., Grosberg, P. and Backer S. 1969. Structural Analysis of 3-D Braided Preforms for
Structural Mechanics of Fibers, Yarns and Fabrics, Composites, Part I: The Four-Step Preforms. J.
Vol 1, New York: Wiley-Interscience. Text. Inst. 81: 491-514.
Kaswell, E.R, ed. 1963. Wellington-Sears Handbook of Lord, P.R and Mohamed, M.H. 1973. Weaving:
Industrial Textiles. New York: Wellington-Sears. Conversion of Yarn to Fabric. Durham, UK:
Ko, EK, Bruner, J., Pastore, A and Scardino, E 1980. Merrow Technical Library.
Development of Multi-Bar Weft Insertion Warp Loos, AC, Weidermann, M.H. and Kranbuchi, D.E.
Knit Fabric for Industrial Applications. ASME 1991. Processing of Advanced Textile Structural
Paper No 90-TEXT-7, October. Composites by RTM. Proc. 5th Textile Structural
Ko, EK, Krauland, K and Scardino, E 1982. Weft Composites Symp., Drexel University,
Insertion Warp Knit for Hybrid Composites. Philadelphia, PA, Dec. 4-6.
Proc. 4th Intern. Conf Composites. McCarthy, S. and Kim, Y.R 1991. Resin Flow
Ko, EK, Fang, P. and Pastore, C 1985. Multilayer Through Fiber Reinforcement During
Multidirectional Warp Knit Fabrics for Industrial Composite Processing. Proc. 5th Textile
424 Textile preforming

Structural Composites Symp., Drexel University, Raz, S. 1987. Warp Knitting Production. Heidelberg,
Philadelphia, PA, Dec. 4-6. Germany: Melliand.
Mohammed, M.H., Zhang, Z. and Dickinson, L. Scardino, EL. 1989. Introduction to Textile
1989. 3-D Weaving of Net Shapes. Proc. 1st Japan Structures. In Textile Structural Composites, eds.
Intern. SAMPE Symp., Nov. 28-Dec. 1. Chou, T.W. and Ko, F.K. Amsterdam: Elsevier,
Pastenbaugh, J. 1988. Aerospatiale Technology. Proc. pp.1-26.
3rd Textile Structural Composites Symp., Drexel Scardino, F.L. and Ko, F.K. 1981. Triaxial Woven
University, Philadelphia, PA, June 1-2. Fabrics. Textile Research Journal 51(2).
Pastore, eM. and Cai, Y.J. 1990a. Applications of Smith, B.E and Block, I. 1982. Textile In Perspective.
Computer Aided Geometric Modeling for Englewood Cliff, New Jersey: Prentice-Hall.
Textile Structural Composites. Proc. 2nd Intern. Spencer, D.}. 1983. Knitting Technology. New York:
Conf. Computer Aided Design in Composite Pergamon Press.
Material Technology, Brussels, Belgium, April Stover, E.R., Mark, w.e, Marfowitz, I. and Mueller,
25-27. W. 1971. Preparation of an Omniweave-
Pastore, eM. and Ko, EK. 1990b. Modeling of Reinforced Carbon-Carbon Cylinder as a
Textile Structural Composites, Part I: Candidate for Evaluation in the Advanced Heat
Processing-Science Model for Three- Shield Screening Program. AFML TR-70-283,
Dimensional Braiding. J. Text. Inst. 81: 480-490. Mar.
Popper, P. and McConnell, R. 1987. A New 3-D Svedova, J., ed. 1990. Industrial Textiles. Amsterdam:
Braid for Integrated Parts Manufacturing and Elsevier.
Improved Delamination Resistance - The 2-Step Tarnpol'skii, Y., Zhigun, I.G. and Polikov, B.a. 1987.
Method. 32nd Intern. SAMPE Symp. Exhib., pp. Spatially Reinforced Composites. Pennsylvania:
92-103. Teknomic. (English translation, 1992).
Potter, K.D. 1979. The Influence of Accurate Stretch Williams, D.J. 1978. New knitting methods offer
Data for Reinforcements on the Production of continuous structures. Advance Composites
Complex Structural Mouldings. Composites, 10, Engineering, Summer, pp. 12-13.
pp. 161-167, IPC Business Press Ltd, July.
TABLE ROLLING OF COMPOSITE TUBES 19
John T. Kanne and Jerome S. Berg

19.1 INTRODUCTION in table rolled tube manufacture.

The most common resin used to coat the
In the field of composites fabrication table
fibers is the epoxy blend family, which is for-
rolling is a major technique for utilizing pre-
mulated for specific product purposes. The
impregnated fibrous tapes in flag or pennant
resin and fiber are combined and advanced
form for tubular structures. The individual
slightly to a selected 'tack' (stickiness) level.
flags become part of the total wall thickness by
rolling the flags around a mandrel. The hard This fiber and resin combination is called
'prepreg'. Prepreg surface tack has an impor-
mandrel provides the support during cure and
tant adhesive quality in table rolling which
defines the inside dimensions of the tube.
permits composite flags to adhere to one
Table rolling is utilized to fabricate a variety
another or the mandrel without slipping dur-
of products including straight tubes usually
under 7.62 cm (3 in) in diameter and up to ing the table rolling operation.
3.66 m (12 ft) long and small diameter tapered Prepreg is offered by specification of resin
type and roll width. FAW from 130-160 g/m2
tubes such as fishing rods, golf shafts and ski
and RC from 30-36% are common in table
~oles. Flags may consist of a wide variety of
rolling. Higher modulus fibers favor a lighter
fib.ers oriented either longitudinally (along the
aXIS of the tube) or offset at a bias angle, hence FAW to ease rolling. The epoxy in prepreg is
catalyzed, so care must be exercised in follow-
the. terms Ilongitudinal and bias flags'. The
ing the prepreg vendor's storage and handling
resm content (RC) and the fiber areal weight
(FAW) of the prepreg define the ply thickness. recommendations. Freezer storage can extend
Since external molds are seldom used for table the. shelf life; therefore, the material generally
arnves on a freezer truck. Because air and
rolled tubes, a variety of polymer compaction
tapes are used to apply an external pressure. moisture are detrimental to the surface tack, it
These tapes provide the external pressure nec- is important to cut and consume the flags as
soon as possible after opening and unrolling
essary to debulk and prevent flag unravelling
~e prepreg. Dry and low tack prepreg can
before cure and to provide some heat driven
compaction during cure. in?uence and ~ggravate flag wrinkles and ply
slIppage, leadmg to voids and dimensional
problems.
19.1.1 FIBERS AND RESIN . Addre~s tack with the prepreg supplier to
Carbon fiber form 234 MPa (34 msi) to 620 MPa find a SUItable resin formulation for the table
(90 msi), glass IE' or 151 type, ararnids, polyeth- roller's manufacturing environment.
ylene and boron are some of the common fibers A partial list of material, equipment and
tooling suppliers commonly utilized in table
rolling is presented in Table 19.1 at the end of
Handbook of Composites. Edited by S. T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 041254020 7 this chapter.
426 Table rolling of composite tubes

19.1.2 DESIGN
Figure 19.1 illustrates the relationship of the
mandrel, prepreg flags and diameters in table
rolling.

19.2 EQUIPMENT
19.2.1 SHEAR/SHEETER
A number of commercially available power
shears have hardened steel blades and
include automatic feed mechanisms for the
material rolls. Safety guards with interlocks
Fig. 19.2 Photograph of prepreg sheeter. (Courtesy
are needed to prevent finger and hand injury of Century Design Incorporated.)
(Fig. 19.2).
The tooling used in this machine is the 'Steel
Rule Die'.
19.2.2 ROLLER PRESS
The steel rule die, the prepreg and an impact
The roller press is a machine used to press sev- sheet (usually a soft plastic like polyethylene)
eral stacked layers of prepreg tape into are passed through the rotating pressure
individual patterns such as tapered pennants. wheels of the roller press. After compaction,

Fig. 19.1 Diagram of tapered mandrel, bias flags and longitudinal flags.
 Equipment 427

the impact sheet is removed to expose a stack

of ready-to-assemble flags which are now
nested between the blades of the die. The roller
press and dies act similarly to a kitchen cookie
cutter (Fig. 19.3). A sharp knife and straight
edge can act as a prototype or for small scale
production.

Fig. 19.4 Photograph of table rolling press.

(Courtesy of Century Design Incorporated.)

like fingers for achieving uniform pressure on

tapered parts. Both flat bed and segmented
bed versions are covered with canvas. This
pad provides sufficient resiliency and friction
Fig. 19.3 Photograph of a roller press including to permit flags to roll without slipping, yet
loading and unloading racks. (Courtesy of Century conform to the mandrel surfaces. A slight
Design Incorporated.) dusting of talcum powder can be used to pre-
vent prepreg from sticking.
Table rolling provides tighter and a more
19.2.3 ROLLING TABLE
uniform compaction of plies than hand rolling.
Originally, in the 1940s, prepreg flags were
hand cut and hand rolled like cigars to pro-
19.2.4 VERTICAL TAPE WRAPPER
duce the tubular structure. The fishing rod
industry was probably first to commonly use A variety of plastic and/or cellophane tapes,
rolled tubular structures. 1.27-2.54 cm (0.5-1.0 in) in width are used to
In the 1960s several devices with a mobile compact the table rolled plies of prepreg.
lower platen were developed. After activating Machines used to apply these tapes must per-
the machine, a pivoted upper platen is low- mit tape tensioning to debulk the product as
ered down upon the mandrel and a linear the tape is applied. Some applications call for
motion activator in the lower platen rolls the multiple passes through the tape wrapper to
mandrel into the prepreg flag. These machines increase the tape pressure for better com-
permit pressure ranges to be established, gain- paction. Additional wraps of tape are needed
ing maximum compaction and increasing the for thicker wall structures. Frequently, two
speed of rolling. The pivoted upper platen types of tape may be used: a release tape and a
permits the combinations of parallel tube or secondary compaction tape. Some tapes have
tapered (cone-like) tube rolling. If the mandrel a release backing which can allow a single
is parallel, then the pivot function of the upper pass of tape. However, these tapes are gener-
platen will be unnecessary. Figure 19.4 illus- ally more expensive. Apply the tape as soon as
trates a commercially available rolling table. pOSSible after table rolling to prevent the flags
Current rolling tables include temperature from loosening. Figure 19.5 illustrates a verti-
controlled platens and platens with piano key- cal tape wrapping machine.
428 Table rolling of composite tubes

[Ij,ilal~

Fig. 19.6 Photograph of horizontal tape wrapping

machine. (Courtesy of Century Design
Incorporated. )

19.2.6 MANDREL PULLER

Tubular parts which have been cured over a

hard mandrel are all subject to mandrel extrac-
tion. The mandrel puller generally connects to
a bolt on the larger shank end of the mandrel.
The end of the composite tube rests against a
stationary block shaped to permit passage of
the mandrel but blocking the tubular part.
Fig. 19.5 Photograph of vertical tape wrapping Mandrel pullers are generally hydraulic or
machine. (Courtesy of Century Design pneumatic. Hydraulic pullers offer a con-
Incorporated. ) trolled extraction speed, while pneumatic
pullers are faster and useful in high volume
environments. Mandrel withdrawal is gener-
ally done prior to tape removal. Figure 19.7
shows a pneumatic mandrel puller.
19.2.5 HORIZONTAL TAPE WRAPPER
The type of mold release used, correct size
This machine represents an alternate to the of the stationary block and the mass plus
vertical tape wrapper. It is used frequently for integrity of the cured part must be carefully
longer, heavier parts and also for very flexible evaluated or end crushing of the part will
mandrels such as fishing rods. The mandrel is occur. Also, thin walled tubes or tubes with a
affixed to a chuck or mechanical coupling high degree of longitudinal plies can crack
which rotates the parts while tape is applied. during mandrel extraction.
The rollers provide support for the part while
motion is in place. The single or even dual tape
19.2.7 CURING OVENS
feed spools move with the tape carriage and
return to restart position (Pig. 19.6). Ovens used for curing the composite tubes can
After cure, the wrapping tapes are removed be either electrically or gas heated and of batch
by slitting the tape longitudinally and peeling or conveyorized design. Temperatures rang-
the tape away from the cured part. Wrapping ing from 121-191°C (2S0-37S0P) are most
tapes are then discarded. common for roll forming prepregs. Consult
 Materials 429

19.3 TOOLING

19.3.1 MANDRELS

The mandrels used for table rolling are usually

hardened steel, sometimes aluminum or even
composite. The mandrels are designed to sup-
port the prepreg during rolling and curing and
provide the inside dimensions for the part.
Recalling that mandrels must be extracted in
'mandrel pulling', some negative taper is ben-
eficial.
Mandrel makers are skilled in the art of
Fig. 19.7 Photograph of pneumatic mandrel puller.
metallurgy. They can select the materials and
(Courtesy of Century Design Incorporated.) heat treatments necessary to create a mandrel
resistent to permanent bending. However, in
many instances the mandrels can be
restraightened if damaged. Hard plated man-
the prepreg supplier for recommendations on drels generally provide a longer life since
appropriate cure profiles. Fine tuning of the scratched or dented surfaces will hamper
cure profile is often needed to optimize partic- mandrel removal.
ular roll forming operations and specific
products. Ovens with thermocouples are use-
ful in determining hot and cold spots, which 19.3.2 STEEL RULE DIES
may indicate oven regulation for uniform tem- These dies incorporate multiple blades
perature control. This ensures a uniform gel embedded in a rigid backing (usually marine
within the part. grade plywood) which cut the material in the
roller press. The prepreg tape (up to 20 layers)
19.2.8 CENTERLESS SANDER OR GRINDER is cut between the cutting die blades and a
polyethylene sheet sandwiched between the
The wrapping tapes can leave a series of spiral rotating press wheels. Dies with one piece
indentations approximately 0.5 mm (0.002 in) blades provide the best and most continuous
deep in the composite tube surface. If a cuts (Fig. 19.8(a». Dies which include weld-
smooth surface is desired for cosmetic reasons ments (as in a triangular shape flag) generally
or for geometry requirements the part can be dull faster since the weldments soften the cut-
surface sanded or ground. A centerless sander ting edges (Fig. 19.8(b» Ramps can be used
basically removes a user defined controlled between the cutting blades in die designs to
amount of surface material. A centerless keep the roller pressure off the blade ends. The
grinder provides a more accurate finish die builder can recommend blade height and
dimension. Centerless grinders are common cutting edge type best suited for the task.
for the high precision required for the tip ends
of golf shafts in which a tolerance of ± 0.5 mm
(±D.002 in) is not uncommon. In carbon fiber 19.4 MATERIALS
golf shaft manufacture, these surface finishing 19.4.1 MOLD RELEASES
techniques are also used to tailor the product
stiffness by incrementally removing material Generally, mold releases for table rolling man-
along the shaft length. This changes the shaft drels consist of two components: a primary
stiffness characteristics. mold release which provides a polymer bond
430 Table rolling of composite tubes

19.4.2 FIBERS
a
The prepreg tapes can be made from longitu-
dinal tows or woven tows of the following
fibers: aramid, glass, carbon and boron. All
these fibrous composite tapes can be cut into
flags and pennants needed for the table rolling
Cutting process. However, because of the brittle nature
blades of boron prepreg, boron is most often cut for
longitudinal flags. Refer to the appropriate
chapter for specific properties of these fibers.
b
19.5 TYPICAL PROBLEMS

19.5.1 VOIDS

Voids are caused by entrapped air which is not

evacuated before resin gelation. The presence
of voids reduces the strength bearing capabili-
Cutting
ties of the part, creates stress risers and can
blades
contribute to surface finishing and cosmetic
problems.
Voids are first minimized by working with
the prepreg supplier to assure a high quality
material with uniform resin content and good
'wet-out' of the fibers. Also, the material sup-
Fig. 19.8 (a) Diagram of steel rule die for rectangu-
lar shaped flags. Ramps prevent blade damage. pliers (prepreg and wrapping tapes) must play
(b) Diagram of steel rule die for a triangular shaped a key role in developing a cure profile for the
flags. Ramps prevent blade damage. Weldments specific process and products.
can cause blade dulling. Voids are increased by flag wrinkles which
are indicative of rolling problems. The capabil-
ity to perform void content checks (ASTM
D3171) and photomicrographs of the laminate
to the mandrel surface to prevent adhesion; can be extremely useful to develop and
and a secondary mold release which acts as a improve tube processing. Laminate photos are
slip agent. also very useful in operator training. Few lam-
The secondary mold release is most benefi- inates are completely void free but void
cial in straight or slightly tapered mandrels contents lower than 1% are possible with table
and is reapplied between subsequent mandrel rolling and tape wrap compaction.
turns. A primary release can lose its effective-
ness after several hundred turns and must be 19.5.2 DRY AND DIFFICULT TO ROLL
stripped off and recoated. A variety of quality MATERIAL
mold releases are on the market. The fabricator
should work with the release supplier to Prepreg dryness (lack of tack) can be due to low
develop a coating program for the specific resin content, resin formulation, ambient condi-
application. Silicone based releases should be tions of the manufacturing environment, the
avoided if the tube is subjected to subsequent age or out time of the material. Insufficient tack
bonding and painting. can cause flag movement during assembly,
 Typical problems 431

wrinkles, voids and parts with a poor surface Warm lay-up and rolling tables can help
finish. Resin content and formulation can be increase material rolling ability and are gener-
adjusted to suit the manufacturing environ- ally adjusted for slight material and
ment. Temperature and humidity control are environmental changes. Off angle plies are dif-
very helpful in maintaining consistent material ficult to roll adjacent to the mandrel and the
tack in the manufacturing shop. Avoid leaving difficulty is magnified by the higher modulus
cut patterns exposed since moisture in the air fibers. Tack tape is a narrow strip of reinforced
greatly affects the material surface tack and adhesive designed to aid the adhesion of bias
sometimes renders it useless. Consuming the plies to the mandrel. Also, solvent based 'tack
material within two days is a good rule to resins' can be applied to the mandrel to ease
follow. application of the first ply. Once the first ply is
.tightly rolled, however, the material tack is
sufficient for subsequent flags.

Table 19.1 Table rolling equipment, material and tooling suppliers in USA

Equipment Tooling
Century Design Incorporated. Mandrels
3635 Afton Road Lynco Grinding Corporation
San Diego, CA 92123 5950 Clara Street
(619)-292-1212 Bell Gardens, CA 90201
(213)-773-2858

Materials Toray
5729 Lakeview Drive, NE
Prepreg Kirkland, WA 98083-2548
Newport Adhesives and Composites (206)-827-9029
1822 Reynolds Avenue
Irvine, CA 92714 Cytec Engineered Materials, Inc.
(714)-253-5680 1440 North Kraemer Boulevard
Anaheim, CA 92806
Fiberite (714)-666-4349
4300 Jackson Street
Greenville, TX 75403 Chemlease
(903)-457-8554 P.O. Box 540083
Orlando, FL 32854-0083
Mold release (407)-425-2066
Frekote Products Dunstone Company, Inc.
Dexter Adhesives and 2104 Crown View Drive
Structure Division Charlotte, NC 28227
One Dexter Drive (704)-841-1380
Seabrook, NH 03874
(603)-474-5541 Steel rule dies
Ontario Die Company of America
Wrapping tapes 2735 20th Street
Flexicon Pacific, Inc. Box 610397
856 North Elm Port Huron, MI 48061-0397
Suite J (810)-987-5060
Orange, CA 92667
(714)-633-9820
432 Table rolling of composite tubes

19.5.3 PART SLIPPAGE DURING CURE 19.5.4 EXPOSED SURFACE VOIDS

The viscosity of the resin drops as the heat of Exposed surface voids after sanding or grind-
cure begins. Occasionally, tapered mandrels ing are indicative of poor rolling practices,
and the constriction of the wrapping tape dur- insufficient lamination pressure and question-
ing the cure can force a part to slip down the able material.
mandrel. Golf shaft design is highly depen- Exposed surface voids are sometimes
dent on mandrel reference position for proper referred to as fiber pulls', which have a wood
I

stiffness and geometry requirements. Slippage grain appearance on parts with longitudinal
can first be minimized by designing a short surface plies.
semi-parallel section in the mandrel (as in the
butt section of the golf shaft). Slippage is also
19.5.6 LONGITUDINAL PLY WAVINESS
reduced by overwrapping the tapes onto the
mandrel at both ends to secure the part. In Tapered parts with longitudinally oriented
addition, the cure profile or the mold release fibers are prone to zones with a wavy or 'fiber
can be adjusted to limit slippage. wash' appearance. The problem is amplified
with multiple taper mandrels and very low
viscosity prepregs. Cure profile modifications
or alternate resins can reduce the tendency of
Ifiber wash'.
RESIN TRANSFER MOLDING 20
Lihwa Fang and S.C. Advani

20.1 INTRODUCTION ready for its removal from the mold when suf-
Resin Transfer Molding (RTM) is a closed ficient green strength is attained. Processes
mold process in which matched male and that are based on similar principles include
female molds, preplaced with fiber preform, Structural Reaction Injection Molding (SRIM)
and different versions of vacuum assisted
are clamped to form composite components.
RTM (Figs. 20.1 and 20.2).
Resin mix is transferred into the cavity
through injection ports at a relatively low pres- RTM offers the promise of producing low
sure. Injection pressure is normally less than cost composite parts with complex structures
~d large .near net shapes. Relatively fast cycle
690 kPa (or 100 psi). The displaced air is
hmes with good surface definition and
allowed to escape through vents to avoid dry
appearance are easily achievable. The ability
spots. Cure cycle is dependent on part thick-
t~ consolidate parts allows the saving of con-
ness, type of resin system and the temperature
of the mold and resin system. The part cures in Siderable amount of time over conventional
the mold, normally heated by controller, and is lay-up processes. Since RTM is not limited by
the size of the autoclave or by pressure, new

Mixing Head
Resin
Catal st

r-- Vent Port

I
I
I Peripheral
I ,.--....--...Gasket
I
I
I
: Pump Unit

Fig. 20. 1 Schematic of the RTM process.

Handbook of Composites. Edited by S.T. Peters. Published

in 1998 by Chapman & Hall, London. ISBN 041254020 7
434 Resin transfer molding

POLYOL

MIX AND METER

o HIGH POUR PRESSURE o HIGH POUR PRESSURE

o LOW POUR Pltl!SSURI! o LOW POUR PItI!SSURE
o IIYIlRAUUC I'ItESUUItE HIGH
o IIYIlRAUUC PRESSURE LOW
OREClltc.

HEAD
HYDRAULICS
MACHINE MODE
o
Atrro
OMAN.
o CALlI.
o WIlI!ICBND
HEAD

Fig. 20.2 Basic construction of a SRIM machine.

tooling approaches can be utilized to fabricate Advantages are:

large, complicated structures. However, the
development of the RTM process has not ful- • Class A surface: Surface definition is supe-
rior to lay-up. In addition, using matched
filled its full potential. For example, the RTM
tools for the mold, one can improve the fin-
process is yet to be automated in operati.ons
ish of all the surfaces.
such as preforming, reinforcement loading,
demolding, and trimming. Therefore, RTM • Close tolerance: Parts can be made with bet-
ter reproducibility than with layup.
can be considered an intermediate volume
• Design tailorability: Reinforcement and
molding process (Krolewski, 1990).
combination of reinforcements can be used
Several unresolved issues in RTM encoun-
to meet specific properties.
tered by composite engineers are in the areas
• Fast cycles: Production cycles are much
of process automation, prefo~g, to~ling,
faster than with layup.
mold flow analysis and resm chemIstry.
• Filler: Filler systems can be used to reduce
During the last decade, rapid advances in
cost, improve fire/smoke performance, sur-
RTM technology development have demon-
face appearance, and crack resistance.
strated the potential of the RTM process for
• Gel coat: One or both mold surfaces can be
producing advanced composite parts. The
gel-coated to improve surface performru:ce.
advantages and associated disadvantages of
the RTM process are summarized. As the • Good mechanical properties: Mecharucal
properties of molded parts are comparable
development of this process is rapid, some of
to other composite fabrication processes.
the disadvantages may be overcome by the
advances made in this technology. • Good moldability: Large and complex
shapes can be made efficiently and inexpen-
 RTM process 435

sively. In addition, many mold materials can design stage. The usefulness of such design
be used. tools is discussed in detail, giving the relevant
• Inserts: Ribs, bosses, cores, inserts and spe- advantages and disadvantages.
cial reinforcement can be added easily.
• Labor saving: The skill level of operator is
20.2 RTM PROCESS
less critical.
• Low tooling cost: Clamping pressure is low The RTM process can be viewed as seven unit
compared to other closed mold operations. operations. The general practice and process-
• Low volatile emission: Volatile emissions ing issues are described for each unit
are low because RTM is a closed mold operation.
process. The worker is not exposed to
chemical vapors as with the lay-up process.
20.2.1 FIBER REINFORCEMENT
Disadvantages are:
Selection of the proper reinforcement type
• Mold design: The mold design is critical should take into consideration loading condi-
and requires good tools or great skill. tion, part geometry (size, thickness),
Improper gating or venting may result in mechanical properties and surface finish. The
defects. quantity of parts demanded also determines
• Mold filling: Control of flow pattern or the selection. The reinforcement normally car-
resin uniformity is difficult. Radii and ries 90% of the load in a composite and
edges tend to be resin-rich. provides over 90% of the stiffness. The rein-
• Properties are equivalent to those with forcement in a composite can be designed to
matched-die molding (assuming proper match the strength requirements of the part.
fiber wetout, etc.), but are not generally as The following characteristics should be con-
good as with vacuum bagging, filament sidered when selecting fiber reinforcements:
winding or pultrusion.
Volume fradion: ratio of the volume of a
• Reinforcement movement during resin
given mass of reinforcement to the volume of
injection is sometimes a problem.
the same component after molding;
In the following sections, the resin transfer
Wash resistance: ability of a reinforcement to
molding process is discussed in terms of the
withstand movement due to fluid motion or
unit operations involved, to familiarize read-
solvation of the reinforcement binder by the
ers with the basic steps of the RTM process.
resin;
The discussion covers details such as materials
of construction, mold design, preforming, cur- Wettability: ability of a reinforcement to reach
ing, and demolding. Processing issues are a condition wherein all voids in the reinforce-
mentioned in each individual unit operation. ment are filled with resin;
Relevant variations of RTM such as vacuum
Sizing: most fibers are coated with size for
assisted resin transfer molding and flexible
better wettability and bonding but the size
molding tools are summarized.
may influence the cure kinetics during the
Process physics is described with emphasis
manufacturing.
placed on the principles that govern the RTM
process; these are applied in the use of com- Most standard reinforcement materials for
puter simulations. Through the design tools composites can be used, but fiberglass, carbon
such as simulation codes for mold filling and aramid are the most common in RTM.
analysis, engineers are able to predict or diag- One requirement is that the reinforcement
nose the problems in gating and venting in the should hold its shape during the injection
436 Resin transfer molding

phase. Therefore, the reinforcements are gen- mally have 4-{) wt. % of thermoplastic binder
erally stitched, woven or bonded together. added. They are thermo formable and can
Reinforcement build-ups in certain areas can therefore be used for highly complex shapes or
easily be included. For example, woven roving when the anticipated volume of production
and fabric can be combined with continuous makes them economicaL
strand mat and chopped strand mat in appli- Different sizings can be obtained on many
cations where higher strengths are required. reinforcements. Sizings can be tailored to the
Hybrid systems composed of high perfor- type of resin system, Sizings are available that
mance reinforcements such as carbon fiber and are compatible with epoxy, vinyl esters or poly-
aramid fiber can also be incorporated in RTM esters. The strength variation with type of
laminates. Surfacing materials called veils can sizing can be as much as 20%, so this factor
be used in the preforms to hide the imprint of needs to be considered in the choice of rein-
fibers, for improved surface finish. Another forcement.
application of surfacing veil is to achieve a
resin-rich skin to improve corrosion resistance.
20.2.2 PREFORM
Stitched fabrics (Fig. 20.3(a» reduce stresses
inherent in the woven roving design and lead For a flat part, the preform can be as simple as
to higher compressive strengths in the com- a stack of reinforcements that fit in the mold
posite. However, other constructions such as cavity. As preforms become more versatile,
8-HS style of weave (eight-harness satin various means of producing preforms are
weave) in Fig. 20.3(b) have been used because available. Currently cut-and-sew is commonly
of improved wetting characteristics and com- used to assemble preforms of various shapes
pressive strength compared to bidirectional for aerospace applications, Other near net
woven fabrics. shape techniques include braiding, spray-up
Continuous strand mat is multi-stranded, and thermoforming (Fig. 20.4).
laid in swirled configuration. The mats nor-

'.
~
.~ ~ ~

,- ,~ .~ .~ ',," '~
l'

" .-.;"
~ ~,
~
y . , ,':, "',,)0, .;
-' - ", ~, ~-

"".
~
'''',Jl.~~ :--
....
,j....
",' ij
'-:-
'
"'W .::;: ~,
-.ii:''''
-;"
':
~~ ., ."
'"-, '" .f

-'" ,..,
~.' ,';.
-~: .;
'",
";1 '~
-<-~- ":, ~....; :;0:':,::-;
's ,.,'" '":-;:y
-, .,~ >'"
~- ..'f ~' : <' ''
~
• '-. ;r \') :...\;-.' .. --';;'."":"
- ,,;:: ~F -~-
", .lLc ",,;,":. t,- 1;;--,0-
" .•-L

~ ~
-

(8) (b)

Fig. 20.3 (a) Stitched fabric; (b) eight-harness satin weave,

 RTM process 437

blank preparation unloadingltrimming

heating zone stamping zone
zone zone

(a)

-
heating/curing
resinlbinder'--'"

cooling

demolding forming
screen

(b)
directed fiber preform

Fig. 20.4 (a) Four station thermoforming preformer; (b) spray-up process.
438 Resin transfer molding

If prefabricated preforms are not used, then Edge definition: The edges of the composite
some means must be found to hold the layers will be resin rich if the preform is not cut to fit
of reinforcement together as they are built up closely to the edge of the cavity or inserts;
on the tool surface. For example, unidirec-
Fiber distribution: Uniformity of fiber content
tional reinforcement is subjected to washing
in preforms, without excessive thinning, wrin-
(washing is unplanned reinforcement move-
kles or folds, is important;
ment due to resin movement) if proper
precautions are not taken to prevent it. To Permeability: A measure of resin distribution
improve conformance of fibers, a tacky resin into the cavity. This quantity is also affected by
(e.g. epoxy), dissolved in suitable solvent (e.g. fiber volume fraction.
acetone), can be used as a spot glue to hold the
reinforcement layers together. The tacky resin Prefabricated preforms can be further bonded
will be washed out during the resin injection together, with or without a core, to achieve
cycle and will not interfere with the cure of the part consolidation. For structural composites,
part. Sometimes veil can be used to hold the this eliminates the need for fasteners and
layers and prevent washing. adhesives to assemble discrete parts. New
Advantages are: thermoformable reinforcement mats can be
used for highly complex shapes or when the
• Fast loading: Preforms allow fast loading of
anticipated volume of production makes them
the mold.
economical (Carvalho, 1991).
• Precise fiber placement: Preform placement
Design of preforms should go hand in hand
can be made precisely without misalign-
with part design. For example, preform comers
ment. This allows high quality, close
are sensitive to radii of the shape. Figure 20.5
tolerance composites for advanced applica-
shows the thickness reduction of preform over
tions molded by the RTM process.
different radii. The preform thickness does not
• Net shape preforms: If thermo formable
reinforcements are used, the stamped pre- change appreciably compared to those around
forms have excellent dimensional stability. the comer when the radius is made larger than
a critical value. However, if the radius is less
• No additional tool: For low production vol-
ume of the composite, the tool for
preforming can be the same as the tool for 0.10 r---.....----."..,....--.......
,, -__._----.
molding. ,,
. , .
0.08 ' : Vetrotex CertainTeed
The only disadvantage with use of preforms is --------f------ -f U750
that there is an additional unit operation. ,,
With the obvious advantages, use of pre- 0.06 --.. ~----~ .. ----- ..
,, ,
-I --------r-.,, . . . . ---r . -------
forms is advisable when volume of production
allows their economical use. When designing : I l
0.04 ____ ...... _~ ________ ~ __ .. ____ ~ ________ ~ _______ _
I

fiber preforms, following issues should be

: : : :
considered: [] exp~riment~1 ! !
0.02 ---- ......... -f------.. -rOo ....... -_.. ~ -- ---- .... ~ .. -- ..... ---
Comers: The fiber in the bent comer of a pre- ,, ,,
,
form tends to move to the inside of a radius. ,
This can cause channeling of flow that leads to 0.0 L..-_...!...._---!.._----':-_...!...-:::::::....c
0.5 0.6 0.7 0.8 0.9 1.0
poor mold filling patterns;
Drapability: This characterizes the ability of a RlRm
fabric or reinforcement mat to conform to con- Fig. 20.5 Effect of corner radii on preform thickness
tours of the tool; (Rm is the recommended radius).
 RTM process 439

than this critical value, dramatic movement of (Fig. 20.6) is necessary to conform to the con-
the fibers to the inside of the radius occurs. As tours of the tool. This drapability problem,
a result, channeling becomes dominant in the therefore, has a two-fold significance in RTM.
mold filling stage and induces irregular flow Because of the fiber rearrangement, the
patterns. nonuniformity of fiber distribution should be
The edge of a preform is another source of accounted for in the design of the composite.
the race tracking of resin. In order to avoid Fiber volume fraction and orientation are no
the channeling effect, the preform edge longer that of the unreformed reinforcement.
should be cut to fit the edge of the mold cav- Further, such preforms exhibit different char-
ity. The task of obtaining a good edge acteristics to resin flow. Designers should
definition is normally difficult because of the account for this change in determining the
bulkiness of the layers and inter-layer move- location of vent ports relative to an injection
ment (sliding, rotating) during the mold port.
closing when prefabricated preforms are not In practice, to modify the permeability of
used. preforms, various flow inducing media or
Preforming of fabrics over tool geometry mechanisms have been suggested. Application
other than simple flat type will induce shear of such high porosity to the preform or inclu-
deformation in the fiber reinforcements. For a sion of a runner system in mold design can
biaxial reinforcement, shearing of the weave alter the mold filling pattern.

Fig. 20.6 Draping fabric on a mold causes shearing of the weave.

440 Resin transfer molding

20.2.3 RESIN SYSTEM AND INJECTION Low viscosity: High viscosities can cause
mold pressures that are too high in both the
The resin used in the RTM process forms the
mold and the injection unit. Raising the tem-
matrix in the composite after solidification.
perature of the resin system is effective to
The solid structure is a result from polymer-
lower its viscosity, but pot life may be
ization. To select the resin system, one must
adversely affected.
take into account of the rheological change
and resin cure kinetics. Sufficient pot life: This is the time it takes
The formulation of the resin system the resin system's viscosity to reach a level
depends on many factors. For example, the that no longer be comfortably handled by the
resin system can be combined with promoters, equipment.
fillers, internal mold releases, pigments, etc.
Tg point: The glass transition temperature
Typical fillers, such as clay or calcium carbon-
snould be as high as possible. As a rule of
ate, may reduce cost. The optimum viscosity
thumb, the glass transition temperature should
for RTM should be less than 500 cP s. Mixing is
be at least 30°C (50°F), and preferably 55°C
normally required to form a suspension.
(lOO°F), higher than the service temperature
Properties requirements (mechanical,
chemical, fire retardancy, etc.) can also affect Toughness: Toughness in a resin system is
resin selection; the resin mix can be formu- exhibited by its tensile elongation. If sufficient
lated to meet specific needs. Attributes to look damage tolerance is required, the elongation
for in resin systems are: should be at least 3%.
• consistent reactivity; Young's modulus: This modulus must be over
• ability to wet out the reinforcement; some threshold value or the composite com-
• rapid cure after gel. pression strength will be less than the
optimum value. A high tensile modulus is
The ester-type resin mix is combined with an
required to adequately support the fiber rein-
appropriate catalyst, such as emulsified BPO,
forcement and prevent premature buckling.
MEKP, cumene hydroperoxide, at the mixing
head and transferred into the RTM mold. Low The effect of the resin system on hot-humid
profile additives have been developed espe- performance is important in the composite
cially for polyester resins to improve surface part. The modulus of a typical resin remains
appearance. In addition, epoxies, urethanes, essentially constant until the temperature is
vinyl esters, nylon and other hybrid resins are close to the ultimate Tg when it falls off to
available for RTM. The newer resins may zero.
require modifications to the pumping/ injec- Under wet conditions, the strength of the
tion unit to meter and condition the resin mix resin usually falls off at the same rate as the
prior to injection. These new systems offer a modulus because of the effect of absorbed
range of cost and performance options for the moisture. Absorbed moisture plasticizes the
RTM process. resin matrix and lowers the strength of the
Influencing parameters are viscosity, pot composite in non-fiber dominated directions.
life, tensile modulus, glass transition tempera- The amount of moisture absorbed by the resin
ture, tensile elongation and moisture matrix should be small, normally less than
absorbance. In considering a new resin system, 2%. This limits the amount of mechanical per-
the choice of the proper resin system for RTM formance degradation at elevated
must satisfy the following system criteria. temperatures.
Failure to meet these criteria usually means One final topic to consider is the injection of
that the resin system is impractical for RTM. the resin system (schematics shown in Figs.
 Mold materials 441

20.1 and 20.2). Items to control in the resin mix

to assure a consistent, smooth running process
include:
• resin mix temperature;
• ratio of catalyst or curative to resin;
• resin mix viscosity;
• amount of air entrained in the resin mix.
Presence of air in the mix can lengthen the
a) ____________ ____________
gel time!induce porosity in the composite
~~ ~

and! or affect the mix viscosity.

Most successful production resin transfer
molding operations are now based on the use
of resin! catalyst mixing machinery using pos-
itive displacement piston-type pumping
equipment for accurate control of the resin to
catalyst ratio. Back pressure at the mix head
may change when a mixed resin is injected
into a cavity filled with the fiber reinforce-
ment. Static mixers greatly simplify the b
process and are easily cleaned at the end of the
injection cycle. A static mixer sends the pro-
portioned resin and catalyst through flexible Fig. 20.7 (a) Matched mold with rigid halves;
hoses to an injection head employing a (b) matched mold with a flexible mold half.
motionless mixer to thoroughly blend the
materials together immediately prior to injec-
tion step. Gate and vent: This critical part of the mold
design should allow complete wetout with
minimal resin wastage.
20.2.4 MOLD
Mold sealing: A perimeter gasket is necessary
RTM mold design and construction is the most to keep void content low. Tight sealing is
critical factor in successful resin transfer mold- important when vacuum is used.
ing. The mold must be constructed so that
resin reaches all areas. RTM molds require spe- Heating/cooling: A typical RTM cycle consists
cial considerations compared to other of a wide range of temperature for initiating
composite tooling. Figure 20.7 shows two pos- the chemical reaction, curing and final
sible configurations in RTM processing. The demolding. Hence, proper heating! cooling
mold must be designed to account for the fol- channels need to be designed.
lowing factors:
20.3 MOLD MATERIALS
Mold materials: The material of construction
dictates life cycle of mold, temperature control The low pressure requirements of RTM allow
and press requirement. the use of more types of mold materials than
can be used in other composites manufactur-
Cavity design: The RTM mold should consoli-
ing. The choice between metal molds and
date as many assembly steps as possible. A good
polymeric composite molds is chiefly one of
design should take advantage of this ability.
volume and processing temperatures. High
442 Resin transfer molding

volume and high temperatures dictate metal steps. Therefore, the mold designer should
molds. incorporate this rule in the design of the mold
Steet the most suitable mold material, pro- cavity. Instead of joining several substructures
vides superior face life. Aluminum is good for or onto a major structure after molding, it is
construction of prototype molds since the structurally more effective and efficient to
metal is easy to machine, is lightweight and incorporate them into the part before fabrica-
has a reasonably high heat transfer rate, but tion. This can be easily achieved by joining
also galls easily. Cast aluminum and spray- substructure preforms when practical.
metal tooling are currently available and can In production the number of molds or cavi-
be used for higher volume applications. Cast ties required is determined by needed
copper alloys are being considered for use in throughput. This should take into account the
RTM molds due to the potential for increased cycle time. For small parts, the designer can
throughput via heat management and better incorporate several cavities in a mold.
durability. High surface quality with excellent dimen-
Composites, for example reinforced poly- sional control can be achieved by
ester and epoxies, are most frequently used for electroplating the mold face with nickel. The
making RTM molds. They can be expected to appearance surface of a part is usually placed
last for approximately 2000 parts (Isorca, on the bottom of the mold. Pinholes are more
1992). Higher production volumes may justify likely to collect on the top surface. Mold
the use of higher cost spray-metal or metal preparation is similar to that used for hand
tools. lay-up. Anew tool must to be treated with sev-
In some cases, the mold must be backed up eral coats of release agent.
in order to maintain its shape. Conventionally, New mold materials provide flexibility in
the backup can be done cost-effectively with mold design for RTM. For example, to demold
core material or steel frames to add rigidity to a part with vertical sides, it is common to
the cross section and to support composite allow several percent draft in the vertical
mold faces. The closure of the mold is dimension. Flexible silicone rubber has been
achieved by mating of the mold surfaces used for RTM molds in the form of a bladder
against a perimeter gasket. Therefore, guide mold half which is capable of being inflated or
pins are usually employed to align the mold deflated depending on the process require-
halves both laterally and vertically to keep ment. During mold filling, the flexible mold
resin from leaking. wall is pressed against the rigid wall by inflat-
Advancement in adapting composite tool- ing the bladder with a pre-determined
ing to the needs of RTM is underway. For pressure. During the injection cycle, the mold
example, lengthening the life of the composite can deform to enhance resin flow. Upon com-
tool face is desirable and effective to maintain pletion of mold filling, the flexible tool can be
quality while keeping costs low. The factors further inflated to consolidate the composite
that cause deterioration of the mold face are component. Part removal in this case is easy
temperature fatigue and attack by solvents or since the flexible half can be deflated. This
mold release agents. An electrolytically or technique allows fabrication of complicated
vapor deposited nickel shell is a new tech- parts that are not ordinarily possible to
nique that will extend face life. demold.

20.4 MOLD CAVITY DESIGN 20.5 INJECTION PORT AND VENT DESIGN
One of the most important design rules for The injection port allows the resin to be trans-
RTM parts is to reduce the number of assembly ferred into the mold (Fig. 20.1) and its design
 Heating and cooling design 443

may be critical. The location of inlet ports must when vacuum is used. Sealing the mold to
allow the resin to reach all areas without achieve cavity pressure of 690 kPa (100 psi) or
bypassing part of the reinforcement. Air vents higher is necessary if the void content of the
help control internal pressure, bleed out air part is to be kept low. The only practical way to
and provide a visual indication of mold filling. accomplish this is to use O-rings. Machining
Race tracking, or channeling, in the mold is the face of the mold to close tolerances is pro-
usually the reason why the resin bypasses hibitively expensive. It is also usually
areas of the reinforcement. Since the resin will impossible to maintain the mold absolutely flat
not flow backwards, this tends to create dry to achieve a metal-tight seal.
patches. The engineering way to ensure com- O-ring design is well established. The slot
plete initial wetout is to gate the mold has to be cut so that the O-ring can deform
correctly in the design. This may be difficult when the mold is closed and maintain a seal.
even for an experienced mold designer. Use of Either square or round O-ring grooves can be
computer simulations as a design tool has used. The type of O-ring material used
become popular in conventional injection depends on the maximum temperature the 0-
molding. Without an engineering design tool, ring will experience during the fabrication
gates and vents can be put in the mold after cycle and the type of solvent used to clean the
molding some trial parts, but many trial runs mold. Nitrile rubber material can be used sat-
may be prohibitive in some applications. In isfactorily up to 120°C (250 0 P). Over 120°C
the next section, new engineering tools silicone rubber can be used to temperature
adapted for RTM mold filling will be dis- approaching 177°C (350 0 P). If help is needed
cussed to overcome the problem. in sealing around inlet or outlet tubes, tacky
Mold designers have found that RTM sealant can be used. This type of sealant is use-
molds must be vented to allow the air within ful for making an O-ring where grooves do not
the mold to be pushed out by the resin. Gate at exist.
the lowest point and vent at the highest point
is generally a good design practice.
20.7 HEATING AND COOLING DESIGN
Experienced designers may use symmetry to
design the inlet ports and outlet vents to The mold should have good temperature con-
remove entrapped air. Venting ports must be trol. The RTM mold should be able to heat and
placed to draw the resin through sections of cool the part during the fabrication cycle. Most
the part that are difficult to wet out. They are resin systems cure faster at elevated tempera-
best placed at dead ends where the resin tures. During demolding, lowering the
would not flow by itself. temperature is sometimes helpful in removing
After the resin has finished bleeding, both the part. Even molds that are intended for
injection and venting ports must be sealed off. room temperature-cured resins should be well
This allows pressure to build up in the mold, insulated so that environmental conditions do
and forces the resin to further wetout other not change the gel times and viscosity of the
sections of the part. This packing stage allows resin. Some molds are heated or designed to
the part to gel under pressure, decreasing void go into ovens to achieve faster cures at higher
content in the finished part. temperatures.
Normally, the mold is heated and cooled
using either hot water or oil. The mold is con-
20.6 SEALING THE MOLD
structed to allow the heating/cooling fluid to
The perimeter gasket seals the edges of the flow through channels (Pig. 20.8) in its interior.
mold to prevent loss of resin and injection pres- The fluid is heated and cooled by conventional
sure. In addition, it is an absolute necessity means, such as a gas-fired heater and heat
444 Resin transfer molding

heating
//////////////
channel - - - " Too hh

top mold
platen ~~Ia z

resin 0000000 I
flow
-OOKOOO
Fig. 20.8 Heating/cooling by flow channels in the RTM mold.

exchanger. For larger molds, the heating and mold is filled, the pumping system is shut off
cooling times will be longer if the heat transfer and immediately flushed, and the part is
area does not increase in proportion to the allowed to cure.
weight. At some point, the production cycle Successful configurations demonstrated in
time becomes limited by the rate at which heat the industry show a common factor: that is,
can be added or removed, and becomes inde- the flow of resin is symmetrical about the vent
pendent of the curing characteristics of the ports, in a manner such that the volume of air
resin system. left in the reinforcement decreases. This com-
Under development is low thermal inertia pression effect helps sweep the remaining air
technology that allows the tool face to be out of the part.
heated by electric wires buried in the face. The When the flow path is arranged in such a
construction of the mold face is such that the way that the resin flows into a configuration
heat flows into the mold face and not outward with increasing volume, there is a tendency to
toward the mold support structure. This is bypass part of the reinforcement. This situa-
accomplished by use of a foam core that insu- tion can happen when core material is used.
lates the bulk of the mold from the tool face. For example, when there is reinforcement on
This novel technology, if successful, will allow either side of a core, it is possible that slight
a more instantaneous transfer of heat where it misalignment in the core thickness will cause
will do the most good - at the mold face. dry spots in the part. To overcome this prob-
lem, the resin must be introduced on either
20.8 MOLD FILLING side of the core simultaneously. Holes may be
drilled through the core to allow the resin sys-
Resin injection is to pump the base resin sys- tem to flow to the other side. When this is
tem to a mixing head through either a single or done, the core floats on the wet reinforcement
two pot system. Impingement mixing of the
and equalizes itself.
components occurs in the mixing head. The When the injection pressure is too high or
catalyzed mix is then pumped through a static reinforcement tends to move in the mold, the
mixer which completes the mixing of the two following remedies must be considered:
components. The injection nozzle is attached
to the injection port on the mold and the resin Multiple gates: partition the mold along the
system is injected into the mold to pack the flow path such that travel distance for resin is
mold to a predetermined pressure. When the reduced.
 Curing 445

Runner system: allows the delivery of resin to ing (Mosher, 1995). Note that the tooling must
various parts of the reinforcement quickly be large enough to accommodate the perime-
without using high injection pressure. ter gasket, air vents, injection ports and guide
pins.
Flexible mold wall: allows the deformation of
the bladder wall to facilitate mold filling.
20.9 CURING
There are several techniques to modify the
flow patterns. Application of high porosity To convert a resin system into useful products
media on the preform or inclusion of a runner it must be cured or cross-linked by chemical
system in mold design can alter the mold fill- reaction into a three dimensional network. The
ing pattern. This is helpful in reducing reaction usually involves either a step growth
injection pressure or displacing air. polymerization, a chain growth polymeriza-
All resin movement must be accomplished tion, or a combination of both. The
within the time allowed before the onset of accompanying rheological change in the
gelation. Additionally, the resin injection process is shown in Fig. 20.9 (Macosko, 1989).
process should not cause movement of the The curing step constitutes a major portion
reinforcement and should be done at low pres- of a typical RTM cycle. During curing, rheo-
sure so that the mold will maintain its shape logical property changes of the resin system
without requiring massive backing. and heat transfer between the mold wall and
Vacuum may be used to facilitate filling the the resin dictate the cure cycle.
mold and simultaneously assist in removing Simultaneously, modulus and strength begin
air from the laminate. This requires good mold to build up at a rate depending on the type of
sealing and the use of a vacuum pump. resin and catalyst used and the chemical kinet-
Vacuum up to 740-760 mm Hg (29-30 in Hg) ics of the resin system. Curing can continue
has been reported in assisting RTM mold fill- after the part is demolded.

Matrix

FILLING CURING POSTCURE

~ ______________ ~ __________________ ~Time

Tfill Tgel 'fcure

Fig. 20.9 Rheological change during the curing process. (Reproduced from Macosko.)
446 Resin transfer molding

Cure cycle is dependent on part thickness, If the adhesion to the mold face is too strong,
the ratio of catalyst or curing agent to resin even exceeding the strength of the composite,
and the temperature of the mold and the resin it can be reduced by spraying release agents,
system. In some cases, the part is removed normally fatty ester soaps or waxes, on the
from the mold immediately after gel occurs. mold surface.
The part must develop sufficient green After the two mold halves separate, the part
strength for handling prior to its removal from can be removed from the cavity. Part removal
the mold. Green strength is the strength a com- methods range from the use of plastic/
posite exhibits after the resin gels, but prior to wooden wedges and rubber mallets to the use
complete cure. Gel time is the interval of time of knock out pins. A mold designed for low
between introduction of catalyst or curing throughput with hand operated clamps pro-
agent to a thermosetting resin and the forma- ducing a relatively simple, lightweight part
tion of a gel. Typical gel times range from would most likely be removed using a wedge
several minutes to about an hour depending and mallet. Sophisticated hydraulic ejection
on the factors mentioned above. systems can be used for high volume, complex
The glass transition temperature, T , for an or heavy parts. To be pushed out, the part
RTM resin system depends on thermalghistory. needs enough green strength to survive con-
For a given temperature, the Tg increases dra- siderable bending stresses.
matically with time until it levels off. As the The most common test for sufficient bend-
curing temperature is raised, the T reaches a ing strength is to fold over a comer of the part
steady-state value at a faster rate. the steady- immediately after demolding. If the comer
state value for Tg is a function of the curing survives the bending without cracks or a
temperature, and usually approaches the cur- crease, the part is accepted. Otherwise, mea-
ing temperature. However, the limit is sures to improve its green strength include
bounded by the degradation temperature of any of the following steps:
the resin system.
• allow the part more time to cure in the
mold;
20.10 DEMOLDING AND POST PROCESSING • increase the mold temperature;
The minimum the curing step must accom- • modify or change the resin system, e.g.
increase the catalyst level.
plish is to develop sufficient green strength so
that the part can be removed from the mold. There is often excess resin at the edges of the
While cost is an important factor, it is not the part and in the vents. Considerable trimming,
only criteria in choosing a method to remove a part of the post processing, is common when
part from an RTM mold. For example, part reinforcement is clamped in the parting line.
weight and complexity, and throughput are Trimming is required for almost all items
important considerations. In many ways, the made by the RTM process. Accurate preform
choice of ejection methods parallels the choice placement and precise alignment can reduce
of clamping methods. the labor in this step.
A few precautions are required to facilitate Postcure, one of the post processing opera-
demolding. Before opening the mold halves, it tions, is used for various reasons. A molding
is necessary to release the part from one mold cycle including postcure can increase produc-
surface. The force required is approximately tion throughput. While postcuring in an oven,
that to overcome the adhesive force between the temperature is not restricted to that
the mold and the composite. Typically, tears of allowed for the mold materials. Therefore a
surface skin or flash, both resin rich, can be higher conversion of reactive groups can be
found around the comers or edges of the part. achieved. It can also prevent the reaction
 Process physics and use of simulations as a design tool 447

exotherms of a resin system from damaging a

composite tool. It is important to hold the part
shape during the process of postcuring and
cooling to prevent distortion or warpage.

20.11 PROCESS PHYSICS AND USE OF

SIMULATIONS AS A DESIGN TOOL
The processing defects addressed in the previ-
ous section are often caused by lack of a
systematic treatment in RTM part design and
process planning. Among the unresolved
issues in RTM encountered by composite engi-
neers, those related to the physical processing
have developed rapidly during the last
decade. The advancement in RTM technology Fig. 20.10 Complex structure manufactured by
demonstrates the potential of RTM becoming a RTM.
primary process for producing many compos-
ite parts.
In this section, the issues in reinforcement in the molded part and therefore is of extreme
preforming, alternative tooling, mold flow importance.
analysis, and cure kinetics are revisited. The Tooling and mold construction are critical
focus is on the use of models to describe and factors in successful RTM. By considering sev-
enhance the understanding of the physical eral alternative configurations, both the
phenomena. The models are built on the injection pressure and the filling time can be
experimental evidence and observations, the reduced. These alternative designs are valu-
goal being to reduce the scope of experiments able as the injection pressure tends to rise
in the engineering applications. Reducing rapidly when inhomogeneous fiber distribu-
engineering experimentation is achieved by tions are present as a result of preforming.
combining three elements: mathematical mod-
els, numerical methods and computer
20.11.1 PHYSICS GOVERNING RTM
software, into a simulation. One example of
PROCESSING AND NUMERICAL SIMULATIONS
such software is LIMS (Liquid Injection
Molding Simulation) (Advani et al. 1993) Darcy's law for flow through porous media is
which has been developed specifically for conventionally used to describe the resin flow
mold filling of complex structures in RTM and in the fiber reinforcement. The generalized
can be used also as a design tool for manufac- form expresses the superficial velocity of resin
turing of complex structural parts as shown in flow in terms of a factor, which is permeability
Fig. 20.10. divided by fluid viscosity, multiplied by pres-
The topics will be presented in the order sure gradient. This expression together with
found in the unit operations of RTM. Draping the mass conservation in the mold are solved
of reinforcement plays the role of distributing together using various numerical methods. A
fibers in a way that depends on the tool geom- typical example of this method combines finite
etry. Simulation of reinforcement draping element method with control volume method
allows an engineer to estimate the fiber con- (Bruschke and Advani, 1990a, 1990b, 1991,
tent distribution. This distribution can change 1994; Advani, 1994; Young et al., 1991a, 1991b).
the volume fraction as well as the orientation The solution is moved forward in time after
448 Resin transfer molding

the pressure field is obtained during the filling surface. The length of the cell segment can be
process. changed as a result of slippage to accomodate
The pressure solution obtained from the this effect.
mold flow analysis can be used to position the A dome shaped part will serve as an exam-
gate and vent. This lends a design engineer ple of this draping simulation. First, a square
'infinite' options when facing the task of mold bidirectional mat is draped. The workpiece is
design. The design rules are no longer initially configured so that warp and weft
restricted to the rule of symmetry used by tows are perpendicular to each other. Then
experienced designers to position the inlet and draping starts at an arbitrary point on the tool.
outlet ports. Instead, a composites engineer The initial constraints used in this case study
would be able to optimize the overall design are prescribed along the central tows in both
based on criteria such as minimizing the injec- the warp and weft directions. The length of the
tion pressure. cell segment is assumed to be constant.
In the draped configuration shown in Fig.
20.11, the degree of deformation varies from
20.11.2 PREFORMING
cell to cell. The minor angles in the preform
For bidirectional mats, woven or stitched, range from 90° to a minimum of 35°. The shear
draping an arbitrary tool surface depends on also results in fiber volume fraction increase
two deformation modes: shear deformation up to 70% for the dome. This information can
and inter-yam slip (Potter, 1979). A mat of this assist a designer in material selection, setup of
nature is treated as a net that consists of many processing conditions and part design: a
cells (Van West, 1990). Therefore, draping over process engineer can use this information to
a surface of double curvature requires the net find out where to make necessary cuts in order
to map on the surface by changing the internal to accommodate for induced deformation. As
angles in each cell. The four sides of a cell are a rule of thumb, formability of preform mat
made up of fiber tows. These tows, under the relies on absorption of such deformation by
preforming condition, are inextensible. At the reinforcement material. A good material
high deformation regions in a reinforcement, can withstand high deformation without
slippage may be necessary to drape the tool wrinkle formation.

Fig. 20.11 Draped dome.

 Process physics and use of simulations as a design tool 449

20.11.3 ALTERNATIVE TOOLING or the preform is driven by the pressure differ-

ence. Therefore, the equation of motion is a
One benefit of this process is that it can con- function of the bladder as well as the rein-
solidate several complex three dimensional forcement material (Lucey, 1992). On the
parts into one molded piece. The key to preform part, the compressibility of the rein-
accomplish this is tool design. From the design forcement in the thickness direction plays a
point of view, a flexible mold wall is very major role. On the bladder part, factors such as
desirable to mold certain parts with difficult- inertia, damping, and rigidity of the elas-
or impossible-to-demold geometry. While a tomeric material can also be included when
hard tool makes clamping and demolding dif- they are significant.
ficult, the flexible mold provides a convenient From lateral compression tests, the
alternative for mold design of these types of load-deflection curve of the fiber reinforce-
parts. Figure 20.10 shows an example of possi- ment material behaves like a nonlinear spring.
ble features which may be molded using this The elastic constant of the preform depends on
concept. In this part, one can easily see the its state. Preform permeability is a function of
small draft angle and the stiffeners which can fiber architecture and porosity. Since the
make demolding difficult. Moreover, the porosity of the preform changes with the
beads and the flanged opening in the bulk- thickness, the permeability can be expressed in
head of this frame are features that are terms of the cavity thickness.
impossible to mold using conventional rigid Figure 20.12 was obtained from the numeri-
molds. cal simulation of two cases: one with rigid walls
To avoid unconstrained wall movement, and the other with a flexible mold wall. In the
the bladder pressure is higher than the injec- case with rigid mold walls, the pressure drops
tion pressure. The motion of the flexible wall linearly with respect to the flow distance. This is

6xl05
....1::t==t::: :=r-£~ ~f~~~~*~~ ~~~~!~~~~~~f~~~~ ~~~~i~~d~~~~F~
Sxl0s : :

4xl0s

-...
tiJ
a.
CD
:::l
en
3xl0s
en
...
CD
a.
2xl0s

lxl0s

o 0.2 0.4 0.6 0.8

x
Fig. 20.12 Pressure distribution in the 1-0 mold near the end of mold filling for flexible and rigid tool.
450 Resin transfer molding

caused by the constant permeability of the pre- pressure drop and overall filling time which is
form inside the mold. The pressure curve for the impossible to attain simultaneously in conven-
case with a flexible mold wall reflects the fact tional tooling (Fong and Advani, 1995).
that the fluid flow in the filled region exhibits a
smaller pressure drop. This reduction is benefi-
20.11.4 GATING, VENTING AND VOID
cial to the molded parts as it causes less fiber
CAPTURE
washout and preform deformation due to the
resin. In this section, computer simulations for RTM
Figure 20.13 shows the results of computed mold filling are discussed to overcome the gat-
gap thickness of the 1-D mold with a flexible ing and venting problem. Mold filling
mold wall. The straight line shows the thick- simulation is an effective way of positioning
ness in a rigid tool. From this distribution, one injection and vent ports. Gating and venting
can see that the gap height is a function of time are critical in the mold design because they
during the filling process as well as a function determine whether complete wetout is achiev-
of pressure. Near the injection gate, the resin able under normal operating conditions.
pressure balances the applied pressure from A gate designed at the lowest part and vent
the bladder and increases the gap thickness to at the highest point is generally a good prac-
its maximum in the 1-D mold. As a result, the tice to allow the air within the mold to be
resistance to the incoming flow has reduced pushed out by the resin. Experienced design-
significantly as shown in the previous figure. ers may use symmetry to design the inlet
Through the numerical study, the potential ports and outlet vents. However, the picture is
of the flexible tool design has been demon- often complicated by the geometry or the
strated. It has the advantage of reducing the presence of inserts. The engineering way to

0.005 "'-'--'-"""-"--"r--T-'--'-"--"r--T-r-,-..,-.,--"
.....L...___
: _:..
: : :
--t--+--+.

o 0.2 0.4 0.6 0.8

x
Fig. 20.13 Gap thickness variation in the 1-D mold near the end of mold filling.
 Process physics and use of simulations as a design tool 451

ensure complete wetout initially is to gate the the design tool. The situation would be much
mold correctly in the design. more complicated if mold filling is coupled
Figure 20.14(a) shows a square plate with with phenomena such as preform deformation
two cutouts in the part. The injection port is and channeling in the comers and along the
first positioned at the center of the lowest part. edges.
The flow fronts corresponding to the gate At the microscopic level, heterogeneities
design are indicated by the curved contours. always exist in the preform media. For exam-
Contours in this figure indicate different time ple, the fiber tow may have a permeability
steps. For example, the contours closer to the several orders lower than that of the inter-
gate represents area that is filled first and the stices. Therefore, micro-voids form when the
contours closer to the vent the last filled orientation of the fiber tow does not allow the
region. As a result of colliding flow fronts in displacement of the air inside the tow. A novel
the middle and top portion of the part, the fig- approach in mold filling analysis is reported
ure demonstrates the capturing of dry patches by modifying the equation of mass conserva-
or macro-voids. These voids can degrade the tion to account for the fluid absorbed by the
properties of the molded composite signifi- fiber tows (Fong and Advani, 1994). Void
cantly. Void capturing is important in the entrapment inside tows is found to be depen-
process simulation to avoid formation of such dent on the microstructure, the vent pressure,
defects. Figure 20.14(b) shows an alternative and the ratio of the difference in the perme-
design that eliminates venting in the middle of ability of the tows and the permeability of the
the part. As a result of injection in the comer, preform (Pillai and Advani, 1994;
the vent port has to be positioned differently. Ranganathan et al., 1994).
This demonstrates the power and simplicity of

(a) Injection Port (b) Injection Port

Fig. 20.14 Design of injection ports: (a) central injection; (b) comer injection.
452 Resin transfer molding

20.11.5 SENSOR CONTROLLED INJECTION Figure 20.16 shows the pressure calculation
from the mold filling simulation. The pressure
Sensor controlled injection is multiple injec-
drops linearly in the one dimensional flow. As
tion in an 'intelligent' way without involving a
the flow front progresses from the inlet toward
complex control algorithm. It requires place-
the vent under a constant flow rate boundary
ment of gates along the flow path at a number
condition at the inlet, the pressure build-up
of locations. The injection gate is also a sensor
looks like the schematic shown in the lower
capable of detecting the arrival of resin. These
left figure. For the multiple gate with sensor
gates are then activated or deactivated in the
controlled injection, the pressure at the first
order of first on, first off, and, therefore, allow
gate increases up to a limit when the flow front
the mold to fill in a series of steps.
hits the next sensor. When the next gate is
For example, Fig. 20.15 is a simple mold
open, the previous gate is shut off. So the pres-
:-vhich has four injection gates. To help visual-
sure build-Up is only limited by the length of
IZe the concept effectively, a 1-D mold is used.
the interval. Therefore, the maximum pressure
The T-column represents different time stages
seen in the mold is only a fraction of the pres-
in the filling process. In this example, only one
sure compared to the lower left figure.
~ate is allowed to open at a time. As the injec-
tion starts at Tl, the first gate is open and the
remaining three are closed. As the flow front
progresses through the mold, it hits the second
gate location at time T2. The injection unit
shuts off the resin to gate one and opens the
sec?nd injection gate. Instead of having the
resm flow through the whole length in the
mold, the length is divided up into a number
of intervals. Therefore, the overall flow resis-
tance decreases as the effort required is for the
resin to flow from one gate to the next closest
gate in the flow path.
Flow Length Flow Length

CONTROL SCHEME
Fig. 20.16 Representation of pressure during mold
filling.
T1 ON OFF OFF OFF
T2 OFF ON OFF OFF
T3 OFF OFF ON OFF Table 20.1 shows the results from two sets of
T4 OFF OFF OFF ON computer simulations. For either case, only
one gate is open at any time during the mold
filling stage. The first column uses constant
flow rate and the second column uses constant

Table 20.1 Comparison of processing parameters

Gate with flow front sensor

PIF single gate rr

T single gate

Single-gate injection 100% 100%

Fig. 20.15 Control scheme for a I-D model. Sensor-controlled injection 48% 36%
 Process physics and use of simulations as a design tool 453

pressure. If only one injection gate is used, the 20.11.6 MOLD FILLING WIlli RESIN DELIVERY
pressure under the constant flow rate bound- SYSTEM
ary condition will reach a maximum.
Compared to the sensor controlled injection Conventionally, an injection port serves as a
with four gates, the pressure at the gate is only point source where fluid is pumped. The
48% of the pressure reached by the single gate drawback of a point source is that the pressure
injection. In terms of filling time, the two value tends to rise rapidly to an extent that
molds are subject to a constant pressure could be detrimental to the preform. By
boundary condition. Results show that the extending the point source into other forms
mold filling for the single gate injection takes proves to be effective in reducing the pressure
almost three times that for the sensor-con- build-up. To implement this concept, one can
trolled injection. use multiple point sources as discussed previ-
An example is shown in Fig. 20.17, which ously. A line source has been popular in
elaborates on how one can utilize a sensor to vacuum assisted RTM because of its ability to
eliminate a dry spot during molding. In Fig. fill the mold using 1 atm of pressure.
20.17(a), where no sensors are implemented A line source may be modified to serve as a
and the injection gate is at the location as runner by allowing more fiberfree space in this
shown, a dry spot will appear in the middle of delivery system. This is the channeling effect,
the part. However, an extra gate in the middle now used to advantage in mold filling. Further
as shown in Fig. 20.17(b), if triggered at the extending the fluid source may possibly yield
point the fluid reaches the midframe, can pre- a 'plane' source. The actual implementation of
vent this void, as indicated by Fig. 20.17(c). a plane source may include a high-porosity
This feature is incorporated in a numerical layer in the stack-up of the reinforcement
simulation such as LIMS and can be systemat- mats. The layer can possess a permeability
ically studied for a given geometry to decide several orders higher than that of the fiber pre-
the best strategy when in situ sensing capabil- form. The result of this is a three-dimensional
ities are incorporated in the fabrication phase mold flow with fluid propagating rapidly
(Liu et al. (1995». through the spreading plane or surface first

t vent

Fig. 20.17 Use of sensors to eliminate dry spots: (a) no sensor; (b) extra gate sensor; (c) void prevented.
454 Resin transfer molding

followed by percolation of the resin through of Application Support Laboratory, Fiber Glass
the thickness of the preform. For three-dimen- Reinforcements Division, Vetrotex CertainTeed
sional flow, venting the mold may become less Corporation, 1991.
Chou, T.W., 1992, Microstructural Design of Fiber
intuitive. In practice, vacuum assistance can
Composites, Cambridge: Cambridge University
provide part of the solution. Press, UK.
Isorca Inc., 1992, Introduction to Resin Transfer
Molding, Society of Plastics Industry,
20.12 CONCLUSIONS Composites Institute.
Fong, 1., J. XU, and Lee, L,J., 1994, Analysis of ther-
Resin transfer molding is a practical process
moformable fiber mat preforming in liquid
for much of the composite industry. The qual- composite molding: study of deformation
ity of RTM molded parts can equal that by modes and reinforcement characterization,
conventional autoclave processes and its eco- Polym. Comp., 15, 134.
nomic advantages are obvious. Although the Fong, 1. and Advani, S.G., 1994, The role of drapa-
underlying principles of RTM appear at first to bility of fiber preforms in resin transfer
be simple, this is often not the case. The chal- molding, Amer. Soc. Comp., Proc. 9th Tech. Con!,
1246.
lenge for RTM is to bring together the
Fong, 1., and Lee, L.J., 1994, Analysis of fiber mat
disciplines of preforming, mold design and preforming in liquid composite molding, pre-
process development with existing fibers and forming induced effects on mold filling, J. Rein!
resins. This can be best achieved through an Plas. Comp., 13, 637.
understanding of the physics governing RTM Fong, 1., Varma. RR and Advani, S.G., 1994, Use of
and by current simulation technology. process models and simulations as design tools
in molding polymer and polymer composites,
The Pacific Conference on Rheology and Polymer
REFERENCES Processing (PCR'94), Kyoto, Japan.
Fong, 1. and Advani, S.G., 1994, The role of dual
Advani S.G., Bruschke, M.V. and Pamas, R, 1994, permeability of fiber preforms in mold filling
Resin Transfer Molding, in Flow and Rheology in simulation of resin transfer molding, Proc. 1st
Polymeric Composites Manufacturing (Ed S.G. Intern. Coni Compo Engng, New Orleans, LA, 17.
Advani,) Amsterdam: Elsevier Publishers, Ch Fong, 1., Liu, B. and Advani, S.G., 1995, Modeling
12, pp. 465-516. and simulation of resin transfer molding with
Advani S.G., Bruschke, M.V. and Liu, B., LIMS 3.0: flexible mold walls, 50th Ann. Con!, SPI, Compo
Liquid Injection Molding Simulation, User Inst., Session 3-A.
Manual, CCM Report, University of Delaware, Krolewski, S. and Busch, J., 1990, The competitive
Newark, DE 19716. position of selected composites fabrication tech-
Bruschke, M.V. and Advani, S.G., 1994, A numerical nologies for automotive applications, Proc. 35th
approach to model nonisothermal, viscous flow Intern. SAMPE Symp., pp. 1761-1771.
with free surfaces through fibrous media, Intern. Lee, S.M. International Encyclopedia of Composites,
J. Num. Methods Fluids, 19, 575-603. 1991, New York: VCH, 1991.
Bruschke, M. V. and Advani, S.G., 1991, RTM: Filling Liu, B., Bickerton S. and Advani, S.G., 1994,
simulation of complex three-dimensional shell- Modeling and simulation of RTM - venting and
like structures, SAMPE Quarterly, 23(1), 2-11. void formation, Proc. Intern. Con! Compo Engng,
Bruschke, M. and Advani, S.G., 1990, A finite ele- p.17.
ment/ control volume approach to mold filling Lucey, A.D. and Carpenter, P.W., 1992, J. Fluid Mech.,
in anisotropic porous media, Polym. Comp., 11, 234, 121.
398-405. Macosko, c.w., 1989, RIM, Fundamentals of Reaction
Bruschke, M.V. and Advani, S.G., 1990, Mold filling Injection Molding, New York: Karl Hanser
of generalized newtonian fluids in anisotropic Verlag.
porous media, Transport Phenomena in Material Mosher, P., 1995, An introduction to vacuum-
Processing, ASME Trans. HID 132,149-158. assisted resin transfer molding (SCRIMP), 50th
Carvalho, RL., Personal communication, Manager Techn. Coni, SPI, Compo Inst., Session 8.
 References 455

Pillai K and Advani, S. G., 1994, The role of dual Vanwest, B.P., Pipes, RB., Keefe, M. and Advani,
permeability of fiber preforms in resin transfer S.G., 1991, The draping and consolidation of
molding, Proc. 9th Am. Soc. Comp., p. 17. commingled fabrics, Compo Manujng, 2, pp.
Potter, KD., 1979, Composites, Ig 161. 10-21.
Ranganathan, S., Wise, G.M., Phelan, P.R, Jr., Young, W.B., Rupel, K, Han, K, Lee, L.J. and Liou,
Parnas, RS. and Advani, S.G., 1994, A numeri- M.J., 1991a, Polym. Comp., 12, 30.
cal and experimental study of the permeability Young, W.B., Han, K, Fong, L., Lee, L.J. and Liou,
of fiber preforms, Proc. 10th ASM/ESD Adv. M.J., 1991b, Flow simulation in molds with pre-
Compo Conj., 309. placed fiber mats, Polym. Comp., 12,391.
Scheidegger, A.E., 1974, The Physics of Flow through
Porous Media, University of Toronto Press.
Tucker, CL., 1989, Fundamentals of Computer
Modeling Jor Polymer Processing, New York: Karl
Hanser Verlag.
FILAMENT WINDING 21
Yu.M. Tarnopol'skii, S. T. Peters, A.I. Beil'

21.1 INTRODUCTION Filament winding is a natural way to com-

bine two-dimensional reinforcement and, with
A winding operation is the basic fabrication
additional processes and devices, three-
technique for forming load-bearing structural
dimensional reinforcement. Advanced
elements made of polymer matrix-based
pr~e~ses, combining filament winding and
fibrous composites, which have the shapes of
braldmg, allow fabrication of spatially sewn
bodies of revolution. A semifabricated product
structures.
(~cured preform) of previously impregnated
The most important groups of wound arti-
filaments, strands, tapes and fabrics is wound
cles are: thin-walled shells (their thickness is
layer by layer with controlled tension onto the
negligible compared to their radius); com-
mandrel or previous layers. By varying the
pound structures, including three-layered and
angle of filament or tape placement, it is possi-
multi-layered shells with a light foam or hon-
ble to control the reinforcement fiber angles
eycomb plastic filler; and thick-walled
within the same layer and through the thick-
structural elements. For thin-walled shells, it is
ness of the composite wall. During winding,
most important to optimize the reinforcement
fiber tension generates pressure between lay-
configuration. For compound surfaces, contact
~rs of uncured composite; this pressure
pressure on the interface is a design parameter.
influences the compaction and void content of
Also, there is the problem of monolithicity for
th~ .ar~cle and contributes to more complete
thick-walled structural elements, which is
utilIzation of the strength and stiffness of the
closely related to the problem of control of
reinforcing fibers. If the contact pressure is
residual stresses. Most composites cannot sus-
insufficient for compaction of the material,
tain a significant internal pressure over time
additional layer-by-Iayer compaction of a
without leaking (weeping) through the other-
semifabricated product must be employed.
wise sound composite wall. The use of internal
The wound article must be converted by
liners made from rubber, plastic or metal can
chemical and/or thermal means to the fin-
provide a leak tight structure. The semifabri-
ished article. With heat treatment, the usual
cated (uncured) article has extremely low
method, the temperature can be constant or
strength in the radial direction. Thus, the ratio
can vary with time. The mandrel defines the
of elastic modulii along and across fibers can
internal shape of the article. It is removed after
reach 103• As additional circuits are wound
curing if the mandrel is not an element of the
with tension onto the earlier 'soft' laminate the
structure. The winding process is illustrated in
Fig. 21.P2.
previous circuits are compressed with a redis-
tribution of preset tension.
The process is typical for all types of filament
wound articles with specific processing
Handbook of Composites. Edited by S.T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 approaches for each3• Controlling the winding
 Introduction 457

FORCE COMPACTION (ROLLING.

PRESSING). PARTIAL CURING

CURING (HEAT BUILDUP.

CURING. COOLING).
q=f(t) EXTRA COMPACTION
~ •••••• t •• ~~

..........
==(((-~~
T"=f(t)

~MOVAL FROM MANDREL

[8((((,
ARTICLE

SEMIFABRICATED
MATERIAL

Fig. 21.1 Stages of winding process.

process may include specifying fiber tensioning to the geodesic line, against the static friction or
and extra compaction pressure (internal or viscous resistance of the binder, or mechanical
external), or geometric parameters (configura- obstacles, such as pins, to keep the filaments on
tion of reinforcement lay-up along the length non-geodesic curve.
and through the thickness of the article). 'The Larger components will result in changes to
reinforcement configuration is determined by the system of residual stresses. To calculate,
the design requirements of the article and can analyze and control the values of residual
be achieved with high accuracy. Winding, as a stresses, it is necessary to find the interrelation
way of obtaining preset reinforcement configu- between the winding parameters (mainly ten-
rations, is restricted in terms of available sioning) and the properties of the finished
winding angles. Filament lay-up on geodesic product. By controlling filament tensioning
lines on an external surface is easy. (A geodesic and the reinforcement configuration, it is pos-
line is a line linking two preset points of the sur- sible to control strength and stiffness of the
face along a shortest path.) In non-geodesic material, residual stresses and pressures on
winding, filament tension can shift the filament the mandrel or previous layers.
458 Filament winding

21.2 HISTORY OF WINDING following stages: winding of a preform; heat-

Historically, the developments in winding ing of a preform with the mandrel up to
temperature of binder polymerization; curing
technology can be traced back to 1370 BC
at constant or variable temperature; cooling to
There is an ancient Egyptian flask in the
below the glass transition temperature and
British Museum made by winding glass fibers
then to room temperature; and removal of the
from the melt onto a dissolvable mandrel.
article from the mandrel.
Embalmed mummies were wrapped with
As a rule, thermosetting binders are com-
tapes soaked in resin that were the natural pre-
monly used in winding, although recently
cursors of polyester resins. Rope wrappings
progress has been made with advanced ther-
were used for reinforcing the bamboo powder
~oplastic binders. Among the reinforcing
rockets in ancient China. Guns with wire
wound barrels for reinforcing and proper fIbers, the most widely used are fiberglass, car-
bon, organic (aramid and polyethylenef,9.
stress distribution were fabricated in Russia as
early as 1869 and in Germany from 1900. In the They can be used in the form of single fila-
ments, rovings, woven and unwoven tapes,
development of fiber winding technology
there have been three important stages4-6: the strands and woven fabrics.
adoption of the first patent in 19637; the cre-
~tion of NOL-ring sample4 for strength testing 21.3.2 MATERIALS FOR WINDING
m 1955; and the publishing of the first mono-
There is a distinction between winding with
graph wholly devoted to composite winding
prepreg, when the reinforcing material is
technology by Rosato and Grove in 19648 . All
impregnated with a binder by an outside ven-
these works addressed the development of
thin-walled articles rather than the task of con- dor and subsequently partially polymerized to
verting a semifabricated or uncured material a 'B' stage and 'wet' winding, and when the
processes of reinforcement impregnation with
int.o a .monolithic article, which is the primary
the binder and the winding operation are per-
obJective for modem composites technology.
The history of analytical development formed on the same equipment simultaneously.
includes the paper by L. Euler on thread equi- The rarely used process of 'dry' winding
librium on a rough surface, papers by P. involves winding of the reinforcing material
Laplace and G. Lame on equilibrium of shells without any binder on the mandrel, followed
loaded by pressure, a paper by A.W. Gadolin by vacuum or any other impregnation of the
article with the binder and curing. This latter
~n compound cylinders with preliminary
tightness and his formulation of the problem process will see increasing attention since it is a
o~ tension wrapping of pressure vessels by
c?n.venient and inexpensive method for pro-
vIdmg a stable preform without edges for resin
wue and papers by A.P. Minakov on thread
transfer molding.
equilibrium on an arbitrary rough surface. The
winding mechanics of an extendible wire have The winding operation of thermoplastics
been described by R.Y. Southwell. The state of can involve, in the case of in situ consolidation,
dilution of the high viscosity of the matrix
~e art of the winding mechanics of compos-
~aterial by heat or organic solvents. It is phys-
Ites has been reviewed in detaiF.
Ically analogous to the effect of impregnation
in 'wet' winding or preliminary heating of a
21.3 TECHNOLOGY OF WINDING prepreg. ~l~ough a 'curing' stage is missing
21.3.1 STAGES OF PROCESS
m the wmdmg operation of thermoplastics,
heat treatment and compaction, as a rule, will
The production cycle for most filament lead to significant changes in macromolecular
wound composites can be subdivided into the structure. These changes can result in physical
 Technology of winding 459

shrinkage analogous to chemical and thermal Winding of more complex shapes aggra-
shrinkage in the curing and cooling of ther- vates the problem of pattern closure. To obtain
mosetting plastics. The analysis of winding a continuous layer, it may be necessary to vary
techniques and general engineering winding the bandwidth during winding. Also, it may
theory for thermosetting plastics are surveyed be advantageous to vary the fiber volume and
in detail6 • thickness by alternate compaction techniques.

21.3.3 GEOMETRY OF WINDING 21.3.4 EQUILIBRIUM CONDmONS OF

FILAMENT ON THE SURFACE
The shape of the article to be wound is deter-
mined by the mandrel. The thickness of the Winding along geodesic lines is easier than
article is controlled by the number of circuits along non-geodesic lines. However, to obtain a
wound (taking into account their compress- continuous covering on the surface it may be
ibility during winding and shrinkage during necessary to use non-geodesic winding. This
cooling). The reinforcement configuration is presents the problem of permissible deviations
determined by controlling the wind angle, from the geodesic lines. The angle of geodesic
which is the angle formed between the tangent deviation can be designated by v, i.e. the angle
to the filament and tangent to the intersection between the main normal v to the curve, along
line of the surface to be wound with the plane which the filament has been laid, and the nor-
parallel to the running axis of winding. For a mal to the surface n at the same point. The
body of revolution it will be the angle between curve on the surface is defined by the radius of
the filament and the wind axis. curvature Rc and the radius of twist Rt and the
The formation of a continuous layer with radius of normal curvature Rn (projection of
finite thickness and bandwidths, and with curvature radius of the curve on normal n to
proper fiber geometry, requires a coverage the surface):
solution. If the bandwidth, the number of cir-
cuits to close and the winding angle are 1/Rn = cos vlRc
carefully chosen, the adjacent circuits along and radius of geodesic curvature Rg:
the axis will be linked in butt joints, forming a
smooth continuous covering. If the bandwidth 1/Rg = sin viRc
is greater, the adjacent circuits will form a lap The equilibrium equation for a filament on
joint resulting in a rough surface. If the band the surface has the form (Fig. 21.2):
width is smaller, there will be a gap between
the adjacent circuits along the axis. This gap dN + F +--JQ
ds '
=0 (21.1)
can be partly or completely filled up by either t

fiber or matrix of subsequent layers. Some

structures may be amenable to an incomplete RN + Fn + #Qn - P = 0 (21.2)
winding pattern. Wmding of subsequent cir- n

cuits is feasible either directly on the

preceding ones (winding of ribbed and honey- : +Fg + #Qg =0 (21.3)
combed structures, Fig. 21.1), or on the rough g
outer surface formed from several preceding where N is tension of filament, s is the arc coor-
layers. In the second version, the reinforcing dinate on the curve, F represents surface
fibers undergo additional bending and inter- distributed forces, related to a unit length, # is
weaving and the structure obtained may have the linear mass density of filament, Q repre-
more voids and lower fiber volume than in the sents mass distributed forces, referred to a mass
case of winding with butt joints. unit, including dead weight and inertial forces,
460 Filament winding

-
ft,

Fig. 21.2 Condition of filament equilibrium on the surface.

l"is the basis vector directed along the tangent of pressure. The pressure p on the convex sur-
to the curve and P is the nonnal surface reac- face comes from tension of the wound
tion related to a length unit of filament. material and is unique to filament winding; p
Equation (21.1) describes the redistribution = P fb, where b is the width of the material to
of tension along the filament length due to fric- be wound. For the simple case of circumferen-
tion, viscous resistance from lower-lying tial winding of cylinders with a radius R, this
circuits, viscous resistance of above-lying cir- pressure is calculated according to the for-
cuits and inertial effects. Equation (21.3) mula:
represents the force, which tends to shift the fil-
N
ament onto a geodesic line and viscous p=- (21.5)
bR
resistance of both the lower-lying and above-
lying circuits. In the simplest case of the H this pressure is insufficient (if R is high), it is
friction law IFg I ~ kP where k is the friction necessary either to increase the fiber tension to
coefficient, from equation (21.2) (at Fn = Qn = 0) a value not exceeding the allowable strength
and equation (21.3) (at Q = 0) the following of the fiber band or tows or to employ addi-
relationship can be deriveJ.: tional methods. For helical angle winding (at
Rn an angle l" to the axis) of a cylinder, it follows
R = tan v < k (21.4) from equation (21.2) and the relationship
g

The ultimate permissible angle of geodesic Rn = Rc = Rf sin2",

..,..
deviation is equal to v ~ arctan k.
To obtain a monolithic (fully compacted) that
material with successful chemical processes
(matrix crosslinking) in any type of composite N . 2
p=-Sffiq> (21.6)
processing, some pressure must be maintained bR
through the completion of the curing stage of The greater the angle deviation from the cir-
the matrix. There are several potential sources cumferential the less the compaction pressure
 Technology of winding 461

generated by the fiber band or tows being Helical winding

wound.
The wind eye reciprocates parallel to the axis
There are a variety of methods for creating
of the rotating mandrel and the article (Fig.
additional pressure (internal or external, local
21.3) and is the most common technique used
or general) by means of rollers, power trowels,
for tubular structures. By controlling the ratio
autoclaves, wrapping of a fiberglass or metal-
of rotational and translational speeds, it is pos-
lic tape under tension, wrapping of thermally
sible to control the wind angles of the
shrinking materials, extensible or squeezing
reinforcement. Three other techniques can be
mandrels, rubber bags, etc. Internal pressure
used: one or more wind eyes are stationary,
can be applied by an expanding elastomer, by
while the mandrel rotates and translates, the
pneumatic pressure to an expanding mandrel
wind eye rotates around the translating man-
or by the selection of a mandrel material with
drel, i.e. winding with a stationary whirling
high circumferential expansion properties.
arm type winder which is one of the widely
used methods of continuous fabrication of
21.3.5 REINFORCEMENT CONFIGURATIONS pipe; a wind eye rotates round a stationary
mandrel and executes a translational motion
There are three basic types of reinforcement
along its axis. Winding with a whirling arm
configurations: circumferential, helical and
type winder is a popular method of insulating
polar. There are up to six degrees of freedom
metallic pipelines. Circumferential winding is
between the mandrel and the wind eye in typ-
a particular case of helical winding with a
ical advanced winding machines today and
wind angle of close to 90° (related to the band-
more motions are possible. Wind eyes can
width and the mandrel diameter).
have three additional degrees of freedom
(three rotational motions). The combination of
several motions may allow more effective or
rapid reinforcement placement on a complex
surface.

Circumferential or hoop winding

The wind eye is stationary, except for travel
along the length of the mandrel at a band-
width per revolution and the mandrel together
with the article rotates about one axis, i.e. for
winding of rings, discs, profiled rings and
Fig. 21.3 Lathe type winder.
discs (the ultimate case of profiling is the
winding through an opening), short pipes,
cones and other bodies of revolution. During
Polar winding
circumferential winding, the shape of a cross
section is decided either by the shape of the This type of winding combines several differ-
mandrel or by localized inserts, or by varying ent winding processes. It has the same
the bandwidth additional hoop layers. Some combination of motions as helical winding,
large and small-scale articles have been fabri- but the shortest axis is the axis of rotation. This
cated by rotating the wind eye about the technique combines two rotational motions.
stationary mandrel. Circular rings, ellipses, During simultaneous rotation of the mandrel
ovals and other shapes can be fabricated by a around its axis and the wind eye around an
circumferential winding technique. axis nearly perpendicular to it, a race-track
462 Filament winding

pattern is generated (Fig. 21.4). This process instead of translation motion of a wind eye,
enables the fabrication of vessels with differ- there is a reciprocating motion along a curvi-
ent sized polar openings or closed ends. There linear trajectory. One version of winding with
are a number of alternative techniques. First, a whirling arm-type winder is when a wind
the wind eye executes a rotational motion in eye rotates and moves along some curvilinear
two planes. It is a very complex technique and trajectory.
is used only for unique articles. Second, the Fourthly, there is a group involving two
rotational motion of the mandrel is in two rotational motions executed alternately (a pla-
planes (Fig. 21.5), sometimes supplemented by nar-polar winding). Applications of this
a reciprocating motion of a wind eye or turn- method include chord winding (Fig. 21.6) of
ing of its head. composite flywheels.

Fig. 21.4 Racetrack winder.

Fig. 21.6 Scheme of chord winding. Numbers des-

ignate the order of chord lay-up: 1-1',2-2',3-3' for
chords on the right side of mandrel; 1'-2,2'-3 for
chords on the reverse side of mandrel; D is a disc
mandrel; H is a pendulum mechanism.

Fig. 21.5 Polar winding by the mandrel rotating in

two planes. The fifth group consists of three rotational
motions. Kinematics of this process may
involve a combination of two rotations around
Thirdly, there is a group involving three nonperpendicular axes and a combination of
motions, two of which are rotational in two two rotations around two perpendicular, but
planes and one is translational. To this group noncrossing axes (winding of a torus).
belong, for example, a race-track type of wind- Even more complicated types of winding
ing, where a wind eye executes a motion along including combinations of circumferential, heli-
a closed but not circular trajectory. The trajec- cal, polar winding with axial lay-up and/or
tory shape affects the variation of the winding braiding have been described2,lO,ll. One of the
angle over the surface of the article. Pendulum examples of fabricating large-size articles by a
winding is analogous to lathe winding, but combined method is shown in Fig. 21.7.
 Technology of winding 463

Fig. 21.7 Combined winding of large-scale articles.

Winding pattern control explosion of the sporting goods business all

over the world. Many golf shaft manufactur-
In the simplest winding machines, control of
ers ha:e studied or are studying the possibility
the ratio of speeds of different motions is done
of us~g filament winding rather than flag
by ~gid kinematic coupling (usually gears or
w~appmg for production of carbon/graphite
chams). With modern winding machines dif-
reinforced golf shafts. Filament winding offers
ferent motions can be regulated individually
lo~e~ labor costs, because of the possibility of
with greater accuracy by computer control. In
wmdmg many shafts at once (Fig. 21.9), the
the gear and chain machines, looseness and
possibility of lower material costs with unlim-
wear of the mechanical elements have been
ited possibilities for varying the axial and the
occasionally responsible for poor winding pat-
torsional stiffness along with the advantage of
terns. Wmding inaccuracies can result from
no seams, such as would be generated by flag
combined fiber delivery and winding machine
wrapping with unidirectional tape, along the
posi~on~l err~rs. These errors can be analyzed
shaft's length. Tennis racquet manufacturers
by wmdmg Wlth a controlled material such as
are also considering a filament wound rac-
fine copper wire, over the mandrel' surface.
quet, again because of the expected lower
Typical feedback resolutions on computer con-
labor costs and ease of control of stiffness.
trolled winding machines range from
Along with the new machines much better
1Q-4m/bit to 1(T5 m/bit for linear axes and
fiber delivery systems have been offered, with
from 0.02 to 0.001 degrees/bit for the rota-
electronic fiber tensioners with roll diameter
tional axes. The speed sensitivity has been
sensing, provisions for setting tension from
greatly impr~ved with the newer control sys-
the main control panel and rewind capability.
tems by tracking both velocity and position for
The machine manufacturers have devel-
ea.ch .axis. This makes it possible to change
oped menu-driven, user friendly computer
wmdmg speeds during the winding.
programs for which the user inputs need only
The standard for the industry now is the 6-
be part geometry, wind angle, band width and
axis filament winding machine (Fig. 21.8).
number of circuits desired.
These have been built in the USA, UK and
One of the most helpful innovations intro-
Germany and are marketed worldwide. These
. duced by the filament winding machine
machines, in general, can be programmed by
manufacturers has been the ability to work
the manufacturer, by third party software sup-
wi~ the machine control computer programs
pliers or by the customer. The newest venue
off-line on a common PC within the standard
for the machine manufacturers has been the
464 Filament winding

Fig. 21.8 6-Axis computer controlled filament winding machine. Machine motion indicated by arrows.
(Courtesy of McClean Anderson, Inc.)

operating systems such as DOS or Windows. 21.3.6 MANDRELS

This allows the design engineer to participate The mandrel, which determines accurate inter-
in the selection of machine parameters such nal geometry for the component, is generally
as number of circuits to close and the number the only major tool. Low-cost mandrel materi-
of crossovers, which influence structural als such as cardboard or wood can often be
properties. . . used when winding low-cost routine parts.
For much of the history of composIte fabn- For critical parts requiring close tolerances,
cation by filament winding, only the large expensive mandrels designed for long-term
aerospace manufacturers had computer pro- use may be required. For high-temperature
grams to predict slipping or bridging of the cure 315°C (600°F), graphite mandrels with
wet fiber band on a non-geodesic path. Now low thermal expansion may be advantageous,
these programs are available from third party however some attention should be paid to the
suppliers and they include 3-D display of the potential difficulties of mandrel removal. Gas
mandrel and the fiber path, the friction at containment pressure vessels often require
each section and the deviations from geodesic metal liners because composites are porous;
and desired winding angles. They also dis- these metal liners can also serve as mandrels12•
play the more mundane information, such as Mandrels are either removable or nonre-
laminate thickness, winding time, length of movable (remaining as a part of the wound
fiber used and laminate weight for any figure structure). Removable mandrels are classified
of revolution and general nonsymmetric according to the removal technique as:
parts (Fig. 21.10). The programming extends
to nonaxisymmetric parts also, such as • entirely removed (for example, tubular
elbows. mandrels with or without taper and with
release agent);
 Technology of winding 465

Fig. 21.9 20-Spindle golf shaft filament winding machine. (Courtesy of ENTEC, Salt Lake City, UT, USA.)

• collapsible (the mandrel is disassembled low-melting temperature alloys. Of these

and removed piece by piece); materials, the water-soluble sand mandrels
• breakable or soluble (plaster, sand or salts). and the breakout/washout plasters are most
commonly used.
The selection of a mandrel involves several
trade-offs. These include part size and com-
plexity, size of openings, resin system and its
curing cycle and the number of components to
be fabricated. The basic requirements for a
mandrel, whether it will be removed from the
part after winding or remain as a part of the
structure, are:
• It must be stiff and strong enough to sup-
fD d3t-$$ 1111

port its own weight and the weight of the

applied composite while resisting the fiber
tension pressure from winding and curing.
• It must be dimensionally stable and should
have a thermal coefficient of expansion
greater than the transverse coefficient of the
composite structure.
Components such as rocket motor cases or
pressure vessels have closed or reduced-area
end openings that require the use of dissolv-
able or fusible mandrel materials. Some of
these materials are: water-soluble sand man- Fig. 21.10 Cadwind display of complex mandrel
drels, breakout plasters, soluble plaster, and fiber path. (Courtesy of Material SA, Brussels,
soluble or meltable salts, eutectic salts and Belgium.)
466 Filament winding

Water-soluble mandrels are primarily used account these fields, their analysis, develop-
in rocket motor cases and pressure vessels ment and application of the methods of
where mandrel removal through small open- controlling the fields are important aspects in
ings is desired. The sand/polyvinyl alcohol the manufacturing of defect-free structures,
(PVA) mixture is cast into female molds that especially thick-walled components.
have preassembled components, such as insu-
lation, wind axis, lightening tubes and polar
21.4 THIN-WALLED STRUCTURAL
bosses. The sand mixture is cured and the two
ELEMENTS
mandrel ends are assembled and bonded.
Spider/plaster mandrels are often used to pro-
21.4.1 TYPICAL ARTICLES
vide a high-tolerance mandrel surface. The
plaster is cured, then overwrapped with tetra- The basic problems in the design and manufac-
fluoroethylene tape or other separator film. ture of thin-walled composite shells are related
Following cure, tooling is removed, the plaster to the optimization of shape and reinforce-
is chipped or washed out and the release tape ment configurations13 • Fibrous composites
is removed, leaving the desired inside dimen- show their best performance under tensile
sions. Metal-supported plaster is generally load acting strictly along the fibers, while the
used for relatively large parts of 3-6 m. matrix serves only to distribute the load uni-
Segmented collapsible mandrels, used for long formly among fibers. A rational reinforcement
tubes, are specialized and expensive, but the configuration for a thin-walled composite
advantages of their reusability and the continu- structure is similar to that of a guyed structure
ous winding process renders expensive tooling working under the same loads. The winding
worthwhile for high-production applications. technique is a natural process and is the most
The main considerations for mandrels are widespread production process for thin-
stiffness, hardness, strength, coefficient of walled structural elements.
thermal expansion and heat resistance. The most widely used elements are cylin-
Mandrel removal from the part is equivalent drical shells, such as pipelines which operate
to application of the pressure of opposite sign under a combination of external or internal
and new shearing stresses at the interface pressure, bending and axial tension. . .mmpres-
between the part and the mandrel and must be sion. Cylindrical sections of missiles,
done with some care. cylindrical shells with semi-spherical closed
ends used as tanks or pressure vessels,
toroidal shells for electrical applications; and
21.3.7 RESIDUAL STRESSES
conical shells, which include nose fairings of
The residual stresses caused by the tension of aircraft and rocket motors, are also common.
the reinforcement and compaction undergo
essential changes during elevated temperature
21.4.2 SHAPE CONFIGURATIONS OF WOUND
exposure. The thermoelastic stresses as a result
PARTS
of the difference in thermal expansion of the
mandrel and composite article and the Filament-wound shapes generally include
anisotropy of thermal expansion of composite cylindrical, spherical, conical, or dome-end
in a statically indeterminable type of structure configurations. These bodies of revolution best
are added to the system of residual stresses. exploit the advantages of high-speed winding.
The majority of wound articles end up with Spherical shapes contain the maximum
these stresses. The resulting fields of residual possible volume with minimum surface area.
stress in wound articles are comparable in Hence this shape is commonly used for pres-
scale to fields of allowable stresses. Taking into sure vessels. Because the radius of curvature is
 Thin-walled structural elements 467

equal in all directions, the best reinforcement 21.4.3 SPECIFIC FABRICATION FEATURES
pattern is an isotropic combination of wind
angles. This pattern is readily obtained by One of the main problems in obtaining the
winding a series of great circles, each stepped designed shape for thin-walled composite
out from the preceding pattern by the width of structures is the problem of warpage. The
one winding band. bending stiffness of structure is proportional
Cylinders and shafts can be wound on a to the cube of the thickness, while residual
cylindrical mandrel by providing an adequate stresses, as a function of thickness, increase
pattern reversal length at each end. The cured more slowly. Therefore, warpage can be a
part can be removed from the mandrel by problem with thin-walled articles. The reasons
using a stripping die. Wmding patterns consist for warpage are:
of hoops or helical patterns as desired to react • Inhomogeneity of material properties in
to design loads. For torsion-type loading, ±450 structures can be ascribed to local imperfec-
helicals provide the most efficient load path. tions, inaccurate wind angles, inhomogeneity
Thick-wall effects are a major consideration in of temperature field during cure and
the design of composite shafting. changes in binder composition or volume
With flat cylinders, the width is just a frac- within the wall.
tion of the thickness. These parts can be • Mandrel removal; pressure and shearing
fabricated using hoop windings alone. A typical stresses on the interface between the man-
application is in flywheels. Thick-wall effects drel and article can be distributed
and delaminations are potential problems that nonuniformly as a result of different ther-
must be carefully considered. Fabrication tech- moelastic behavior of the mandrel and
niques using programmed winding tension, composite article, especially for complex
high strain resin systems and low temperature shapes.
cures help to overcome these problems. • Anisotropy of composite shrinkage and dif-
The combination of cylinders with domes ferent relaxation rate of components of
is typical of rocket motor cases and pressure residual stresses can cause inhomogeneity
vessels. The winding patterns and dome and anisotropy of relaxation characteristics.
shapes must be carefully chosen to prevent
fiber slippage and to react adequately against In the production of thin-walled articles it is
the pressurizing forces on the domes. Ideal, most important to have a high degree of wind
or isotensoid, domes or end closures for fila- angle fidelity.
ment-wound vessels have a general surface
of revolution that requires a numerical solu- 21.4.4 SANDWICH AND COMPOUND lHIN-
tion. The two types of dome contours WALLED STRUCTURES
generally considered are geodesic (constant
stress) and polar (planar) ones, each of which Sandwich structures incorporate a low density
requires its own derivation. The only geomet- filler to increase bending stiffness to increase
ric variables affecting the dome shape are the bending moment of inertia. Foam and honey-
radius of the cylindrical portion of the pres- comb plastics are most often used as low
sure vessel and the radius to the center of the density fillers. These materials serve not only
filament band at the polar opening. There are to increase the separation between the load-
several winding considerations. Winding pat- bearing layers, but also to act as sound and
terns must be stable, the dome must be heat insulation. Sandwich structures are
shaped with properly sized openings and widely used in aviation and construction.
there must be balanced stress fields in a cylin- Typical design problems are associated with
drical section. optimization of the thicknesses of load-bearing
468 Filament winding

layers and the filler and selection of the rein- material properties. To define these changes, a
forcement angles. When winding the outer simplified hypothesis about the progression of
load-bearing layer the winding tension has to the stress state has been developed.
be low enough to avoid damage to the low Experiments with tensometric (instru-
density filler. Stress distribution during wind- mented) mandrels have served as an impetus
ing depends essentially on the compliance of for the development of an engineering theory
the mandrel6• In sandwich structures the outer (Fig. 21.11). The distinction between the wind-
load-bearing layer of the filler serves as the ing process of a semifabricated composite
mandrel for the outer skin lay-up. article and that of an isotropic metallic tape was
made by examining an 'integral force parame-
ter' - the dependence of the pressure - on the
21.5 THICK-WALLED STRUCTURES
mandrel related to the number of circuits n
being wound. A significant part of the applied
21.5.1 ENGINEERING THEORY OF WINDING
pressure is consumed in the deformation
The development of engineering winding process of the lower-lying circuits. By using
mechanics must involve the study, description instrumented mandrels it has been possible to
and sequential consideration of the peculiari- evaluate the variation in winding pressure for
ties of composite behavior under loading every stage of the process. In the curing
perpendicular to the plane of reinforcement. process, constant pressure on the mandrel is of
During the winding process the physical- special interest and makes it possible to
mechanical properties of the composite develop several variations of the theory of
undergo great changes. The composite proper- residual stresses, omitting the polymerization
ties remain linear and practically do not change stage. It establishes the dependence of the resid-
in the reinforcement direction. The compliance ual stresses fiJr and O"ao on the geometry of the
in the transverse direction is essentially nonlin- article hlR, material anisotropy 132 = Ea IE"
ear during the stages of winding, heat build-up winding angle qy(x, r), the number of circuits n,
and curing of the article and can vary by three winding tension N(x, r) and parameters of the
orders. The possibility of applying a universal curing process (q is contact pressure, T is tem-
rheological model to such a material is beyond perature and 't is curing time).
the present state of the art. The engineering
approach assumes that the material behavior in
each of the stages of the process follows a spe-
cific rheological law. Passing from one stage of
cure to another results in an abrupt change in

9res
l

Fig.21.11(b) Pressure variation on the mandrel. 0-1

is a winding stage; 1-2 is heating after winding; 2-3
n
is heating prior to curing; 3-4 is polymerization at
constant temperature; 4-5 is cooling on the man-
drel; Non/R is the ultimate pressure (- - -) on the
Fig. 21.11(a) Relative pressure (6) on the mandrel mandrel in elastic tape winding, R is a radius of the
in winding.·1 is a metallic tape winding; 2 is a com- wound article, qres is a residual pressure on the man-
posite winding. drel after cooling.
 Thick-walled structures 469

Residual stresses in thick-walled articles

differ essentially from those in thin-walled
structures, both in nature and magnitude. In
thin-walled articles these stresses can cause
the loss of monolithicity and reduction in
load-carrying capacity. a.)
To analytically describe the force variation
in the winding process, the circular models
were based on the simplest linear approxima-
tion of the nonlinear strain diagram or "" cr (Fig.
21.12). The basis of the solution is that the
process of continuous tension winding can be
modeled by a system of concentric rings,
which are first mounted on the mandrel and
Fig.21.12 Models of winding stages. (a) a linear cir-
then on each other under contact pressure
cular model (1965); (b) a spirally circular model
equal to the tensioning of the ith layer. The (1971); (c) approximation of the deformation char-
range of thicknesses and operating pressures, acteristics of winding materials used for a nonlinear
were broadened when it became impossible to circular model (1973); PI' Pw PIlI are coefficients of
consider only one linear section of the stiffness anisotropy of the composite in the process
stress-strain curve of the material. This of winding; (d) a model of a growing body, R =
required model refinement. The modification R(w), w is a parameter of growth.
of this model has played an essential role in
the development of technology of winding
thick-walled articles. It was established that
21.5.2 METHODS FOR CONTROLLING
the total winding pressure on the mandrel had
RESIDUAL STRESSES
increased nonlinearly, approaching asymptoti-
cally an ultimate value
Programmed winding
No
P "" -R arctan sinh me Programmed winding involves variation of
IC i
winding tension in accordance with a speci-
where IC = P(c/Ri ), Ri is the inner radius of the fied program with the aim of eliminating fiber
ring to be wound and c is the band thickness. buckling and compensating for thermoelastic
The critical number of circuits ncr was estab- tensile stresses which arise during cooling of
lished beyond which the winding pressure on the article with the mandrel. Crack formation
the mandrel did not increase. and loss of monolithicity can occur at this
stage. By properly selecting the correct tension
program it is possible to transfer the radial
stresses or into compression and eliminate the
Where excessive force conditions of winding danger of fiber buckling.
appear, a circular area through the thickness of The range of variation of tensioning Nmax is
the article appears, where the circuits pass into limited by the strength of the semifabricated
compression leading to instability in the form article to be wound, the contact pressure q =
of irregular curvatures and buckling of the N rna/ R and by the overall dimension of the
reinforcement. These defects are common to part to be wound R. New methods have been
thick-walled wound articles. employed for developing thick-walled wound
structures: winding with layer-by-Iayer cur-
ing, introduction of compensating interlayers
470 Filament winding

and winding with extra pressure. These meth-

ods, along with programmed winding, have
facilitated the design of thick-walled articles
with wall thickness comparable and some-
times equal to the inner radius. Another
means is reinforcing thick-walled articles in
the radial direction with a radial reinforce-
ment. This method makes it possible to
significantly increase shear stiffness and trans-
verse tensile strength.

Winding with extra external or internal

pressure
TIlis method can be accomplished by layer-by-
layer compaction in the winding process, by
spinning rollers, by wrapping fiberglass or
metallic tape and by extensible (Fig. 21.13)
mandrels with programmed compaction pres-
sures to force extra compaction of the article.
Compaction of the wound article can also be (a) (b)
performed with these same processes during
heating and curing. Control of residual Fig. 21.13 The extensible mandrel (a) and the pre-
stresses can be achieved by varying the mag- form stretched along the axis (b).
nitude of extra pressure and the time of
removal. For a step-by-step compaction tech-
nique the following operations (winding of a cations of programmed winding. The advan-
part of the article, compacting, pressure tages of layer-by-Iayer winding however
release) are repeated. In this case, the residual result in an significant increase in the dura-
stresses can be altered by varying the interme- tion of the process of manufacturing for
dia te thicknesses of the article and the thermoset-polymer composites. Therefore, a
magnitudes of the applied pressure in each consideration of the number of individual
step. layers based on a theoretical analysis is very
important for process economics. For interme-
diate curing, both the usual curing methods
Winding with layer-by-Iayer curing
and accelerated ones (Le. radiative or catalytic)
The basic idea of this method is similar to step- are appropriate.
by-step compaction in that it decreases the
compliance of the already wound circuits. TIlis
A method of temperature gradients
is a result of partial or complete curing of a
group of wound layers, followed by winding There are methods for controlling the residual
of the next group of layers. As the number of stresses by varying the process parameters in
curing steps, into which the process is divided, the heat build-up and curing stages.
increases, layer-by-Iayer winding approaches Thermoelastic behavior of anisotropic bodies
some idealized ultimate process of winding a differs greatly from the thermoelastic behavior
'cured' material (similar to thermoplastics). of isotropic bodies. The methods of tempera-
The layer-by-Iayer winding extends the appli- ture gradients must compensate, at least
 Applications 471

partly, for the thermoelastic stresses which 21.6 APPLICATIONS

arise from anisotropy of thermoelastic proper-
ties, by artificially developed and controlled 21.6.1 PRESSURE VESSELS
temperature gradients. Different modifica- Most advances in filament winding have been
tions and objectives for using the method of made in the course of development of pressure
temperature gradients are possible. Among vessels, particularly rocket motor cases. In
them are the curing in an inhomogeneous tem- these structures the reinforcement is loaded
perature field, the creation of temperature
only in tension and winding ensures the
gradients during the cooling process, or mod- arrangement of fibers along the directions of
ification of the polymer matrix through the
principal stresses. It uses the main advantages
thickness of the article for the purpose of con- of composites, such as high strength and stiff-
trolling the gradient of heat release during a ness in the reinforcement direction, and avoids
curing chemical reaction.
shear stresses which are dangerous for com-
posites. Rocket motors come in many sizes,
A spatial reinforcement from the graphite/epoxy filament wound case
for the Space Shuttle to pressure vessels as
A number of methods for controlling residual small as 0.152 m in diameter for orbital adjust-
stresses at the expense of variation of the rein- ments to spacecraft. Metal-lined pressure
forcement configuration have been worked
vessels are widely used in aerospace, aircraft,
out. By controlling portions of the elastic, marine and commercial applications. The
strength and thermo-physical properties Skylab Oxygen Tank shown in Fig. 21.14 is an
either by changing their level or generating example of such a structure. The tank, 1.143 m
artificial inhomogeneity, it is possible to affect in diameter and 2.286 m long, operated at
the level and distribution of residual stresses. 31 MPa with a demonstrated life up to 1000
There are methods for the reduction of cycles in temperature service range of
anisotropy of composites, such as an extra -53-71°C.
radial reinforcement by short needles, wind-
ing while braiding a three-dimensional
reinforcing framework, hybrid reinforcement
and decreasing of fiber volume content,
Another method of controlling the residual
stresses is to introduce compliant interlayers
of rigid foam or other material. The third
group of methods creates an artificial inhomo-
geneity by substituting a homogeneous
structure for a compound one which has dif-
ferent elastic and thermophysical properties.
This is done by creating a smooth macro-inho-
mogeneity of properties by varying the bulk
reinforcement coefficient, or the ratio of two
types of reinforcement in hybrid composites,
or by changing the reinforcement angle or
chemical composition of matrix through the
article. The basic methods for controlling the
residual stresses are summarized in Table 21.1.
Fig. 21.14 Skylab Oxygen TankS. (Courtesy of
Brunswick Corporation.)
472 Filament winding

Table 21.1 Methods of controlling residual stresses in wound articles

Method Techniques
Forced A programmed winding
Pressing and rolling during winding process and after it
Chemica-technological Curing in regime of spreading of reaction front
Curing in nonhomogeneous temperature field
Variation of matrix composition through the article
Continuous or layer-by-Iayer curing during a winding
Controlling physical fields Controlling of cooling rate
Method of temperature gradients
Influence of electromagnetic fields at curing
Controlling of reinforcement Additional radial reinforcement
schemes Introduction of thermocompensating interlayers
Variation of reinforcement coefficient
Application of hybrid composites
Creation of artificial nonhomogeneity
Creation of compound structures
Winding with braiding

21.6.2 TIllN-WALLED SUPPORTED amount of attenuation. By increasing the pres-

STRUCTURES sure of a water-filled test vessel, deep water
Thin metal-lined pressure vessels over- depths could be simulated for calibration of
wrapped with Kevla~ reinforcement for the sonic equipment, while the equipment is sus-
Space Shuttle and various medium class satel- pended just a few feet under water. A 2.4 m
lite systems and metal fuel tanks overwrapped diameter acoustic pressure vessel with 0.10 m
by composites for military aircraft are wall thickness rated for multiple cycle life
included here. The applications in aircraft 34.4 MPa working pressure is shown in Fig.
structures include entire composite aircraft 21.15. Special winding programs were
fuselages as well as many secondary struc-
tures. Two prototype fuselage test sections for
Beech Starship with different stiffening con-
cepts have been developed, both filament
winding concepts used high strength graphite
and Shell resin 9400. The first technique used a
filament-wound isogrid network, 1.8 m diam-
eter, 2.4 m long geodesically reinforced test
section to prove the concept. The second
approach used a honeycomb-stiffened struc-
ture with graphite epoxy skins and
lightweight composite tooling.
One of the earliest uses of composites for
marine application was in the use of acoustic
transmission of radar signals. Fiberglass-rein-
forced plastic structures have an ability to Fig. 21.15 Eight foot (2.4 m) diameter acoustic pres-
transmit sound waves with a minimum sure vesselS. (Courtesy of Brunswick Corporation.)
 Applications 473

required in concert with multiple compaction loads. The advantage of this technology has
and cure cycles due to the thickness of the ves- been demonstrated in high resistance to
sel. Composites combining low density, impact loads. The impact failure is localized
stiffness and high strength with superior within several spatial cells of the structure.
sound damping properties are very promising
materials for submarines and deep-submer-
21.6.5 THICK-WALLED STRUCTURES LOADED
gence devices. Filament wound composites
INTORSION
offer cost savings in terms of materials and
process both for primary and secondary struc- Progress has been made in the development
tures on submarines. of thick-walled wound rods loaded in torsion.
A characteristic example is shafting for
ships 15. The advances in the mechanics and
21.6.3 NETTED-RIBBED SHELLS
technology of winding of thick-walled large-
The middle layer of a sandwich composite scale elements have substantially solved this
structures is usually fabricated by an alterna- problem. It is well known that fibrous com-
tive technology. Winding technology can be posites behave poorly when loaded in torsion.
used not only to manufacture facing layers, Ways of increasing load-carrying capacity
but also for middle layers by winding a net- involve increase in shear strength of a ply and
work of helical ribs. Netted-ribbed composite the adherence to the optimal law of shear
shells (Fig. 21.16) are widely used in aerospace stress distribution over the radius of a rod.
because of low dead weight, as antenna Increase in the strength of a ply is accom-
frames, solar batteries, etc. 14 . plished by properly selecting the optimal
angle in the plane of fiber lay-up and the
number of starts and crossovers during wind-
21.6.4 WINDING WITH BRAIDING
ing. Shear strength in torsion T\T = 400 MPa
Winding with braiding allows the manufac- has been attained and a shaft was fabricated
lO

ture of articles of complex shape with with wall thickness equal to the inner radius16 •
sophisticated reinforcement techniques. These Thick-walled wound structures can also be
structures possess high resistance to bending, manufactured by multidirectional braiding of
torsion, tension, compression and to local a system of one tow. The main difficulties of
this process are the complexity of preform
impregnation and necessity of compacting
over the external surface to increase fiber vol-
ume and to ensure monolithicity of the article.
This promising technology can also be used
for braiding over the external surface of unidi-
rectional thick-walled elements, to reduce the
danger of delamination due to low transverse
tensile strength. Another possibility to
increase the torsion strength of thick-walled
products is the spatial placement of straight
fibers along the surface contours of one-sheet
hyperboloids, whose centers are shifted along
the longitudinal shaft axis for a specified step.
The slope of straight contours of one-sheet
Fig. 21.16 Carbon-epoxy filament wound isogrid hyperboloid can be governed by the wind
space structure. angle of the applied fibers, which are originally
474 Filament winding

parallel to surfaces. The spatial arrangement Wu=WujV =kv atu

v 0
of reinforcement results not only in the
increase of transverse tension and interlami- where Vo is the overall flywheel dimension.
nar shear strength and in the localization of Advanced composites are the most promising
fracture within several spatial cells2,17. materials for inertial energy storing systems
because of high specific strength, high tensile
21.6.6 COMPOSITE FLYWHEELS strength along reinforcing fibers and compara-
tively safe modes of failure. The possibility of
The ultimate storage energy WU of a rotating quick release of all stored energy is one of the
body is proportional to tensile strength of the advantages of flywheels. The highest mass
material at" and body volume V , Le. (Wu = energy storage capacity is achieved with rim
kfatuV) and is independent of the material den- flywheels made by a circumferential winding
sity p. The ultimate angular speed of flywheel technique (km ~ ).5). This is true only for thin
depends on the material density and conse- rings. Increasing volume energy storage
quently, composite flywheels are high-speed capacity requires growth of the radial thick-
structures. Mass energy storage capacity Wm U
ness of the rim.
is proportional to specific strength of material With increasing radial thickness the prob-
lem becomes bidirectional, Le. not only
circumferential but also radial stresses must be
considered. The low transverse tension and
where m is flywheel mass and volume energy shear strengths of unidirectional composites
storage capacity Wvu is limit the ultimate rim thickness. This is why

Fig. 21.17 A full-scale specimen of the hull of underwater vessel of reinforced plastic before test in a pres-
sure chamber19.
 References 475

previously developed methods of increasing Fabrication, 2nd printing. Covina: SAMPE.

load-carrying capacity of thick-walled articles 6. Tarnopol'skii, Yu.M. and Beil', A.1. 1983,
appear to be extremely useful in the design Problems of the mechanics of composite wind-
ing. In Handbook of Composites, (eds. A. Kelly and
and fabrication of composite flywheels 18 • The
Yu.N. Rabotnov), Vol. 4, Fabrication of
high effectiveness of methods of increasing Composites, (eds. A. Kelly and S.T. Mileiko) pp.
transverse tensile strength of composites by 45-108. Amsterdam: North-Holland.
improving the matrix-fiber adhesion, i.e. by 7. Concise Encyclopedia of Composite Milterials, 1989,
plasma surface treatment of carbon fibers, (ed. Anthony Kelly) Oxford: Pergamon Press.
may be of greater value for thick wall fabrica- 8. Rosato, D. and Grove, C.. 1964. Filament
tions. Use of chord winding is another way to Winding: Its Development, Manufacture,
Applications and Design. New York: John Wiley.
increase radial tensile strength and provides a
9. Tsai, S.w. 1994, Theory of Composites Design.
solution to solve the following problems: how Dayton, Paris and Tokyo: Thlnk Composites.
to connect the rim, shaft and hub of the fly- 10. Textile Structural Composites. 1989, (eds. T.W.
wheel and how to increase shear strength and Chou and EK. Ko) Composite Materials Series,
stiffness. Vol. 3, (series ed. B. Pipes) Amsterdam: Elsevier.
11. Tsyplakov, O.G. 1974 (Part 1), 1975 (Part 2),
Scientific Fundamentals of Technology of Fibrous
21.6.7 DEEP SUBMERGENCE DEVICES Composite Materials. Perm: Perm Book Publisher.
12. Humphrey, w.n. and Peters, S.T. 1987, Filament
Deep submergence composite devices with the winding. In Engineered Milterials Handbook, Vol.
depth of immersion up to 2000 m (manned) and 2, Engineering Plastics, pp 368-377. Ohio: Metals
up to 6000 m (unmanned)19 have been built. A Park.
hull with 22 m length, 8 m diameter and 250 t 13. Obraztsov, LE, Vasiliev, V.V. and Bunakov, V.A
total mass was entirely manufactured by wind- 1977. Optimal Reinforcement for Composite Shells
of Revolution, Moscow: Mashinostroyenie.
ing on numerically controlled machines. The
14. Bunakov, V.A and Protasov, V.D. 1992, Netted
full-scale hull (2.4 m diameter and 17 t mass) for composite cylindrical shells. In Proc. 1st
an underwater vessel made up entirely of fiber- USSR-USA Symp., Mechanics of Composites,
glass plastic (metal fraction in the structure less Vol.2, Composite Structures, (ed. Yu.M.
than 0.1 'Yo) withstood an external pressure Tarnopol'skii), pp. 89-98. Riga: Zinatne.
350 MPa is shown in Fig. 21.17. Tests methods 15. WIlhelmi, G., Appelman, W. and Loo, E 1986,
for wound articles have been described in detail Composite shafting for naval propulsion sys-
in References 15 and 20. tem. Naval Engng., July, 129-136.
16. Tamopol'skii, Yu. M. and Zakrzhevskii, AM.,
1994, Thick-walled composite rods loaded in
REFERENCES
torsion. Mechanics Compo Mater 30(4) 40-47.
17. Tarnopol'skii, Yu.M., Zhigun, I.G. and
1. Peters, S.T., Foral, R.E and Humphrey, W.o. Polyakov, V.A. 1992, Spatially Reinforced
1987, Filament winding. In International Composites. Lancaster: Technomic.
Encyclopedia of Composites, (ed. S.M. Lee) pp. 18. PortnoY. 1989. Composite flywheels. In
503--518. New York: VCH. Handbook of Composites, (eds. A Kelly and Yu.N.
2. Structural Composites. A Handbook, 1994, (eds. Rabotnov), Vol. 2, Structures and Design, (eds.
Yu.M. Tarnopol'skii and V.v. Vasiliev) C.T. Herakovich and Yu.M. Tarnopol'skii) pp.
Lancaster: Technomic. 531-582. Amsterdam: North-Holland.
3. Peters, S.T. 1987, Filament winding. In 19. Advanced Materials and Processes. 1993.
Engineered Milterials Handbook, Vol.1, Composites, Scientific Technical and Advertising Article
pp. 503--509. Ohio: Metals Park. Collection 1, 70-72. St. Petersburg: CRISM
4. Handbook of Composites, 1982, (ed. G. Lubin) Prometey.
New York: Van Nostrand Reinhold. 20. Tarnopol'skii, Yu.M. and T. Ya., Kincis. 1985.
5. Peters, S.T., Humphrey, W.D. and Foral, R.E Static Test Methods for Composites. New York: Van
1991, Filament Winding Composite Structure Nostrand Reinhold.
FIBER PLACEMENT 22
Don O. Evans

22.1 INTRODUCTION When starting a course, the tows are

restarted and compacted onto a surface. As the
Fiber placem:nt is a unique process combining
course is being laid down, the processing head
the dIfferential tow payout capability of fila-
can cut or restart any of the individual tows.
ment winding and the compaction and
This permits the width of the band to be
~ut-restar~ capabilities of automatic tape lay-
increased or decreased in increments equal to
mg. Dunng the fiber placement process,
one tow width. Adjusting the band width
individual prepreg fibers, called tows, are
avoids excessive gaps or overlaps between
pulled off spools and fed through a fiber deliv-
adjacent courses. At the end of the course, the
ery system into a fiber placement head (Fig.
r~maining tows are cut and the head is posi-
2.2.1). Here they are collimated into a single
tioned to the beginning of the next course.
fIber band and laminated onto a work surface
During the placement of a course each tow
which can be mounted between a headstock
is dispensed at its own speed, allowing each
and tailstock.

Individual Tow Payout

With Controlled Tension

Fiber Placement Head

- - - - Band CoIl mator

Tow Restart Rollers -----lI~-. 1II1II

Collimated Fiber Band

Compaction Roller

Fig. 22.1 Fiber

placement head.

Handbook of Composites. Edited by S.T. Peters. Published

in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7
 Introduction 477

tow to independently conform to the surface normal to the surface while the machine is
of the part. For example, when the head lami- laminating tows. The machine also has 24 pro-
nates a curved path, the outer tows of the fiber grammable electronic bidirectional tensioners,
band pull more length than the inner tows. which are mounted in a creel. These tensioners
A rolling compaction device, combined provide individual tow payout and maintain a
with heat for tack enhancement, laminates the precise tension. The fiber placement head is
tows onto the lay-up surface. This action of mounted on the end of the wrist. The head
pressing the tows onto the work surface (or a precisely dispenses, cuts, clamps and restarts
previously laid ply) adheres the tows to the individual prepreg tows.
lay-up surface and removes trapped air, mini- To increase productivity some machines are
mizing the need for vacuum debulking. equipped with dual mandrel stations (Fig.
Figure 22.2 is a diagram of a Cincinnati 22.3). This setup allows two sets of lay-up
Milacron 'Viper' Fiber Placement System tools to be placed on the machine, ensuring a
(FPS). This system has seven axes of motion constant supply of work for the head. During
and is CNC controlled. The machine consists routine manual operations such as hand lay-
of three position axes (carriage, tilt, crossfeed), ing small plies, inspecting plies, or vacuum
three orientation axes (yaw, pitch, roll) and an debulking on one part set, the head simply
axis to rotate the mandrel. All of these axes are shifts to the other mandrel and picks up the
necessary to make sure the processing head is program where it left off.

X-axis (arm crossfeed)

V-axis (arm tilt)
Z-axis (carriage longitudinal movement)
C-axis (parVmandrel rotation)
I-axi$ (head yaw)
J-nis (head pitch)
K'aris (head roll)
A-axis (redirect roller angular position)
Q-axis (low restart linear position)
Fig. 22.2 Fiber placement
system.
478 Fiber placement

rest of the tape is despooled. This will cause

the tape to eventually break. During part fab-
rication this backing film is removed before
the fiber reaches the fiber placement head.
When selecting a tow that is to be used by a
fiber placement machine, it is important to
consider the cross sectional area of the fiber.
This cross sectional area will be a major factor
in determining the ply thickness of the lami-
nate. This is because the width of the tow is
fixed and the thickness of the tow varies
depending on the cross sectional area of the
fiber and the resin content.
The following equations can be used to
determine the thickness of a given tow. Since
Fig. 22.3 Fiber placement maclUne dual station.
resin content is usually given as a percent of
resin by weight, it is first necessary to calculate
the resin volume fraction, which is the percent
of resin by volume.
22.2 MATERIALS 1
(22.1)
rvf = ( rd )( 1 )
A fiber placement machine can dispense
prepregged fibers that are commonly used by
1+ if (1/(1 - rwf») -1
the aerospace industry such as carbon, Kevlar® where rvf = resin volume fraction, rwf = per-
and glass. These fibers need to be impregnated cent of resin by weight fraction, rd = resin
with a resin and formed into tows or slit tape. density (g/mm3), if = fiber density (g/mm3).
The width of tow or slit tape used by fiber Using the resin volume fraction the thick-
placement range from 3.2 mm (0.125 in) to ness of fiber can be calculated using the
6.4 mm (0.250 in) After the tows are impreg- following equation:
nated with a resin matrix they are flattened to fa
a desired width and wound onto a 7.6 cm (3 th = fw (1 - rv/) (22.2)
in) diameter by 28 cm (11 in) long core in a
helical pattern. A typical length for a 2.3 kg (5 where th =tow thickness (mm),fa =fiber cross
lb) spool of prepregged IM7-12K tow, 3.2 mm sectional area (mm2),fw = tow width (mm), ruf
(0.125 in) wide, is 3350 m (11 000 ft). = resin volume fraction.
Slit tape is fabricated by running a 7.6 cm For example the thickness of a Hercules
(3 in) wide tape through a slitter creating IM7 fiber impregnated with a Hercules 8551-7
smaller widths of slit tape. These slit tapes are resin containing a 32% resin content by weight
then wound onto a number of 7.6 cm (3 in) and flattened to 3.2 mm (0.125 in) would be
diameter by 28 cm (11 in) long cores. When the 0.14 mm (0.0055 in).
slit tape is wound. onto the spool a backing To increase or decrease the thickness of the
film which is wider than the slit tape must be tow, the cross section of the fiber must be
added. If the backing film is not used, the slit changed by adding or subtracting 6K or 12K
tape cannot be removed from the spool fiber bundles. With slit tape this is not a prob-
because of stringers, which will occur during lem because the thickness of the tape is
the despooling operation. The edge of the slit determined when it is prepregged and before
tape separates and stays on the spool while the it is slit to width.
 Tooling considerations 479

The tow width of the material is very guided through a fiber delivery system and
important in controlling the gap between the head, but high tack is needed when it is being
prepregged tows. For example if the fiber compacted onto the surface.
placement head is designed to lay down tows Materials that have a low tack can be
that are 3.2±O.38 cm (0.125±O.0l5 in) wide, the despooled with a fiber tension of 0.45 kg (lib)
tows will be compacted onto the surface in or less. These low tensions are achieved
3.2 mm (0.125 in) spacings. If the tow is exactly because the resin does not stick to the spool or
3.2 mm (0.125 in) wide, there will be no gap the components of the fiber delivery system.
between the tows. If the tows are 2.5 mm This lower fiber tension is needed while fiber
(0.100 in) wide, there will be a 0.63 mm placing concave areas. A higher tension will
(0.025 in) wide gap between the tows. If the cause the fiber to bridge over concave areas.
tows are 3.8 mm (0.150 in) wide, there will be Materials with low tack levels also have less
a 0.63 mm (0.025 in) overlap. tendency to deform or rope while being pulled
Figure 22.4 shows a typical width his- through the fiber delivery system. They also
togram of a spool of tow material. To develop transfer less resin to the components of the
this histogram a spool of tow was run through fiber delivery system and head. This reduces
a measuring device, which measured the the number of times that these components
width at 1100 locations. The average width is need to be cleaned because of resin build up.
3.20 mm (0.126 in) and the standard deviation Resin build up in the head causes it to mal-
is 0.13 mm (0.005 in). A typical width his- function.
togram of a slit tape will show an average The tack of most resins can be reduced by
width of 3.18 mm (0.125 in) and a standard lowering their temperature. To accomplish
deviation of 0.08 mm (0.003 in). Maintaining this, most fiber placement machines are
the quality of tow and slit tape will greatly equipped with an air conditioned creel. The
improve the quality of the composite laminate fiber placement head also has some means of
and the reliability of the process. cooling the components that come in contact
The ideal fiber placement material has no with the tow material. To achieve the desired
tack at 21°C (70°F) and high tack from 27°C to tack required to adhere the tow material to the
32°C (80°F to 90°F). Low tack is needed when surface, the area where the material is com-
the material is being pulled off the spool and pacted onto the surface is heated. This is
accomplished by blowing hot air at the nip
point of the compaction roller (Fig. 22.1).
Average: 3.20 mm
Std. Dev: 0.13 mm 22.3 TOOLING CONSIDERATIONS
.
........................ .
Maximum: 3.56 mm ..
Minimum: 2.74 mm The tooling for fiber placement provides the
R
c # of Samples: 1100
shape of the part being fabricated and a suit-
~. able surface for the machine to lay tows onto.
~ rt also has features that can be used to associ-
ate the electronic data's coordinate system,
which describes the laying surface, to the
machine's coordinate system. The electronic
data must accurately represent the finished
2.03 2.54
tool. The smaller the surface of curvature is,
Tow Width (mm)
the more important it is that the electronic data
accurately represent the finished tool. If the
Fig. 22.4 Tow width histogram. electronic data does not accurately represent
480 Fiber placement

the lay-down surface the tows will be placed

in the wrong location and there is a chance
that the machine could collide with the tool or
part. It is also important that the electronic
data is in a format that can be used by the
offline programming system. If data is in the
wrong format, the length of time it takes to
program the part is increased and sometimes
there is a loss of surface accuracy. If no elec-
tronic data exist then the tool needs to be
accurately measured and this information
needs to be converted into electronic data that
can be used by an offline programming sys-
tem.
Tooling for fiber placement needs to meet
criteria that are not necessarily required for
other processes. The tool surface must have
sufficient strength so that it does not deflect as
the head compacts tows onto it. When the tool Fig. 22.5 Tooling features.
surface deflects, the tows will not be placed in
the correct position and they will buckle after
the tool springs back into position. Tows that having to rotate the mandrel and they must
are buckled will no longer carry the desired also be described in the electronic data coor-
design loads. Another consideration is that the dinate system used by the part programmer.
tool surface needs to extend 152 mm (6 in) past The probed coordinates are used by the offline
the largest part boundary. This extended area software to create a transformation matrix
is used by the head to roll off and onto the part that will convert the part's electronic data
as it is starting and ending a course. This coordinate system to the position of the tool in
extended area also needs to be part of the elec- the machine's coordinate system.
tronic data, because the offline software also If the tool has to be removed from the
uses this area as it is generating the courses machine and then replaced in after each part,
when deciding where the ply boundaries are. it is important that the tool is designed with
This extended area is also used as a place to features that allow it to be located in the same
apply the bagging sealant tape for vacuum position each time it is put back in the
debulking (Fig. 22.5). machine. This will eliminate repeating a time
The tool must also be designed with fea- consuming alignment procedure. If the tool is
tures that can be used to associate the part's mounted between a headstock and tailstock it
electronic data coordinate system to the is desirable to attach the ends of the tooling
machine's coordinate system. This can be shafts to the headstock and tailstock with non-
accomplished by perpetually molding or locking tapered adapters. This makes it easier
machining three cross hairs onto the tool sur- to load the tool and when the tapers are tight-
face, in the area of the 152 mm (6 in) ened they are self locating. The tapered
extensions. To provide accurate results, the adapters also need to incorporate a keyway on
cross hairs need to be placed as far apart as the headstock end so the tool can be located in
possible. These three cross hairs are probed by the radial direction. A tool mounted to a flat
the machine to detect the location of the tool plate should be located by a three-point locat-
in the machine's coordinate system, without ing system incorporating pins or keys.
 Ply shape 481

22.4 PLY SHAPE length. The exterior ply boundaries can the
extended in the areas where there are mini-
Fiber placement is used to fabricate simple
mum cut length problems. These extended
parts such as flat panels or complicated parts
areas can later be cut off. Interior plies must be
such as an inlet duct, which is square on one
reshaped to match the fiber angles.
end and round on the other (Fig. 22.6). The ply
The designer must also consider the end
shapes can be any size or geometric form. A
ply shape can also include interior ply bound- locations of the tow. Each tow is cut at a 90°
aries which create holes. angle; because of this the ending angle of the
tows may not match the angle of the ply
When generating ply shapes, the designer
must consider the shortest tow length the boundary. Figure 22.8 shows three different
ply boundary conditions for a 45° ply. The
machine can lay down. This length is the dis-
tance from the start of the lay-down point to
where the tow is cut in the head. Figure 22.7
shows exterior and interior ply shapes of an Outer Ply BoundolY ..s'

Boundaries With Missing Tows

Adjusted Boundaries With No Missing Tows

Fig. 22.6 Fiber placing an inlet duct. Fig. 22.7 Ply boundary adjustments.

aircraft structure. The black areas are the areas amount the tows can cover the ply bound-
that cannot be fiber placed because the length aries is specified by the percent of ply
of the required tow is less than the shortest boundary coverage. A 100% value is typically
allowable fiber length. These areas could be used on the larger ply boundaries that are
laid in by hand, or the ply shape could be later trimmed to shape. This makes sure that
adjusted, as in Fig. 22.7, so that the required all of the fibers go to the edge of the bound-
fiber length is longer that the shortest cut ary. A 50% value is used on interior ply
482 Fiber placement

the head laminates over a contour, some tows

must dispense more length than others. This
allows the fiber placement machine to steer
PLY BOUNDARIES the fiber band.
Steering can be defined as the binormal
radius of curvature along a fiber path. An
example (Fig. 22.9) shows 0°, 45° and 90°
courses laid on a cone shaped mandrel. On
the right side of the figure, the same courses
are shown on the unwrapped mandrel sur-
face. On the unwrapped mandrel surface it is
easier to see which courses are being steered.
The 0° course is straight showing that no
steering is necessary. On the 45° course the
amount of steering changes as the course is
0% 50% 100% laid from the larger end to the smaller end of
the cone. The steering radius on the larger end
is 58.4 cm (23.0 in). It decreases to 20.3 cm
(8.0 in) on the smaller end. On the 90° course
Fig. 22.8 Ply boundary conditions.
the center line steering radius of the 2.54 cm
(1.0 in) wide course is 28.7 cm (11.3 in). The
90° course also illustrates the need for differ-
boundaries. The 50% value allows some of ential payout. The fibers on the inside steering
the fibers to go past the ply boundaries; but it radius are 32.14 cm (12.653 in) long. The fibers
makes sure that the correct volume of mater- on the outside steering radius are 35.11 cm
ial has been placed in the ply. The percent of (13.821 in) long. This is a 2.97 cm (1.168 in)
ply boundary coverage is specified in the difference in length.
offline programming software. Tape is typically laid along it's 'natural
path' , but can be steered by small amounts. A
tape laying machine can steer a 7.6 cm (3 in)
22.5 STEERING
wide tape along a 20.3 m (800 in) radius with
The ability to steer the fiber band allows the small amounts of buckling. The buckling
designer to maintain a constant fiber angle on occurs because the fibers on the outside steer-
a complex surface, or align the fibers with ing radius are in tension and the fibers on the
some applied stress. Steering is made possible inside steering radius are in compression. A
because of differential payout and because the typical fiber placement machine using
impregnated tow has enough tack to over- 3.2 mm (0.125 in) wide materials can steer a
come any sliding forces. Differential payout fiber band along a 63.5 cm (25 in) radius
of the individual tows is one of fiber place- without buckling the individual tows. When
ment's key features. This feature is required to steering a radius smaller than 63.5 cm (25 in),
allow each tow within the fiber band to main- the tows will begin to buckle if laid on a flat
tain a unique length as the laminating action or a convex surface or 'Venetian blind' if laid
of the head pulls the tows off the spools and on a concave surface. 'Venetian blinding'
compacts them onto the part surface. When occurs when the fibers on the inside steering
the head is required to laminate a curved radius of the individual tows are adhered to
path, the outer tows of the band will pull the surface and the outside steering radius
more length than the inner tows. Similarly, as fibers are in the air.
 Dropping and adding tows 483

a DEGREE COURSE
ON MANDREL
a DEGREE COURSE
FLAT LAYOUT

45 DEGREE COURSE
ON MANDREL
45 DEGREE COURSE
FLAT LAYOUT

STEERING: 27.4 R
LENGTH: 32.14 em

90 DEGREE COURSE
ON MANDREL 90 DEGREE COURSE
FLAT LAYOUT

Fig. 22.9 Steering.

the compactor lays the ends onto the surface

22.6 DROPPING AND ADDING TOWS
of the mandrel.
To start a course, the delivery head is posi- Each tow is cut at a 90° angle. When adja-
tioned to the start point and the tows are fed cent courses are placed next to each other, this
out under the compactor. While following the can leave a triangular gap or overlap (Fig.
fiber path, the delivery head can change the 22.10). The designer can use the off-line soft-
fiber band width by adding or dropping indi- ware to specify the percent of overlap between
vidual tows. This feature allows all or any courses. A zero percent specification will leave
combination of tows to be removed or added a triangular gap between courses. In this case
to the band width in increments equal to one the outside tow of the course being laid is
tow width. With this capability, it is possible dropped when it touches the adjacent course.
to decrease the band width to prevent adja- On the other extreme, a 100% overlap will
cent courses from overlapping each other. A leave a triangular overlap and no gaps. The
course is ended by cutting all of the tows as 100% overlap specification can leave a part
484 Fiber placement

TRIANGULAR GAP

, ,
/;// 0% OVERLAP
,.'
~~~"~
,,#'

50% OVERlAP

TRIANGULAR vvc."'U"t' -

100% OVERlAP

Fig. 22.10 Overlap between courses.

'bumpy' because of the overlaps. The 50% which is to be manufactured by fiber place-
specification is most commonly used because ment. The first is concave surfaces and the
the smaller gaps and overlaps tend to average second is areas with small radii of curvature.
out each other as the laminate is built up. The When considering a part with a concave sur-
other technique that helps average these trian- face area, the designer must make sure the
gular gaps and overlaps is to offset each ply by fiber placement head can fit into the concave
one and a half tow widths. This ensures that area without hitting the surface of the part.
the small triangular gaps and overlaps do not There are some techniques that can be used to
align on top of each other. overcome some of these limitations. To help the
head fit into a concave area the offline software
has a feature known as collision avoidance.
22.7 SURFACE GEOMETRY
The software knows the part and head geome-
Fiber placement can automate the fabrication try. It constantly checks to see if the two are
of many composite part geometries that in the colliding. If they collide, the software will rock
past could only be laminated using hand lay- the head off the surface normal away from the
up. The types of surface geometries that can be collision. Figure 22.12 shows a part with a
fiber placed range from fan blades to full 3600 small concave area being fiber placed. As the
asymmetrical shells (Fig. 22.11). There are two head stays normal to the surface and
surface features that the designer needs to pay approaches the concave area, it will collide into
special attention to when designing a part the part, as shown in the upper exploded view.
 Surface geometry 485

Fig. 22.11 Fiber placing a fan blade.

SURFACE NORMAL
COLLISION POINT

TOOL SURFACE

FIBER PLACEMENT
HEAD

NO COLLISION AVOIDANCE

SURFACE NORMAL

10° COLLISION AVOIDANCE

Fig. 22.12 Collision avoidance.

486 Fiber placement

To correct this problem, the software rocks 22.8 INSPECTION

back the head 100 off normal, as shown in the Fiber placement is a very repeatable process
lower exploded view.
and requires only a small amount of in-process
There are limits to how much the head can
inspection, but it is important to perform a rig-
be rocked off the surface normal. If the head
orous first article inspection. This inspection is
hits on both front and the back sides, the soft-
used to verify that the part program fabricates
ware cannot avoid the collision and the area a part that meets all of the design require-
should be redesigned. Rocking the head to the
ments. This detailed first article inspection
front and back slightly affects the effective
should not need to be repeated unless the part
applied compaction force and the minimum
program has been changed.
cut length. Rocking the head sideways also
The first article inspection should start with
affects the effective applied compaction force
a dry run of each ply. Dry running a program
and requires extra compactor compliance.
means to run the machine through the pro-
Examples of small radii of curvature are
gram without laying down tows. Dry running
shown in Fig. 22.13. On the convex surface
of a program verifies that the electronic data
the compactor cannot completely come in

SMALL CONVEX RADIUS

Fig. 22.13 Small radii of curvature.

contact with the surface. The part can still be describing the tool surface matches the actual
fiber placed but the number of tows in the tool surface. It also verifies that the transfor-
fiber band should be reduced to match the mation matrix, used to associate the part's
compaction line width. On the concave sur- electronic data coordinate system to the tool's
face the radius should be increased to allow position in the machine's coordinate system, is
the compactor to contact the surface, other- correct. This is accomplished by watching the
wise the tows will bridge over the small compactor as it follows the part surfaces. Dry
radius of curvature. running also verifies that collision avoidance
worked properly.
 Further reading 487

Table 22.1 Inspection criteria

Discrepancy Criteria Correction procedure

Gap A gap that is greater than Add a tow to fill the gap.
2.54 mrn (0.1 in) wide and
is longer than 76.2 mrn (3.0 in).
Wandering tow at start A wandering tow that leaves a Pick up the tow and reposition.
and end of course gap greater than 2.54 mm (0.1 in)
wide and is longer than 50.8 mrn
(2.0 in). If the tow is captured under
another tow, do not repair the
captured tow.
Twisted or folded tow A twisted or folded tow that is longer Remove twisted or folded tow and
than 76.2 mrn (3.0 in). replace.
Wrinkles at start of course A wrinkled tow that is wrinkled more Pick up the wrinkled end of the tow
than 19.05 mrn (0.75 in). and straighten it out.
Bridging of tows Bridging is more than 1.57 mrn Use a heat gun and hand pressure
(0.062 in) high. to remove bridging or vacuum bag
the part for a maximum of 15 min at
full vacuum.

The next step in first article inspection is to scale and lay it on the ply normal to the fiber
load the machine with tows and fiber place direction. All the gaps along the 30.5 cm (12 in)
each ply. After each ply is fiber placed, the ply distance are measured and summed. If this
is inspected for the correct fiber angle, ply loca- value exceeds a pre-determined design crite-
tion, band-ta-band overlap and for missing rion the area is reworked.
tows. The easiest way to inspect for fiber angle
and ply location is to have a Mylar template FURTHER READING
that has the ply boundary plotted onto it and a
line for the fiber angle. The template needs to Barth, James R. 1990. Fabrication of Complex
be located with alignment marks or pins that Composite Structures Using Advance Fiber
Placement Technology. 35th Intern. SAMPE
are part of the tool. A Mylar template will not Symp., 2-5 April 1990.
work on surface geometries with curvatures in Enders, Mark 1. and Hopkins, Paul C. 1991.
both directions. For these geometries, a formed Developments in the Fiber Placement Process.
inspection tool should be fabricated. 36th Intern. SAMPE Symp., April 1991.
After the first article inspection, subsequent Enders, Mark L. 1991. The Fiber Placement Process.
parts need to have each ply visually inspected Intern. Conf Camp. Mater., (ICCM/8), July 1991.
for excessive gaps and overlaps, lost tows, Evans, Don 0., Vaniglia, Milo M. and Hopkins, Paul
C. 1989. Fiber Placement Process Study. 34th
twisted tows, wrinkled tows and bridging
Intern. SAMPE Symp., 8-11 May 1989.
tows. Table 22.1 is an example of a typical in- Evans, Don O. 1993. Design Considerations for
process inspection criteria. Another criterion Fiber Placement. 38th Intern. SAMPE Symp.,
used to evaluate gaps is to take a 30.5 em (12 in) 10-13 May 1993.
PULTRUSION 23
Brian A. Wilson

23.1 INTRODUCTION ment and manufactured the equipment to pro-

duce structural elements by the method. A
The word pultrusion is used to describe a com-
typical pultrusion machine is shown in Fig.
mercial fabrication process for the production
23.1. The process has a relationship to extru-
of fiber reinforced composite elements. First
sion, which is used primarily with metals and
mention of the process is recorded in a patent
describes the process of forming a shape using
in 1951 with much of the early work in the
a closed die and pushing normally hot metallic
1950s attributed to W. Brandt Goldsworthyl.
materials through the die. Pultrusion differs in
He performed much of the process develop-

Handbook of Composites. Edited by S.T. Peters. Published Fig. 23.1 Typical pultrusion machine. (Courtesy of
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 W. Brant Goldsworthy and Associates Inc.)
 The basic process 489

that it takes advantage of the strength of the curing energy in the heated die to cure the com-
longitudinal fibers in the section to pull the posite into a hard structural product and using
extruded shape through the forming die and the resulting shape as a unidirectional strength
the heated curing die to create a constant cross element for commercial structures.
section structural element from a composite The primary advantages of pultrusion are
materials system. Hence the name pultrusion. as follows:
Use of this process has now expanded both in
the USA and around the world with many • production is continuous;
manufacturers equipped to produce the simple • labor requirements are low;
structural elements which are the main pultru- • material scrap rate is low;
sion products. The process was labeled for the • the requirement for support materials is
first 30 years or so of its existence as a com- eliminated, i.e. breathers, bleeder doth, sep-
mercial production method. This was owing to arator film, bagging film, edge tape.
the nature of the method, using longitudinal
fibers and a simple thermoset resin system to
23.2.1 VARIATIONS IN PROCESSING
produce a structural shape which had its pri-
mary strength in the longitudinal direction and The original process of pulling a bundle of
properties in the transverse and third axis unidirectional fibers through a curing die
relating only to those of the resin system. The remained without variation until the middle
ability to incorporate three-dimensional of the 1960s. At this point, pultruders, both in
strength first occurred in the mid 1960s when it Europe and the USA, developed the process of
became possible to include layers of mati fab- adding fabric in strip or sheet form and fiber
ric and circumferentially wound layers of fiber mat to the pultrusion system in order to pro-
both within the body and on the surface of the vide transverse strength and shear strength in
structural shapes. Multi-directional reinforce- the comer sections of complex pultruded
ment was used extensively starting in the shapes. In addition, the concept was devel-
1970S2,3. These developments, plus the intro- oped for a filament winding head to be added
duction of thermoplastic resin matrices, have to the pultrusion machine to provide a hoop
brought the pultrusion process into the manu- wrap around a pultruded form. This process is
facturing arena of the aerospace, defense and known as 'pull winding' and is primarily used
aircraft businesses as a relatively inexpensive with tubular or rod shapes. It is not feasible to
and repeatable method of producing a constant add the filament winding process to a struc-
cross-section structural element4, 5. tural shape that has a concave curvature, such
In this chapter, the basic process, variations as angles, channels and I-beams. A typical
on the process, pultrusion equipment, materi- process of pull winding is shown schemati-
als including fiber, fabric, resin matrices, cally in Fig. 23.212.
additives, tooling, curing, structural connec- During the 1970s, the concept of pull form-
tion methods, equipment manufacturers, ing was developed by Brandt Goldsworthy.
pultrusion fabricators and potential markets Three versions of this forming process were
and applications will be reviewed6-11. developed.
The first was actually named'curved pultru-
sion' by Goldsworthy Engineering. The method
23.2 THE BASIC PROCESS
was developed under a NASA contract and
The basic pultrusion process as originally con- resulted in a pure, constant radius section with
ceived consisted of creating a cured composite a constant cross section. A curved die was used,
shape by pulling a bundle of resin impregnated with a reduced radius, smaller than the
fibers through a heated die, providing sufficient required part radius. This allows for a degree of
490 PuItrusion

COUNTER-ROTATING
WINDING HEADS
MANDREL
AND
SUPPORT

LAYER 1. LAYER 2. LAYER 3. LAYER 4. LAYER 5. FIBERS

WOUND FIBERS FIBERS WOUND FIBERS
FIBERS

Fig. 23.2 5-Layer tube on double head pull winder. (Courtesy of Pultrex Ltd.)

spring back after the product is removed from response to a specific production require-
the die13. The die was split along its length, in ment14, 15. The pull forming process still uses
the vertical plane. One half of the curved unit the basic principles of pultrusion since it has a
was fixed in place. The internal radius portion continuous strand of fiberglass roving within
of the die was further split in half vertically, the product and this roving is used to pull the
along the central radius plane, forming two product through the sequence of the forming
quadrants. Each of the quadrants had a slightly process. Goldsworthy Associates have
longer circumferential length than the fixed designed two versions of this process, a
half. The two moving quadrants moved past curved pull forming and a straight pull form-
the fixed part of the die at the processing speed ing. These two versions of the process are
of the pultrusion. When each moving quadrant briefly described in the following paragraphs
had exited the plane of the fixed die, it moved but it should be noted that a major difference
rapidly in a circular fashion to contact the end between pull forming and pultrusion is that
of the following quadrant as it too passed pultrusion is a generic process which can be
across the face of the fixed die. used to make many different cross-sectional
shapes and products by changing the die in
the machine (which creates the cross-sectional
23.3 PULL FORMING
shape) and the materials of fabrication. Pull
The pull forming process is a highly sophisti- forming however, is essentially a custom
cated variation of the pultrusion process in process which must be tailored to a particular
which both curved and straight product sec- product design.
tions are fabricated with the ability to change
cross-sectional dimensions of the product. It
23.3.1 CURVED PULL FORMING
can be thought of as a combination of pultru-
sion and compression molding. The pull In the curved version of the pull forming
forming process was developed by process, a curved structure is manufactured
Goldsworthy Associates and was created in using a selected combination of roving and
 Pull forming 491

mat/fabric to satisfy the performance require- machine is shown in Fig. 23.3.

ments of the product16,17. The die is a multi- It should be noted that the process does dif-
cavity die. Several of these dies are mounted on fer from pultrusion in that the fiber preform is
the face of a wheel and are curved with the placed into the mold rather than being pulled
radius of curvature of the die matching the through it. Also, while the cross section of the
radius of its position on the wheel. The die is mold can change, the resulting part is a con-
open faced and is also open on the two ends for stant volume design, equivalent to the volume
entry of the fiber preform. The preform is of the preform which is initially compressed
impregnated with resin in the same manner as into the mold.
for pultrusion and the fiber and mat/fabric
combination is pulled into the curved mold or
23.3.2 STRAIGHT PULL FORMING
die by the rotation of the wheel. As the preform
is pulled into the die, a stainless steel band Straight pull forming was also invented by
moves against the open face of the die and com- Goldsworthy Associates in response to a spe-
presses the impregnated preform into the cific customer requirement for the automated
contour which is in essence a compression manufacture of hammer handles. In this
molding cavity. After the die is closed, it is process, the fiber is pulled as a preform through
heated to cure the resin and following cure, the the automated machine. Many single cavity
stainless steel band peels away from the face of dies are mounted on a belt. As the fiberglass
the die/mold. A fixed pin which is mounted roving moves through the system, a section
adjacent to the wheel strips the product out of of bulk molding compound (BMC) is cut
the mold and into a storage bin. When the stain- from a rope of this material, is opened up
less steel band is moved against the mold it is longitudinally and placed around the roving
held in place by an automated clamping system and clamped onto it. The BMC is then com-
until the product is fully cured at which point pressed in a small briquetting press which
the clamping system releases the band which travels with the fiber preform as it moves
then retracts into a 'parking' position. The through the system. Following the compression

Fig. 23.3 Curved pull forming process. (Courtesy of W. Brandt Goldsworthy and Associates Inc.)
492 Pultrusion

molding of the BMC, the press releases and the process variation has not been successful to
balance of the fiberglass roving is fed in and date and further development will be required.
encapsulates the BMC material. Following the
encapsulation, a shrink film is placed around
23.4 PROCESS EQUIPMENT
the wet package and the process goes through a
heated tunnel which shrinks the film around The pultrusion fabrication machine consists of
the product. Finally, the product is run through six different parts (Fig. 23.4): the creel; the
the frame of a C-shaped die in a press which resin bath; the forming die; the heated curing
provides the final compression molding and die; the pullers; the cut-off saw.
curing of the hammer handle. The product is
pulled from the die and cut off and the dies con-
23.4.1 CREEL
tinuously exit the belt and are recycled back to
the front end of the process. The creel is the beginning of the pultrusion
This type of pull forming is a highly auto- process and is the material storage system from
mated, sophisticated process which requires which the fibers and mat, veil or fabric are
considerable capital investment for the design, drawn in the correct sequence to match the
manufacture and set up. Each potential prod- design requirements of the structural shape.
uct, which would be a candidate for the pull Since pultrusion is a long run continuous
forming process, must have its design analyzed process, fiber rovings are provided in the max-
and the process modified to respond to the cus- imum size configuration possible. Continuous
tom requirements of the particular product. glass rovings are normally provided in 'center
With the increasing popularity and explo- pull' packages between 14 and 23 kg (30 and 50
ration of the resin transfer molding process in lb) in size. These center pull packages are typi-
recent years, the pultrusion process has been cally stored on a bookshelf style creel. These
reviewed for the potential of developing a creels have from three to six shelves and are
repeatable, precise cross-section of a preform capable of storing anywhere from 45 to 120
using a loose fiber bundle and an adhesive packages of this type of fiberglass. These creels
binder in place of the resin system. This are normally mounted on casters to create a

Mat
Roving racks
racks

Pulling
mechanism Cutoff saw

Resin tank Finished

Preforming product
guides

Fig. 23.4 Schematic of pultrusion process. (Courtesy of W. Brandt Goldsworthy and Associates Inc.)
 Process equipment 493

mobile system. The casters are provided with a plied on 'way wound' packages on standard
foot locking device to enable them to be locked diameter cardboard cores. A typical carbon
in place when this is required. The glass roving fiber package mounted in a payoff creel is
is pulled vertically from the package through shown in Fig. 23.5. These fiber packages are
ceramic-lined holes in the shelves above. The designed to provide fiber payoff from the exte-
roving is collected above the creel and turned rior of the fiber package and hence avoid
90° by means of a ceramic textile type thread twist. While fiberglass is normally supplied in
guide and then moved forward to the material the heavy center pull spools, it can also be
accumulating section just prior to the resin obtained in the outside payoff type package.
bath and forming die. A series of ceramic All of the cardboard cores are a standard three
guides or rollers can be provided in the fiber inch diameter with the exception of the aramid
path in order to tailor the tension in the fibers which is 90 mm (3.5 in) inside diameter.
to the required level. The pulling of fiber from This style of package requires the use of a
the center of the package will automatically multiple spindle creel design in which the
insert a twist in the fiber as it is led into the pul- packages are oriented normally horizontal but
trusion machine. To eliminate this, some usually with a slight upward camber. Creels
fiberglass rovings are provided in a center pull are available with package positions up to
twistless condition where the natural twist has 1500 or 2000 on a single creel or combination
been off-set by a 'built-in' reverse twist. of creel frames. A loaded multiple spindle
Continuous fibers of fiberglass, carbon, creel is shown in Fig. 23.6. Ceramic guides are
aramid and polyethylene are normally sup- used to move the fibers to the front of the creel

Fig. 23.5 Carbon fibre spool on package holder with quick braking action. (Courtesy of Texkimp Ltd.)
494 Pultrusion

(12-24 in) on cores of 75-100 mm (3-4 in) in

inside diameter. This special mat/fabric creel
must be able to accommodate rolls of these
dimensions and resulting weight and permit
the rolls of material to be installed in a varying
sequence of locations to match the design of the
final structure. This type of material creel will
normally provide horizontal feed. However, if
vertical mat or fabric is required, then an inde-
pendent custom unit must be provided. These
units are usually of a carousel-type configura-
tion. Figure 23.7 shows a typical mounting for
feeding veil rolls into the system.

Fig. 23.6 Loaded multiple spindle creel. (Courtesy

of Texkimp Ltd.)

and into the pultrusion system. Tension

requirements in this type of system are usually
provided by spring attached tension straps
which rub on a pulley attached to the package
holder. This provides a braking action. The
straps can be either independently loaded
with weight or can be connected to a central
mechanical system, such that tension for all of
the spools can be varied by a single adjust- Fig. 23.7 Feeding external veil material into pultru-
sion system. (Courtesy of Creative Pultrusions Inc.)
ment handle. Fiber packages for this type of
system are normally 5 kg (11 lb). With this
small size of package, it is normal for a large Clearly, the overall creel system for materials
number of packages to be located on this type supply must be able to provide any combina-
of creel assembly. In addition, package simula- tion or arrangement. It should be noted that in
tion bars are normally provided with these every fiber and fabric system used for pultru-
types of creels to give a uniform tension and to sion, there should be sufficient continuous
eliminate the possibility of a fiber from a small roving in the system to sustain the required
or 'almost used' package touching and abrad- pulling force. As the various materials travel
ing a near-by forward mounted, full package. forward towards the resin application area it is
Stationed immediately after the fiber creel is important to control the alignment of the vari-
the specialized type of creel which is designed ous fibers and fabric/mat strips which are
to accommodate rolls of mat, veil or fabric. going into the configuration. This will prevent
These materials are usually supplied in roll knotting and twisting of the fiber reinforce-
form with diameters between 305-610 mm ment and also will ensure that the various fiber
 Process equipment 495

materials remain in the correct relationship to bath is frequently used which matches the pre-
each other and are placed in the correct zone of form shape or orientation of materials which
the pultruded product according to the prod- pass through it. These types of tailored chan-
uct design requirements. This can be best nels or baths can also be used to orient the flat
accomplished by the use of orifice plates, creel materials properly. This method permits the
cards combs and rollers (grooved and flat) to resin impregnation to take place without mov-
precisely and accurately place all the materials. ing the reinforcements away from their
The materials commonly used for these guides optimum path or shape.
and rollers are primarily titanium oxide
ceramic, stainless steel, tool steel alloys and
23.4.3 VERTICAL PULTRUSION
chrome plated steel. These materials are also
used for the various rollers associated not only The vertical pultrusion process should be men-
with guiding the sheet materials but also with tioned at this point since the primary difficulty
controlling them during their passage through in creating a vertical pultrusion is the place-
the resin impregnation zone. ment and use of a resin bath. In the
arrangement for vertical pultrusion the equip-
ment used is essentially similar to that for the
23.4.2 RESIN IMPREGNATION
horizontal method18 • Most equipment operates
Virtually all pultrusion processes utilize a in the vertical position with the exception of the
resin impregnation bath to facilitate the resin bath and roving creels which are generally
impregnation of the resin into the fiber struc- positioned horizontally and the roving is fed in
ture. The position of the resin impregnation the regular manner, horizontally through the
bath in the production line can be varied and resin impregnation bath. This bath is located
the manner in which the resin is applied to the above the entrance to the forming die and the
fiber can have many different versions. A resin fibers exiting from the bath are turned 90°
dip bath is most commonly used. During this across a roller and then proceed vertically into
process the fibers are passed over and under a the forming die. The advantage of using this
series of rollers or bars which both spread the vertical procedure is that a uniform arrange-
fiber to more easily accept the resin and also ment of impregnated fibers can be achieved
provide a massaging effect which 'kneads' the across the section being formed. The effects of
resin into the fiber bundles and structure. gravity are removed from the fiber arrange-
The wetting speed of the fibers depends ment. During the horizontal process, gravity
upon their pretreatment and on the resin for- plus any nonuniform tensioning across the
mulation. Wetting is also affected by the type fiber group will result in some fibers sagging
of sizing agent on the fiber, the possible pres- under their weight with resin and hence not be
ence of remaining lubricant on woven fabric properly located as they enter the forming and
and finally the type of binder which is used in curing dies. An additional advantage of the ver-
mat and veil products. The resin bath is uni- tical process is that it is possible accurately to
formly used for products that utilize all roving locate the internal mandrels which are neces-
in their construction or for products that are sary for hollow shapes and tubes. This is
easily formed from the flat fiber ply which particularly important in the fabrication of
emerges from the resin bath. However, in heavy, thick wall tubes. In the horizontal
many of the more sophisticated products process, these internal mandrels will frequently
which are now made from pultrusion, it is deflect under their own weight and cause a
impractical to dip all of the materials in the nonuniformity of wall thickness around the
resin bath. When vertical mats are required or hollow profile or tube. In addition, the vertical
hollow profiles are produced, a tailored resin process allows the use of several different fiber
496 Pultrusion

entry points to the forming process with multi- fabricators to provide pultruded tubular struc-
ple resin baths. In the horizontal procedure, the tures both round and square23, 24. The process
multiple entry points have to come from the side combines the standard continuous unidirec-
of the main ftber path and it is not as easy to coa- tional fibers of the pultrusion process with
lesce the fibers into a single merged preform. hoop wound continuous fibers. The longitudi-
nal fibers are used for axial and bending
resistance while the hoop fibers are used for
23.4.4 USE OF THERMOPLASTIC MATRIX
hoop tension and compression resistance. The
RESINS
combination of the two processes of pultru-
Thermoplastics cannot be applied to roving sion and winding provides virtually unlimited
easily using the standard resin bath, even with possibilities. However, the increasing com-
heating. Thermoplastic resins impregnated on plexity of the process limits these
roving are generally available as preimpreg- combinations. Normally, longitudinal fibers
nated (prepreg) materials and are supplied by are positioned at the inside and outside sur-
specialty companies who are experts in the faces of the tube with one or two hoop wound
prepreg process. Thermoplastic matrices layers positioned internally in the tube wall.
improve the toughness of the composite and These hoop wound layers are not truly 900
this is their major end use advantage. In gen- hoop wound layers because of the slightly
eral they have high softening temperatures helical nature of the winding and the lateral
and high physical properties coupled with a movement of a wound roving, one band width
low fluid viscosity in their melted form. The of advance with each circumferential pass.
prepregs are normally prepared using solvents Depending on the diameter of the tube, the
and this provides some difficulty against envi- winding angle is typically anywhere from
ronmental regulations in order to remove and 80-870 • The winding can be performed in
dissipate the majority of the solvent. Some sol- both clockwise and counterclockwise direc-
vent has to be retained in the process in order tion and in addition to the hoop tension and
to have the material be sufficiently pliable for compression resistance can also provide a
machine and manual handling. In addition to degree of torque strength to the tubing.
the improved toughness of the thermoplastic Single and double head pull winders are used
composite, an important advantage of thermo- with these techniques. The pull winding
plastic pultrusions is the ability to heat and process is shown in Fig. 23.2 and the proce-
reshape the product after forming l 9--22. dure for the use of these systems is self
explanatory. The important control feature of
the pull winding process is the control of the
23.4.5 RESIN BATH LIFE
winding speed and position with respect to
In a continuous pultrusion process, the pot life the linear motion of the pultruded form
of the resin should be several days. However, through the machine. This is achieved by
if the volume of the resin bath can be kept sensing the speed of the longitudinal motion
small in relation to the resin being withdrawn, accurately and using a microprocessor con-
shorter pot-life resin, i.e. 3-4 h can be used. trol to coordinate the speed of the rotating
Shorter pot-life resins result in smaller batches head motor. This relationship defines the
and mixing becomes time consuming. pitch of the winding which is defined as the
linear distance moved during one revolution
of the head. With the microprocessor control
23.4.6 THE PULL WINDING TECHNIQUE
the pitch can be varied, thus providing varia-
The pull winding process was developed in tions in fiber content and angular position, as
Europe and was used frequently by European required. A paper by D.E. Shaw Stuart25
 Process equipment 497

defines the primary advantages of pull wind- used for forming dies include Teflon®, high
ing over other methods of tube manufacture as: molecular weight polyethylene, chrome plated
steel and a variety of tool steel alloys. The fabri-
• it is a fully automatic continuous process;
cation of the forming die is a custom process
• it is dimensionally accurate and repeatable;
which is best performed at the pultrusion fabri-
• it produces tubes with a good external
cator by a skilled tool maker I machinist.
appearance and finish;
• it can have built-in color;
• it can be made with thinner wall sections
23.4.8 INJECTION PULTRUSION
than conventional pultrusion or filament
winding. In this modification of the resin impregnation
process, the resin is directly injected into either
the forming die or into the initial segment of
23.4.7 PULTRUSION DIES the pultrusion die. The resin is injected into the
die under pressure and is forced into the inter-
Two types of dies are used in the pultrusion
stices of the fiber system. The principal
process: the forming die and the heating or cur- advantage of this system is that it limits the
ing die. Forming is normally accomplished release of volatile resin components and reac-
immediately after the impregnation process
tion products.
although some shaping with the roving and An additional advantage of this process for
mati fabric inserts in a dry condition can take laboratory or research and development pul-
place prior to entering the resin impregnation truders is that it enables a rapid resin change
processing step. Forming dies or guides are without removal and cleaning of all of the
normally attached to the heating or curing die
resin bath components. In addition, dry fibers
in order to provide the correct relationship are not impregnated with resin before entering
between the fOrming and the heated curing the die and therefore they can be positioned
step. For tubular or hollow profile pultruded
accurately, even with complex profile cross
products, a central mandrel support is neces- sections and multiple mat and fabric entries.
sary internal to the fiber form and it is
The schematic of the injection pultrusion sys-
necessary to extend this mandrel as a cantilever tem is shown in Fig. 23.8. A typical resin
through the pultrusion die. It is also important injection pultrusion die is shown in Fig. 23.9.
to resist the forward drag on this central man- In addition to the advantages listed above,
drel which will occur from the surface tension there are several disadvantages which must be
and adhesive forces of the resin on the roving or weighed in a process trade off, before resorting
mat. Materials must form in sequence around to the injection pultrusion method. These dis-
this mandrel and must alternate from one mate-
advantages are:
rial to another to prevent any weak areas at
overlapping joints. The sizing of the slots, holes • a more complicated die design;
and clearances in the forming die must be care- • more parameters to control;
fully designed so that excess tension on the dry • analytical support is required to predict wet
or impregnated fiber is avoided. The fiber is out behavior and design of the resulting
weaker in this condition than in the cured con- dies;
dition and filaments may be independently • fibers in the die are very tightly compressed
broken and distortion of the mat and fabric can and resin penetration is difficult, particu-
take place. The forming die can also be larly with large wall thicknesses;
designed to permit excess resin removaL This • the tightly packed fibers can act as a filter
prevents an abnormally high hydrostatic force and partially remove resin additives, partic-
at the curing die entrance. Materials commonly ularly those in suspension.
498 Pultrusion

Fiber
rack

pulling mechanisms
engaged disengaged

I U

preforming
guides

Fig. 23.8 Injection pultrusion process. (Courtesy of American Composite Technology Inc.)

injection port thermostats

t
23.4.10 POSITIONING OF THE HEATED DIE
?----'Ifr---, The heating/curing die must be firmly
mounted to the frame of the pultrusion
machine in order to react the pulling forces,
which are created in pulling the product
through the process. These forces are typically
in the range of 5440-7250 kg (6-8 tons).
However, some of the larger machines which
matrix are capable of producing parts up to 1.52 m
injection
(60 in) wide and heights of 0.60 m (24 in) (with
wettlng cure
zone zone
narrower parts) can require pulling forces up
to 16320-18 000 kg (8--20 tons). The mounting
method must not only permit these types of
Fig. 23.9 Resin injection pultrusion die. (Courtesy loading but must also be capable of mounting
of American Composite Technology Inc.) height adjustment in order to accurately align
the axis of the die with the pullers which move
the product out of the die. Following align-
23.4.9 HEATING AND CURING DIES
ment and clamping, the mounting system
There are three considerations in the design must also not allow any deflection which
and use of a heated die for the curing of a pul- might provide an angular or dimensional mis-
truded composite section. The first of these is match of the machine.
the positioning of the die relative to the
machine access, the second is the actual
23.4.11 DIE HEATING
method of heating to induce the necessary
energy into the composite material to fully cure Die heating is probably the most critical con-
the resin system. The third includes the design trol parameter of the whole pultrusion
features and construction materials for the die. process. The die heating profile will control
 Process equipment 499

the rate of polymerization of the resin system Use of the platen system will allow change out
and the position of the resin gel point front of dies with very little system down time.
within the die. It also influences the degree of During start up and also during shut down
resin exotherm profile throughout the various periods it is necessary to provide a cooling
wall thicknesses of the pultruded structure. method at the front of the die to prevent early
The curing dies are typically heated with elec- gelation of the resin system. During these
tric strip heaters or hot oil jackets. The thermal phases, the entire die will heat up rapidly
curing using these two methods of heating is without the composite being pulled through it
slow, owing to the fact that the tool steel of the and acting as a heat sink. The cooling can be
die is a poor conductor and limits the rate of done dynamically through cold air impinge-
heat injection to provide a uniform cure. The ment or by using a water jacket or tubes
thermal energy which is required to cure the through which cold water will be pumped.
composite material must all be applied Instead of these dynamic methods, a simple
through the outer surface of the composite static method would be to provide an
shape. The heat input is required to produce unheated zone at the beginning of the die to
gelation of the resin throughout the composite act as a heat sink and conduct away the excess
part. When the composite structure being pul- heat during start up or shut down actions.
truded has thick sections, this necessarily These same unheated or cooled sections may
requires a longer heat input which slows also be useful at the exit of the die to remove
down the pultrusion rate. Increasing the rate contained heat from the product prior to its
of heating to permit a faster pulling speed will exit.
not necessarily solve this problem and may It has been mentioned that the heating pro-
result in premature curing of the outside skin file within the die is the primary factor which
of the profile with early onset of the exotherm controls the speed of throughput of material in
in the resin system and potential overheating the process. This was recognized early in the
and cracking of the exterior surface of the part. development of the pultrusion process by W.
Temperature conditions of the die are con- Brandt Goldsworthy and he added a radio fre-
trolled by the internal placement of quency preheated system to his pultrusion
thermocouples and this can result in individ- machine design26,27. The use of radio frequency
ual control of heating rates in various heating in conjunction with the traditional die
segments of the die. heating methods can significantly increase
Instead of using strip heaters or hot oil jack- running speeds. This technique is basically
ets, it is possible to heat the curing die by limited to unidirectional reinforced rods, bars
means of heated platens and a press. These and shapes and cannot be used if carbon fibers
heated platens will usually have several zones comprise the reinforcement. Data from
of heating control with imbedded thermocou- Goldsworthy Engineering Inc. shows speed
ples to sense the platen temperature. This increases of up to 400% based on a 2.54 mm
method provides a uniform heating condition (0.1 in) wall thickness pultrusion. Speed
to the die. However, since the thermocouples increases of 100% (i.e. doubling the speed) are
are sensing the platen heating temperature accomplished at wall thicknesses of 2.54 mm
rather than the die or product temperatures, (0.1 in).
the temperature differential between these During recent years it has become possible
zones must be well documented. This permits to evaluate the internal die profile with regard
the platen heating temperature to be set to pro- to both temperature and pressure28. The sen-
vide an adequate heating level in the sors which are used to provide that data are
composite product and absorb all of the con- essentially similar to strain gages with output
duction and radiation losses from the system. leads attached to them. The pressure sensor is
500 Pultrusion

a unique development resulting from the heating zones in order to maintain control
music industry. These sensors were originally over the resin gel, curing and cooling of the
used as striking pads in electronic keyboards pultruded section within the die. In addition,
and the manufacturer has now developed if one of the cavities in the multi-cavity tool is
them to measure pressure forces transverse to damaged, the complete tool is not out of pro-
the plane of the sensor. The output from the duction since the remaining useable heating
sensor is transmitted via a pair of thin strain zones can be modified to continue to provide
gage type wires. It is thus possible to insert an acceptable product.
both of these sensors (temperature and pres- The dies are usually made from steel, which
sure) into the fiber resin system prior to can be hardened tool steel, or steel alloy which
entering the curing die. The sensors will then is treated, hardened and plated before use. The
travel through the die internal to the product die must be relatively thick walled so that it
and will record both the temperature and the can be heated uniformly and retain the heat
internal pressure in the die from the entrance input. Thick wall design also reduces distor-
to the exit. Once the sensors have traveled tion under temperature and pressure. It has
beyond the exit from the die, the connecting been suggested that the cross-sectional area of
wires can be cut. The section of product which the steel should be at least ten times that of the
contains the two sensors and the lead wires is pultruded part. The steel used should be hard
then cut out and discarded. The resulting data since it has to withstand the abrasive action of
provides a complete process description from the composite being drawn through it. A typi-
the entrance to the exit of the die. cal hardness is a Rockwell C rating of 30,
which can be obtained with a prehardened
tool steel. Dies are manufactured from multi-
23.5 DESIGN FEATURES AND MATERIALS
ple pieces that are machined and joined
In considering the design of the heated curing together to create a design profile cavity. As
die for a pultrusion system, it is possible to use the various sections of the die are connected
a single cavity tool, a multi-cavity tool or sev- together they must be properly aligned. This
eral single cavity tools mounted in parallel. can be done using dowels for permanent
The choice between these three· will depend alignment or by machining an alignment
upon the size, complexity, dimensional toler- groove on the outside part of the die which
ance and surface quality of the pultruded can then be used for a hard metal key which
product. Individual or single cavity dies are a can be driven into the groove to provide a pos-
frequent choice of pultruders, particularly for itive alignment for the entire tool. The die
medium to large or sophisticated shaped assembly is then match drilled and tapped for
products. Use of the multi-cavity tool or sev- the assembly bolts which hold the parts of the
eral single cavity tools mounted in parallel is die in position against the high internal pres-
generally restricted to very competitive and sure developed during the process. Following
low cost shapes such as round and square alignment and assembly, the exterior surface
cross section rod. The choice between an indi- of the assembled die is ground flat. The inter-
vidual die and a multi-cavity die is frequently nal cavity surfaces are polished using
dependent upon the experience and prefer- polishing wheels and buffing compounds to a
ence of the manufacturing engineer in the high surface finish 0.254-0.762 J..Llll (10-20 !!in).
pultrusion company. The multi-cavity tool can At this point a bell mouth is machined
be an efficient arrangement for production around the entrance of the die to provide a
conditions. In this case the die is two to three smooth entrance for the resin impregnated
times the length of a single cavity tool. The fiber form. The radius will vary in size from
longer die makes it possible to use different a small radius for small simple structural
 Design features and materials 501

profiles to a relatively large radius for a large pultruded profile remains stationery until the
and complex composite profile particularly puller returned to the beginning of its stroke.
where the fiber content is high. The bell mouth Because of the alternating pull and pause
is polished and the finished die is hard chrome mechanism this system was known as the
plated to a thickness of 0.025-0.050 mm 'intermittent puller'. This system is still used
(0.001-0.002 in) with a Rockwell C hardness of on some very early machines, however, it is
approximately 70. The dimensions and surface certainly not in widespread use.
quality of the pultruded product are a direct A modification of the clamping/ pulling sys-
reflection of the condition of the die. Dies will tem has become popular which provides a
thus not be removed from production to be continuous pull. The clamping, pulling and
inspected unless the shape or quality of the unlocking cycles of this system are coordinated
product begins to change. Chrome plated dies by the control system of the machine. The drive
will normally have a life of 61 000-150 000 m system used can be either a hydraulic cylinder,
(200000-500000 ft) of product run with their a threaded ball screw such as is used on lathes,
initial chrome plate application. Up to 305 000 m or a chain drive mechanism. The clamping
(1 x 106 ft) of product have been produced from pads are still formed to match the contour of
some hard chrome plated dies. Chrome plated the profile being pulled. The two puller heads
dies require frequent inspection to insure that must operate in the space originally designed
their internal shape and dimension is main- for the single puller. Consequently, limited lat-
tained. They should also be inspected for wear eral movements of approximately two feet
of the chrome plated surface since the wear each are sustained by the two pullers. The two
process will proceed much more rapidly if the puller system is shown in Fig. 23.10.
tool steel surface is exposed from under the
chrome. Once the die has worn and produces
product beyond allowable dimensional toler-
ances, the die may be replated and repolished
back to original dimensions. This process may
be repeated several times.

23.5.1 CLAMPING AND PULLING

Three different types of clamping and pulling
systems have been used in the pultrusion
industry. Of these three, only two are now
encountered. The original system used on the
earliest machines in the 1950s and early 1960s
employed a single clamp. This clamp was Fig. 23.10 Conventional two puller system.
hydraulically controlled and contoured pads (Courtesy of Pultrex Ltd.)
were used for gripping the part. The unit con-
taining this single clamp was pulled by a
continuous chain for a distance of 3.2-3.7 m Continuous belt pullers are used on basic com-
(10-12 ft). A reversible motor was used to drive mercial machines. These pullers can be used
the chain and following the pull stroke, the with single or multiple cavity molds. The
puller / clamp released the product and cleated chain or caterpillar version of the con-
returned to the beginning of the pulling stroke, tinuous belt machine has many individually
clamped the product and pulled again29 • The contoured puller pads attached along the
obvious disadvantage of this system is that the chain length. The number of these contoured
502 Pultrusion

puller pads depends on the complexity of the aramid fiber to date has been the high pressure
part but generally varies between 12 and 60 water jet and presumably this could be adapted
pads. This large number of pads permits a for use on a pultrusion machine. The inherent
lower unit pressure between the clamping pad disadvantage of the high pressure water jet is
and the pultrusion. The caterpillar type its cost which is from $50000-100000. A com-
machine was designed and built in 1958 by promise solution to this problem would be to
Brandt Goldsworthy, Dennis Franks and Tom cut off the sections as smoothly as possible,
Bailey. Caterpillar type machines are preferred using the diamond saw, to a slightly over
and still widely used in the industry. length condition with very rough ends. The
product length can then be subcontracted to a
waterjet cutting source for final trimming. This
23.5.2 CUT-OFF STATION
will result in some wastage.
Every pultrusion machine utilizes a cut-off saw
to cut the pultruded profiles to the required
23.6 MATERIALS
length for shipment and use. This saw is fre-
quently of a radial arm type as shown in Fig. Fiber properties to aid the designer are shown
23.11, but can be also a chop saw, orbital or elsewhere in this text. Following the selection of
band saw. It is mounted on a platform which the fiber type to suit the required design factors,
moves down the pultrusion exit table at the the fiber must be oriented in the correct direc-
same speed as the pultruded product. Carbide tion. It is understood that all of the fiber types
and diamond tipped saw blades are used for must be available in continuous form in order
the cutting of glass and carbon pultrusions. to be useable in the pultrusion process. The
However, these saw blades are not effective for most commonly used form of continuous rein-
cutting aramid fiber pultrusions. This fiber is forcement is roving. This is available in single
known for its properties of toughness and and multiple strand configurations30• Glass rov-
resistance to abuse. These properties lead to ings are designated by their yield which is the
difficulty in machining, grinding and cutting. number of yards per pound of material. The
The most successful method of cutting two most commonly used versions are at
112 yd/lb or 224 in/kg or 112 or 124 m/kg (56
or 62 yd/lb). The glass rovings are typically
supplied in 18.1 kg (40 lb) hollow cylindrical
packages with a center pull payout. A similar
center pull spool is also available for both
aramid and polyethylene fibers. Carbon fiber is
typically available in either a 3K, a 6K or a 12K
filament. It should be noted that the tow sizes of
the carbon are much smaller than the glass rov-
ing and package weights are 1-2 kg (2-5 lb)
with an outside payoff designed for a package
holder style creel system. New versions of the
carbon fiber roving are available now in 40K,
160K and 320K tows. Use of these tows allows
the fiber to be laid down very rapidly and con-
sequently these versions of carbon fiber are
attractive to pultruders. Typical properties of
Fig. 23.11 Conventional cut-off saw. (Courtesy of fibers used in pultrusion are shown in Table
Creative Pultrusions Inc.) 23.131,32.
 Materials 503

Table 23.1 Typical properties of the major fibers used in the pultrusion process

Property E-glass S-glass KevIar Spectra Carbon Carbon

(Aramid) (Type T300) Inter. modulus
Density, 2600 2491 1470 968 1720 1770
kg/m3
Tensile strength, 3447.5 4585.2 2964.9 1170.0 1896.1 2560.9
MPa
Tensile modulus, 72.4 86.9 131.0 26.0 379.2 473.6
GPa
Elongation at 4.8 5.4 2.3 3.7 0.5 1.81
break, %

All the rovings discussed will yield the highest is porous and its use on the exterior surface of
possible longitudinal properties. Fibers as rov- a pultrusion might well leave porosities or
ings result in the maximum fiber content to be voids in the surface. To counter this problem, a
achieved in pultrusion. If the longitudinal rov- very fine filament, E-glass mat, commonly
ings are used under near perfect conditions, a known as veil, can be used as the surface ply.
65% fiber volume percent level should be Its presence during the pultrusion process will
achieved. In a product which utilizes 100% tend to bring more resin to the surface of the
roving this material is normally in the longitu- pultrusion and this will achieve a smooth, uni-
dinal direction or axis of the pultruded form surface, devoid of porosity or voids. The
product. Properties in the other two directions veil mats can also be placed internally in the
are dependent upon the resin system and the composite and recent improvements in their
mechanical properties of the matrix resins are structural properties have made this possible.
much lower than fiber properties. Transverse The random fiber mats in E-glass are used
strength problems are overcome by inserting in weights of 0.15-0.6 kg/m2 (0.5-2 oz/ft2).
transverse fiber materials into the pultrusion. The inclusion of these mats in the pultruded
This is done either by using fabric or continu- structure means that some of the longitudinal
ous strand mat. The latter is most commonly fibers will have to be removed to allow for the
used. While the fabric is a standard woven tex- volume of the mat, veil or fabric. With the use
tile form, the continuous strand mat has fibers of fabric or mats in the structure, the resin con-
oriented in a random mode, bonded with a tent by volume will increase in order to fill the
thermoset resin binder which holds the mat openings in the mat or fabric. Thus while the
together adequately for processing in the pul- transverse strengths increase, the longitudinal
truded section. While mat is available in any of strengths usually decrease. Mats are also avail-
the fibers which have been previously dis- able in carbon fiber.
cussed, the most common available mat is an The random fiber structure of strand mat
E-glass version that has coarse fibers in an provides fibers in all directions. However, this
open or porous construction. This mat can be random orientation does provide some prob-
used either as a center ply in a pultruded lems in that the fibers may not provide a
structure or on the outer surface of the struc- symmetrical balance within the structure. The
ture. Use of the mat greatly improves the initial solution to this problem was the use of
transverse physical and mechanical proper- woven fabric. However, the lack of tension in
ties. It should be noted, however, that the mat the fabric results in a lower strength capability
504 Pultrusion

of the pultrusion since under load, the fibers 23.6.1 MATRIX RESINS
in the fabric will have to straighten and
Of all of the technology considerations in the
become tensioned prior to being able to
pultrusion process, the most critical material is
accept load. One way of solving this diffi- the resin system and its formulation36-38. Resin
culty is to use non-woven biaxial fabrics
selection controls mechanical characteristics,
which are stitched or knitted together at the
electrical insulation, corrosion resistance,
crossover points. Because of the nature of fab-
operating temperature and flame and smoke
rication of these nonwoven materials, any
properties. It also has a significant effect on the
ratio of fibers in the two directions can be
process speed because of the required cure
provided. It is also feasible to utilize ±45°
cycle for any particular resin. The selection of
fibers in conjunction with the 0 and 90° fibers.
a resin system will also affect the production
The biaxial fabrics are normally used as inter-
cost of the process. The two most commonly
nal plies and not on the external surface. This
used resin types in pultrusion are the isoph-
is due to the nature of the nonwovens in that
thalic polyester and the vinylester. These two
their transverse fibers will tend to be dis-
comprise over 90% of all resins used in pultru-
placed by friction with the walls of the die
sion. Epoxy resins and phenolic resins are also
during pulling. Fabrics using a ±45° orienta-
being increasingly used. Phenolic resins were
tion without any longitudinal fibers are
traditionally avoided by pultruders because of
usually impractical for the pultrusion
their condensation reaction during cure.
process. It should be noted that hybrid com-
Condensation reactions produce large vol-
posites with tailored properties are possible
umes of water vapor and this typically causes
using combinations of the fiberglass, carbon
voids, channels, delaminations and porosity
and aramid materials. The designer will read-
when there is no provision to remove it. While
ily determine the mechanical properties
the pultrusion die does have an entrance and
which are required from the nonwoven or
an exit, nonetheless the system is essentially a
woven fabrics. The rule of mixtures will
closed, pressurized volume. Table 23.2 pro-
apply for combination properties33. vides typical mechanical properties for resin
When pultruded composites are used in
systems most generally used in pultrusion and
outdoor weather conditions, the surface of
several other chapters discuss the broad range
the composite may be degraded with time by
of matrix materials39 •
sun, wind, rain and ultraviolet exposure. In
order to solve this problem, additional resin
needs to be provided at the exterior surface of
23.6.2 PHENOLICS
the pultrusion. This is done by incorporating
the very fine filament veil mats which are In consideration of the use of phenolic resin
typically fabricated from polyester or nylon34 • systems, the disadvantage of the condensation
These veil fabrics are available in a variety of type reaction was certainly sufficient to cause
weights and weaving patterns. They help the delays in the potential use of phenolic resin in
pultruder by prOViding a tough surface mate- the pultrusion process. The amount of water
rial which will protect the die wall from the vapor which is generated in the condensation
abrasive nature of the fiberglass or aramid. In process is very large and it has always been
addition, the resin rich surface is created assumed that a phenolic pultruded structure
without any obvious fiber weaving patterns, would look somewhat like a sponge. However,
plus the veil materials can be screen printed phenolic systems have been pultruded in
with company identification or decorative recent years and a phenomenon has occurred
effects35• which is not well understood. During pulling
of a phenolic/fiberglass structure through the
 Materials 505

Table 23.2 Typical mechanical properties for resins Federal Department of Transport and the var-
used in the pultrusion process ious state departments of transportation have
imposed fire controls on composite materials. It
Property Polyester Vinylester Epoxy is required that they will not burn or stimulate
Density, 1100 1100 1300 combustion, have minimum required smoke
kg/m3 levels and also will not produce toxic fumes
Tensile strength, 77.2 81.2 75.4
under flame impingement and high environ-
MPa mental temperature conditions. Considering all
of the resins in the composite industry, pheno-
Tensile modulus, 3.3 3.4 3.3 lic resins will come closest to matching these
GPa
flammability specifications. Because of these
Elongation at 4.2 4.5 6.3 factors, it is anticipated that the use of phenolic
break, % resin systems in pultruded products in the
Flexural strength, 122 134 115 future will increase dramatically.
MPa
Flexural modulus, 3.2 3.1 3.3 23.6.3 COMPARISON OF RESIN SYSTEMS
GPa
In recent years, there has been use of epoxy
Heat distortion, 77 99 166 and phenolic resins in pultrusion. G.A. Hunter
°C
of Shell Development Company compared the
properties of resin systems42 • He provided a
three zone model of the pultrusion process
pultrusion die, a high pressure jet of steam is within the curing die (Fig. 23.12). The sketch
noted at the exit from the die. How and why provides an excellent background for compar-
the steam is caused to come off in this manner ison of resin properties43 •
is not known. However, in the experiments Of the four primary resin systems used in
which have been run, the resulting pultrusion the pultrusion process, the polyester and
has not had any porosity problems and the vinylester resins account for more than 90% of
processing tests are noted as being successful. the marketplace. Phenolic and epoxy resins
Specific pultrusion grades of phenolic resin make up the balance of the market. In com-
systems are now available from plywood paring resin systems, one should review the
manufacturing corporations such as internal contours and the heating profile of the
Weyerhauser and Georgia Pacific. The avail- heated resin die and examine the change in
ability of these resins would certainly indicate resin morphology as it proceeds through the
that the anticipated processing problems have die. The model of the pultrusion process given
not occurred40, 41 •. The importance of phenolic in Fig. 23.12 which shows the three zones of
resins is in their resistance to fire and their low the heated die and the transition of the resin
smoke and toxicity production under fire con- phase from liquid through the gel zone into
ditions. All forms of composite materials, the solid phase. The first zone shown in Fig.
including pultrusions, are being used increas- 23.12 is where the material enters the die at
ingly in mass transit, aircraft and civil room temperature and expands as it absorbs
engineering applications. In all of these areas heat which causes the hydraulic pressure in
of application, increasing contact of the com- this zone of the die to rise. As the material pro-
posite material systems with the general gresses into zone 2, or the gel zone, it has
public is occurring. Because of this, the fire absorbed more heat, is beginning to cross link
smoke and toxicity requirements of specifica- and changes from a viscous liquid into a non-
tion control groups such as the FAA, the flowing jelly type of material, then into a
506 Pultrusion

ftTHE GEL ZONE"

STRIP HEATER
/'
'\ DIE /'
(\--;~---L-I-Q-U-ID-P-H-A-SE"':d::=""r.:=::-:~-=-_---_~-'~-S-O-L-1-D-P-H-A-S-E--_-+~--I(:

\" - ;-D-IE""""'C========±::±:======,-_ _ _ ......;,..._---i I

ZONE 1 ZONE 3
1
I
lvlSCOUS SHEAR SLIDING FRICTION FORCES

FORCES r-···;~·:~·~·!~~E;:·~·~:~·······l

Fig. 23.12 Three zone model of the pultrusion process. (Courtesy of Shell Development Company.)

rubber-like texture. As the material cures to a gel zone will expand or contract.
hard solid, shrinkage occurs which releases Joseph Sumerak in 1985 quantitatively
the hydraulic pressure forces and the product described the internal dynamics of the pultru-
shape retracts from the internal surface of the sion process. Taking test results from
die. This is zone 3. In this zone, because of the Sumerak's earlier work, Hunter showed the
release of the product from the surface of the relationship of pull loads to processing speed
die, the sliding frictional forces are very slight. for catalyzed and uncatalyzed resin systems
Depending upon the thickness of the part and (Fig. 23.13)44-47. For the uncatalyzed resin case,
the process speed, the bullet-nose shape of the the rising pulling load associated with

REICHOLD 31 - 020
1000
WITH 20 PHR CLAY
900
[] - PULL LOADS ATTRIBUTED TO
VISCOUS SHEAR AND FRICTION OF
800
CATALYZED RESIN
o - PULL LOADS ATTRIBUTeD TO
700 PURE VISCOUS SHEAR OF THE

....
ID
A -
UNCATALYZED RESIN
CALCULATED PULL LOADS
I 600
ATTRIBUTED TO VISCOUS
U)
SHI!AR OF THE CATALYZED
C 500
oC RESIN

0
...I 400
...I
...I
::::» 300
a.
200

100

0
0 12 24 36
LINE SPEED, IN/MIN.

Fig. 23.13 Pull loads compared with line speed for different types of resin systems. (Courtesy of Shell
Development Company.)
 Materials 507

increased processing rate, or line speed, is the die and is proportional to the speed of pro-
result of increasing shear forces over the cessing. Hunter provided evidence that the
length of the die. In the case of the catalyzed pressure loss in zone 3 of the die occurs well
resin and referring back to Fig. 23.12, viscous before the material cools. Thus it is not thermal
shear forces are generated only in the front contraction but volumetric shrinkage due to
portion of the die, i.e. zone 1. Within the gel the cure of the resin. The coefficient of thermal
zone, cohesive forces come into play for a expansion of the steel material of the die also
small length of the die which is followed by enters into this equation. For a differential
the transition to the rubbery cured material temperature of 121°C (250°F), the hottest tem-
which provides substantial friction forces. As perature section of the die for a 12.7 mm
the resin hardens and shrinks away from the (0.5 in) diameter pultrusion will be 0.3% larger
surface of the die, the frictional forces are than the entrance. Thus pressure and volumet-
reduced significantly. It is obvious that the ric shrinkage together play a major role in
pull load is significantly higher for the cat- pultrusion dynamics. Insufficient pressure
alyzed resin system, particularly as the line causes sloughing problems and insufficient
speed increases. This proves that the major shrinkage can cause excessive pull loads. The
portion of the pultrusion loads are generated resin rate of shrinkage affects the rate of pres-
in the gel zone and are cohesive forces and sure decay and is controlled linearly by the
frictional forces resulting from the interface of cure rate of the resin. Thus a delicate balance
the resin and the die. between pressure, cure rate and shrinkage
Sumerak showed that a significant part of must be obtained for a clean pultrusion
the internal pressure does develop inside the process to take place.

22.8' C QUENCH
2

~ , - ~---------
,""
:::E
-
~ I 70' C HEAT SOAK
~ /
~ Oh,~--------------------~.~~----------------~
w
:::E
:::I

o
..J
-1 INITIAL CURE CYCLE

- - - SECOND HEAT CYCLE

·
>
STARTING VOLUME - 9.~98 ML
-2
NET VOLUME LOSS - 0.55 ML
VOLUMETRIC SHRINKAGE - 6.04 %
·
-3

o 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90
CURE TIME, MIN

Fig. 23.14 Volume change of polyester resin during cure. (Courtesy of Shell Development Company.)
508 Pultrusion

23.6.4 SHRINKAGE is fully cured. Similarly, the pressure during

Hunter ran shrinkage tests on a typical poly- the gelation phase ensures that the product is
ester resin and a standard Shell epoxy resin tightly held against the surface of the die and
system42 • The volume change of the polyester consequently a smooth surface will be gener-
and the epoxy resin are during cure are shown ated with the pressure preventing sloughing.
in Figs. 23.14 and 23.15. The data shows that Thus, from a comparison of the test results, it
the polyester shrinks almost twice the degree is obvious that the polyester shrinkage profile
of epoxy. However, Hunter reports that the net is superior to the epoxy in terms of providing
shrinkage is not nearly as important as the gelation and cure under pressure. In addition,
profile of that shrinkage. The polyester contin- the sudden high initial rate of shrinkage fol-
ues to expand after the gel point which is lowing gelation for the polyester resin is also
followed by a high shrink rate that gradually beneficial in that it results in a fast pressure
tapers off. In comparison, the epoxy resin drop and hence frictional force reduction.
shrinks before it gels and continues to shrink In comparison, the epoxy resin begins to
at a steady rate until it is fully cured. This shrink well in advance of gelation and gels
information sheds new light on the under- under a condition of declining hydraulic pres-
standing of the pultrusion characteristics of sure. Thus much of the hydraulic pressure is
epoxy resins compared to polyesters. In most lost by the time that gelation occurs due to the
composite manufacturing processes, the appli- effect of volumetric shrinkage. Following gela-
cation of pressure during the curing phase of tion, the rate of shrinkage is very slow such
the resin is always beneficial to the resulting that it causes the friction forces to only reduce
product. Pultrusion is no exception to this rule gradually. Thus in the gel zone for epoxy
and it is beneficial to generate pressure in the resins there is insufficient hydraulic pressure
pultrusion die up to the point where the resin to prevent sloughing. This explains why a

3
GEL POINT
I 22.8° C QUENCH
2
~----------------~
....I
:i
~~ :
/
aU 1 I
I
"z:
oC o~--
I
I 70° C HEAT SOAK
__________________________ -+~~ ___________
o
w
:i
::J -1 - - INITIAL CURE CYCLE
....I.
o - - - SECOND HEAT CYCLE
>
STARTING VOLUME - 8.954 MLS.
-2
NET VOLUME LOSS - .32 MLS.
VOLUMETRIC SHRINKAGE - 3.57 %
-3

o 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90
CURE TIME, MIN

Fig. 23.15 Volume change of epoxy resin during cure. (Courtesy of Shell Development Company.)
 Materials 509

problem is frequently encountered when 23.6.5 CURE RATE

epoxy resin is substituted directly for a poly-
ester without consideration of the curing and As previously mentioned, shrinkage rate is a
shrinkage properties. Hunter postulates that a direct result of the known cure rate of the resin.
simple solution has been found to compensate It is beneficial in the pultrusion process to have
for these shrinkage characteristics in epoxy a high shrinkage rate to initiate a quick pressure
resin systems. The presence of fillers, whether drop to reduce frictional pull loads. From this
fiber or powdered, in the resin system, reduces point of view, it is important for the epoxy to
the amount of total volumetric shrinkage. Also have a fast cure rate. This will also provide a
the pressure from thermal expansion is shorter gel zone which will result in a faster
directly proportional to the amount of filler or processing rate. Cure rates of polyesters may
fiber reinforcement volume. Thus the increase be varied chemically by changing the amounts
in reinforcement to resin volume ratio, either and types of peroxide catalysts which are used
by fiber or powder fillers, will reduce the to initiate them. It is not simple to change the
shrinkage tendency and the hydraulic pres- cure rate of an epoxy resin chemically.
sure will be increased at the same time. Even Curing agents for epoxy resins are selected
though the epoxy resin shrinks prior to gela- based on the desired performance parameters
tion, sufficient pressure will remain to prevent for the epoxy in the final product or structure.
sloughing. This explains why it is always ben- Some considerations of pot life and manufac-
eficial to have a higher fiber to resin ratio for turing process also influence this selection.
epoxies in the pultrusion process than for Thermal accelerators can be used. However,
polyester resins. the effect of increasing cure rate accelerator

250r----------------------------------------
o EPOXY WI aTD. ACCELERATOR LEVEL

A EPOXY WI DBL. ACCELERATOR LEYEL

200 o
o
POLYESTER WI HIGH TEIIP. CATALY8T

POLYE8TER WI LOW TEIIP. CATALY8T

III
CI
Z
o 150
U
1&1
III

0'--_ _---IL--_ _---I_____-L_ _ _ _....1.._ _ _-J

75 100 125 150 175 200

GEL TIME TEMPERATURE, DEG. C

Fig. 23.16 Gel time compared with temperature of epoxy and polyester resin. (Courtesy of Shell
Development Company.)
510 Pultrusion

content may be to reduce the pot life. Hunter 23.6.6 REINFORCEMENT VOLUME
ran experiments to provide data on gel times The general relationship of glass fiber content
of epoxy resin at different accelerator levels.
to pull loads in epoxy resin system is shown in
In addition, he checked on the gel times of
Fig. 23.18. These data were derived from an
polyester resin using two different cur~g experiment in the Shell laboratories where the
agents. Figure 23.16 shows the data whIch
reinforcement volume was decreased and the
resulted from these two evaluations. The data
pull loads were recorded until sloughing
on the epoxy resin system shows that signifi-
occurred42 • The fiber volume was then
cantly more heat is required to generate the
increased until the sloughing was eliminated
same gel times even though the accelerator and was increased further until pull loads
content is doubled. Figure 23.17 shows the
became too large. The data shows that there is a
viscosity versus time of the epoxy resin at two
plateau in the pull load curve spanning approx-
accelerator levels and two temperatures. This imately 2% of the glass fiber content range. This
graph clearly indicates that pot life is sacri-
is the optimum level for pultruding the
ficed by increasing the accelerator level. Pot
12.7 mm (0.5 in) diameter epoxy rod used in the
life is also affected by temperature and Fig. test program. Below the optimum range,
23.17 illustrates that a small increase in tem-
sloughing occurs owing to the insufficient
perature will reduce the time to double the
hydraulic pressure at the gel point. Above the
initial viscosity by almost half. Heat can be
optimum level, the pull loads rise owing to the
generated during the mixing process .of. ~e high pressure both during and after the gel
epoxy formulations and because of this It IS
zone (referring back to Fig. 23.12). Both poly-
important to minimize the mixing times when
ester and epoxy systems respond similarly to
using high shear mixers that generate heat the different types of reinforcement materials
within the body of the resin system. The
which are contained within the pultruded
implication of the experimental data pre-
structure. For both of these resin systems, the
sented in these figures is that the most
minimum reinforcement level to prevent
efficient method of increasing cure rate in the sloughing when using a continuous mat and
epoxy resin is to increase the die temperature. roving is somewhat less than that for an all rov-
4000 0- STD. ACCELERATOR LEYEL AT I.·C

A- .TD. ACCI!LI.RATOR LEVEL AT 1:1.:1"0

3600
w [J- DBL. ACCELERATOR LEVEL AT a.·c

o 3000 0-
II)
DBL. AceELERATOR LEVEL AT sa.l"e

a.
t- 2500
z
w
U 2000
,;
t- 1500
II)
0 1000
U
II)

> 500

0
o 2 3
TIME, H

Fig. 23.17 Epoxy resin viscosity compared with time and temperature. (Courtesy of Shell Development
Company.)
 Materials 511

1000~---------------------------------'~----i

1/1" DIAIIETER ROUND

PROCESSED AT 12"/IIIN.
112 YLD. BINGLE END ROVING

800
H-" ENDI
11 % CLAY

...GI
Q
0(
800
...
0 TOO LOW. OPTIIIUIi

...... SLOUGHING

OCCURS
RANGE

:)
0.;
400

200 L-__ ~ __ ~ ____ ~~~ ____ ~ __ ~ __ ~ ____ ~ __ ~

78 79 80 81 82 83 84 87
GLASS CONTENT (NON-COMBUSTIBLES), % BY WT.·
~Of WHICH APPROXIIiATELY Z % IS RESIDUAL CLAY

Fig. 23.18 Epoxy resin puUloads compared with glass content. (Courtesy of Shell Development Company.)

ing part. Table 23.3 lists the target reinforcement 23.6.7 DIE TEMPERATURE CONTROL
volumes for the epoxy system for a variety of In polyester pultrusion, there is a wealth of
reinforcement systems. These values were gen- prior experience7 which can be used to pro-
erated by following the same procedure as for vide a temperature set point to produce the
the data in Fig. 23.18. The data shown in Table desired surface and internal quality of the part
23.3 are qualitative rather than quantitative val- with the controller being a thermocouple
ues. They may be used to estimate the required located some short distance from the entrance
reinforcement volume. to the die. This creates a situation which is
Table 23.3 Target fiber volume ranges for epoxy independent of the actual exotherm tempera-
pultrusion ture in the curing process.
For epoxy pultrusion it is vital that the peak
wt.% exotherm be understood and controlled. It
should not exceed 225°C (437°F) in the hottest
All glass roving reinforced composites
region of the part and die. At this temperature
Multi-end type rovings 78 homopolymerization will take place within
Single end type rovings 77-81 the epoxy resin system and the resin does not
Glass roving and continuous mat need the curing agent to stimulate the cure.
reinforced composites The mechanical and physical properties of the
3.175 mm (1/8 in) thick cross sections 64-67 structure are degraded under these conditions
6.35 mm (1/4 in) thick cross sections 71-74 by the presence of the unused curing agent.
For most thin profiles (up to 12.7 mm (0.5 in)
Carbon fiber reinforced composites
thick), a single heating zone is sufficient. The
All unidirectional tows 67-74 thermocouple should be located in the center
(57-65 vol%)
512 Pultrusion

CIT
200C
2 :9 9 9 9 2 1/2' D ROUND
12' IMINUTE

_,
750 LB. LOAD
18' STRIP HEATERS
250
...

""
............
.... _------
- TEIIPERATURE ON THE SURFACE OF THE ROD
- - - - TEMPERATURE IN THE CENTER OF THE ROD

O~----~--~--~~~h---~--~~~~--~-
o 6 30 36 42 48

Fig. 23.19 Die temperature profile of single zone heating. (Courtesy of Shell Development Company.)
of the strip heater to minimize the overshoot in the epoxy resin system. If the temperature
and lag time for the temperature controller. set point is too low, the resin cure rate will
For a 12.7 mm (0.5 in) thick cross section the decrease which increases the size of the gel
temperature set point of 200°C (392°F) on the zone. At the same time, the rate of shrinkage
surface of the die will yield an internal peak and the rate of hydraulic pressure decay is
exotherm temperature of 225°C (437°F). Figure reduced and this results in more pressure
23.19 is a graphic example of this single strip within the larger gel zone which increases pull
heater profile. Figure 23.20 depicts the graph loads. As the die temperature increases, the
of the temperature set point versus pull loads conditions begin to favor reduced pull loads.

700

ID
...I

Q
- 600
C
0
...I
...I
...I
~
II. 500

140 150 160 170 180 190 200

DIE CONTROL TEMPERATURE, ·C

Fig. 23.20 Pull load compared with die control temperature. (Courtesy of Shell Development Company.)
 Materials 513

With section thicknesses up to 12.7 mm (0.5 23.6.8 RESIN MIXING

in) a two zone heating profile can be used and
The best approach to mixing resin system com-
this will probably eliminate the need for radio
ponents is to precisely follow the stated recipe
frequency (RF preheating) to prevent internal
for the amount of material to be added and the
cracking. The pull loads are much lower and
degree of mixing following the addition. As ~e
the temperature decay rate is less than the sin-
ingredients such as catalyst accelerator, pIg-
gle zone heating profile. Thus the two zone
ments, viscosity and extenders, internal mold
version will result in a higher degree of cure
release and fillers are added, the resin is mixed
than the single zone.
for a very short time of up to one minute (Tab~e
For thick pultrusions with sections beyond
23.4). The curing agent is left out of the mIX
12.7 mm (0.5 in), the RF preheating method
until the mix is ready to be added to the pultru-
results in faster processing rates without crack-
sion resin bath. The addition of filler materials
ing. Figure 23.21 illustrates the die te~per~ture requires a high shear mixing and this should be
conditions for the same 12.7 mm (0.5 m) dIam-
minimized since it generates significant quanti-
eter rod pultruded with RF preheating.
ties of heat. Just prior to start up, the curing
Preheating serves to reduce the temperature
agent should be added and blended at a
differential between the die entrance and the
reduced mixer speed (high shear mixing is only
gel zone which results in less volumetric expan-
needed for filler addition.) Following start up of
sion due to temperature and hence less
the system, make-up replenishment resin
pressure. The temperature lag between the sur-
batches should be scaled to the depletion rate
face and the center of the part is also reduced by
for the run. If the resin consumption is 7.51 (two
RF heating, thus the gel zone is smaller which
gallons) per hour, add 7.5 1 (two gall~ns) per
reduces pulling loads. Increase of the process-
hour. This addition will assist the pot life of the
ing rate through the machine will bring the
epoxy resin and will dilute it with fresh resin on
pressure, the gel zone size and the pull loads
a frequent basis. If a small bath size can be used,
back to normal conditions. This is how RF heat-
this will increase the dilution effect of the
ing permits faster processing rates without
replenishment material. Depending on the size
increasing the pull loads beyond the standard
of the bath, this technique can enhance the pot
levels.

CIT
200C
9 :2 9 9 9 9 1/2' D ROUND

I DIE 18' IMINUTE

722 LB. LOAD
18" STRIP HEATERS

, ... - ...
250
, ....
...... ......
....
~200
w
II:
~15()t
......
....... ......
0( --~----.
II:
w100
IL
:II! - - TEMPERATURE ON THE SURFACE OF THE ROD
w
..... _ _ _ _ TEIIPERATURE IN THE CENTER OF THE FlOD

0
0 6 12 18 24 3D 36 42 48
DIE LENGTH, IN

Fig. 23.21 Die temperature profile with RF heating. (Courtesy of Shell Development Company.)
514 Pultrusion

life of the resin by 800% or more. A large master initiated there may be potential for significant
batch of the resin can be mixed and set aside in build up on the surface of the die and this may
advance of the production run. This batch with- . be difficult to remove. The most troublesome
out the curing agent will be stable for up to spots in the die are the low pressure, remote
three days. The curing agent can then be added areas such as a corner or a small radius. There
to a small make-up batch and mixed in just is a standard process used in pultrusion of
prior to the addition to the system resin bath. A purging the die with pure mold release just
typical formulation used in batch mixing is prior to entry of the resin. Experience and
shown in Table 23.4. recent tests have shown that the pure mold
Even with the best of conditions in terms of release is not necessary and may lead to
a small resin bath and frequent addition of related problems. The normal types and quan-
new batches, the resin mix will ultimately tities of mold release recommended for use
become too viscous for good fiber wetting. A with epoxy were determined by tests at resin
good tip is to provide a large hole in the resin suppliers. Any levels of mold release in excess
bath with an appropriate plug. This allows a of those recommended will not provide addi-
quick drain and a clean and refill with only a tional benefit to the process. A high
momentary pause in the process. With this concentration of mold release may result in a
step, the pot life of the resin bath will be reset 'squeeze off' at the die entrance which could
to zero. It should also be noted that the plug in work its way back into the resin bath. If a pre-
the drain hole should not be threaded since a lubrication step is used, this squeeze off resin
very small amount of resin can cure and lock quantity must not be allowed to get back into
up the threads. A preferred plug would be the resin bath.
hard rubber with a rim, similar to the knock Prior to the resin entering the bath, the die
out drain plugs in the floor of an automobile. temperature must be stabilized at the set point.
For most parts an initial throughput speed of
25.4-30.4 em (10-12 in) per min is recom-
23.7 START-UP PROCEDURE
mended until the cured stock is in the pullers,
A key factor in trouble free start up is to use to minimize loads imposed on the dry fiber.
the minimum required amount of reinforce- Process rate increase can then be made gradu-
ment. However, if too much reinforcement is ally and for epoxy resins the processing speeds
eliminated in start up, sloughing will take normally will not exceed 45.7 em (18 in) per
place in the die because of insufficient total min. It would appear that 30.4-35.5 cm
pressure in the gel zone. Once the sloughing is (12-14 in) per min provides the best combina-

Table 23.4 Two part batching for long production runs

Part 'A' Part 'B'

(in order of addition) [PHR, (wt.%)] Epon Curing Agenf®
1. Epon® resin 9310 100.00 9360
2. Epon Curing Agent® 9360 0.67 33 PHR based on resin weight
3. Mold Accelerator 837 WlZ Int. 1846 0.70 OR
4. Zylac 907 0.40 Recompute based on the total weight of Part A:
I

Blend the above for 30 s then add clay (33/121.77) 100 = 27.1
5. ASP-400P 20.00 and use
Blend clay for no more than 5-10 min 27.1 PHR to Part'A:
Courtesy: Shell Chemical Co.
 Additives 515

tion of processing parameters, pull loads and provide a mild scrubbing action that will clean
surface gloss. If RF preheating is used, perform the surface of the die.
the start up without it at a reduced speed
15.2-20.3 cm (6--8 in) per min to minimize pull
23.7.2 SHUT DOWN
loads. When cured stock is through the die and
in the pullers, the RF preheater can gradually Standard shut down procedures have been
be started up. The resin temperature entering developed for polyester resins. The key step is
the die should be monitored and as it reaches to remove the resin bath or bypass the rein-
about 71 °C (160°F) gradually increase the pro- forcement around it. The dry reinforcement
cessing speed. Do not recycle the resin which is should be completely pulled through the die.
squeezed off at the die entrance. None of it should be cut out or removed. At
this point the die will be free from build up
and ready for a restart.
23.7.1 TROUBLESHOOTING
The resin drain from the bath should be
The standard problem encountered with placed in a container in an area with good ven-
epoxy resin pultrusions is poor surface finish tilation and spill protection. A metal or plastic
or sloughing. The reasons for this have been tray with a surrounding high lip will be satis-
discussed earlier in this chapter. If the rein- factory for spill protection. The resin containers
forcement level is low then the cure for this is should only be half full. Eventually when it
to obviously add some reinforcement. does exotherm, it will become hot, expand and
However, if the reinforcement level is in accor- may overflow the container if it is too full.
dance with the specifications, then additional
reinforcement will increase the pull loads
23.8 ADDITIVES
beyond standard. In this case the die tempera-
ture profile is probably too low. There is an Inorganic fillers are used to reduce shrinkage at
instrument on the market known as the polymerization. They also extend the volume of
Gelstar Thermal Analyzer. The thermal ana- the resin phase to provide a low cost formula-
lyzer can be used to obtain a temperature tion. These are primary additives. Fillers can
profile within the die. From this, the size of the frequently constitute the largest proportion of a
gel zone can be estimated. This is proportional formulation, second only to the base resin.
to the lag time between the die and internal Fillers are classified according to their particle
temperature profiles. If the temperatures size, as either coarse fillers or fine. Coarse fillers
appear to be within limits but the lag time is have an average particle size in excess of 8 11m
too large then the processing rate is too fast for and are generally the nonfibrous type with low
the particular cross section within the die. At surface area and low oil absorptions. They can
this point either reduce the process rate or use be highly loaded into the resin and are easily
RF preheating to minimize the problem. These wetted out by the resin system. Their disadvan-
steps will reduce the size of the gel zone. tage is that they tend to provide poor
If sloughing is encountered, the part should compound cohesiveness and to introduce local-
automatically clean itself up. The typical ized resin rich pockets and possibly to increase
purge techniques that are common in the fiber agglomeration during secondary molding
industry can be used with epoxy resins. If a processes (as in pull forming). The large particle
portion of the die refuses to clean up, a trick is size filler can be filtered out by a high density
to insert a copper kitchen 'Chore Boy' in the roving preform which can lead to large voids in
area of the part which has the problem. This the interior of the pultruded structure.
will often push the offending plug out or catch The most common of these fillers are cal-
onto it and pull it out. The copper gauze will cium carbonate, aluminum silicate and
516 Pultrusion

alumina trihydrate. Silica and coarse tales are 23.8.1 PIGMENTS

also examples of coarse fillers. Calcium car-
Pultruded products are normally associated
bonate is primarily used as a volume extender
with bright colors and these colors are nor-
to provide the lowest cost resin formulation
mally created by adding suitable colored
where performance is not critical48•
pigments to the resin system which are then
Fine fillers have an average particle size of
cured into the matrix material. Pigments are
5 Jlffi or less and have high surface areas which generally of three types: (a) dyes; (b) organic
can produce high viscosities in the formula-
pigments; (c) inorganic pigments..These ~~
tions. These fillers provide a high order of
pigments are characterized by theIr own mdl-
cohesiveness and will tend to lubricate the
vidual properties49 • Dyes have good
pultrusion system. They also help to reduce
transparency and acceptable brightness.
localized shrinkage owing to their more com-
However, they have poor heat resistance and
plete distribution within the polymer. Kaolin
tend to migrate in solution. Organic pigments
clays, hydrous alumina silicate, fine tale, col-
also have acceptable brightness and brilliance
loidal silica and precipitated calcium
but are not normally as good as the dyes.
carbonate are examples of fine fillers. Clay
Weather resistance and UV absorption tends
fillers are used to improve corrosion resistance
to be a problem and the colors may deteriorate
and where electrical properties are required.
and fade after long periods of UV exposure.
They also impart a superior surface finish to
Inorganic pigments are generally the materi-
the pultruded product. Alumina trihydrate
als of choice. These are usually natural or
improves flame and smoke generation proper-
synthetic metallic oxides, sulfides, or other
ties and occurs in applications where
salts which are heat treated and converted to a
consumer or governmental codes are imposed
dry powder at 600-1100°C (1l12-2012°F).
to decrease flammability. Inorganic pigments have superior properties of
Fillers used in the pultrusion process
brightness to those of organic pigments or ~yes
should contain less than 0-5% free water con-
and are very resistant to weather and mIgra-
tent and should be uniform and free from tion and have a very high stability under light
contamination. Foreign material in the filler
exposure. The major problem, h~,:ever~ in
may cause localized reaction with off gassing
incorporating these pigment addlti~es mto
of volatile byproducts and voids or could
resin formulations are the effects which they
affect the uniformity of coloring within the have on the cure cycle of the resins. It is possi-
pultrusion. Fillers are mixed into ~e resins in ble for them to be involved in the
quantities up to 50% of the total resm formula-
polymerization reaction during the curing of
tion by weight (100 parts filler per 100 parts the resins and to become chemically attached
resin). Limits of filler addition are based on the
into some of the reactive sites. This has a strong
viscosity of the system which results fro~ ~e effect on the properties of the resulting com-
particle size of the filler and the characte~stics posite material and will require a change in the
of the resin. Wetting agents are sometimes
temperature/time curing cycle. A large per-
used to add a volume of filler material without centage of pultruded composite components
increasing formula viscosity. Wetting agents are used outdoors and UV exposure is a prob-
can be added to the filler by the supplier or as lem. Under these circumstances, titanium
an additive during the formulation process for dioxide, an excellent UV absorber and
the resin. Air release agents are added in the whitener, is used as a part of the formulation.
same manner and will result in more efficient
Its presence would also naturally create a paler
packing by reducing entrapped air in the li~­ color in the pultruded composite and hence
uid resin. They also tend to reduce VOId
additional quantities of inorganic pigment are
content in the finished product.
 Additives 517

normally required to make the color bright. • surface preparation of the composites is
Zinc sulfide is also a UV absorber which is fre- not required;
quently used in pigment systems. Both • inspection of the joint is relatively easy;
titanium dioxide and zinc sulfide seem to have • the joint can be assembled and connected
little effect on the ultimate mechanical and up to full strength rapidly;
physical properties of the pultrusions such as • ability to disassemble and reassemble.
the strength, moduli and impact resistance.
The disadvantages of a mechanically fastened
joint are:
23.8.2 STRUCTURAL CONNECTIONS

Pultruded FRP composites can be joined using

• in preparing the joints, holes must be
drilled and sealed, structural members
various methods of assembly including must be trimmed and any required gusset
mechanical fastening (with plastic or metallic plates must be fabricated;
bolts or screws or by doweling with dowels or • drilling of holes cuts the longitudinal
rivets), mechanically interlocking connections strength fibers in the area of the hole. This
(where molded or laminated inserts lock into can cause high stress levels and stress con-
the sections being fastened), adhesive bond- centrations;
ing, or a combination of these methodsso . • strength of the joint is dependent upon the
When structural components are assembled bearing strength of the composite material;
special attention must be given to the rigidity, • the strength of mechanically fastened
geometry, fabrication and assembly require- joints is also dependent upon the strength
ments involved. It has been determined that of the fastener.
mechanical fastening and/or mechanically
Fastener parameters include:
interlocking connections are most suitable for
structural connections. Some connections • clamping force/installation torque limits;
depending upon the geometry and the stresses • washer size for transmission of load;
developed in connection can be strengthened • fastener size;
by also adhesive bonding in addition to • hole size and tolerance;
mechanical fasteners. Each of these systems • joint type;
have their own advantages and disadvantages • geometry of the fastener layout;
which are discussed in the following para- • composite thickness;
graphs. • rate of loading and the direction;
• static or dynamic loading;
23.8.3 MECHANICAL FASTENERS • failure criteria;
• high stress concentrations around fas-
If properly designed and fabricated, mechani- tener.
cally fastened connections are the most
reliable method of joining pultruded struc-
23.8.4 ADHESIVE JOINTS
tural sections. In recent years a significant
amount of information through empirical test- Adhesive joints have become popular for the
ing and prototyping of connections has connection of the composite materials since
become available on the bolting and riveting there is not degradation of the composite itself
of composites. There are many standard refer- by the bonding process. Adhesives are usually
ences in this field 51- 53 • available in solid, paste or liquid form and are
classified as either inorganic or organic materi-
The advantages of a mechanically fastened als. The majority of the structural adhesives are
joint are: derived from the organic group and can be
518 PuItrusion

thermosetting resins, thermoplastics, or elas- double or single lap joint. The stress can be
tomers. These adhesives are known for their particularly severe in thick, double lapped
properties of being strong, tough, insoluble and joints;
useable over a wide temperature range54. The • adhesive bonded joints take time to cure
primary advantages of an adhesive joint are: while the resin is setting up and hardening.
During the cure cycle, the component parts
• the properties of the composite material
remain intact and are not degraded;
must be restrained in a fixed position.

• adhesives are generally stronger than the

composite material being bonded, conse- 23.8.5 MECHANICAL ADHESIVE
COMBINATION
quently any failure mode is usually forced
into the surrounding primary structural Some connections are stronger using a
material; mechanically fastened/ adhesive bonded joint.
• good distribution of joint stresses; Advantages of mechanically fastened/ adhe-
• adhesive bonding can be used to bond dis-
similar materials;
sive bonded joints are:

• adhesive joints can be particularly forgiv- • higher overall capacities;

greater resistance to environmental and
ing. Flaws in the joints do not generally •
degrade the strength of the joint. The main thermal deteriorations;
less subject to peel stress failure than
reason for this is that the critical stress •
'bonded only' joint;
location in a bonded joint is usually at the
improved fatigue and impact characteris-
end of the overlap and flaws tend to occur •
in the center of the joint where the stress tics;
increased joint rigidity.
level is low; •
• good fatigue and impact loading charac-
teristics.
The disadvantage of mechanically fas-
tened/ adhesive bonded joints is that they are
labor intensive.
Disadvantages of the adhesive bonded joints
are:
23.9 APPLICATIONS FOR PULTRUDED
• surface pretreatment such as cleaning and PRODUCTS
etching must be carefully performed;
Applications for pultruded products are many
• the preparation and mixing of the adhe-
sive in the correct proportions is critical. and varied but generally are commercially ori-
Manufacturers instructions must be fol- ented. The process shows up in the consumer
lowed precisely; and recreation market, electrical equipment
products, corrosion resistance, civil engineer-
• there is a time limit or shelf life of the
adhesive following its preparation; ing and construction and transportation
(automotive, truck, bus and rail).
• the component parts must be carefully
located using holding tools and fixtures
and maintained in position with pressure 23.9.1 CONSUMER/RECREATION
across the joint during the curing of the The combination of strength, stiffness, fatigue
adhesive; resistance and aesthetic design and coloring
• while adhesive joints are normally makes pultruded products ideal for the con-
designed and stressed in shear, there is an sumer recreation market. Applications are:
additional failure mode owing to peel fishing rods, archery bows and arrows, hockey
stress. This is a tensile stress which devel- sticks, tent poles, ski poles, playground equip-
ops a maximum value at the free end of a ment, fence posts and baseball bats.
 Applications for pultruded products 519

23.9.2 ELECTRICAL EQUIPMENT

In this marketplace, strength, electrical insula-

tion and safety are primary attributes of
pultruded products55 • The following are the
most significant applications: fuse holders, lad-
ders, tool handles, electrical conduits56, cable
trays57 and power rail covers for subway trains.
Figure 23.22 shows a set of pultruded ladders
and Fig. 23.23 shows the installation of a
power rail cover for a subway train.

Fig. 23.23 Third rail cover for rapid transit rail sys-
tem. (Courtesy of Creative Pultrusions Inc.)

23.9.3 CIVIL ENGINEERING/CONSTRUCTION

MARKET
In this market the properties which are
required are strength, modulus, corrosion
resistance and nonslip surfaces. Applications
are: gratings, stairs, guard rails, bridges and
platforms58-{iO, crash barriers, ladder cages,
structural supports61-{)3, sign posts and signs,
light poles64,65 and pedestrian bridges.
Figure 23.24 shows an installation of gratings
and hand rails in a chemical plant. Figure 23.25
shows a pedestrian bridge and Fig. 23.26 shows
the Aberfeldy foot bridge in the UK. This latter
application is one of the most interesting devel-
opments in the application of pultruded
sections. This bridge was designed and erected
by Maunsell Structural Plastics Division of
Fig. 23.22 Pultruded ladders. (Courtesy of Creative MaunselI Engineering in London. The bridge
Pultrusions Inc.) spans the River Tay in Scotland and connects
520 Pultrusion

Fig. 23.26 Aberfeldy footbridge. (Courtesy of

Maunsell Structural Plastics Ltd.)

Fig. 23.24 Grating and handrail installation. two sections of a golf course. The bridge is a
(Courtesy of North West Fibre Mechanics Ltd.) double cable stay design and all of the compo-
nents are made of composite materials with the
bridge decking and guard rails being pultruded
products. Maunsell has also installed a com-
posite vehicular bridge at Bonds Mill in
England which was opened in 1994.

23.9.4 TRANSPORTATION MARKET

This market is potentially very large. It
includes automotive, trucks, busses, light rail,
subway trains and passenger trains. The prod-
ucts which are being pultruded for this large
array of industries are as follows: drive shafts
for trucks66, leaf springs67,68, bumpers, frames
and cross members, transportation container
bodies, roll up doors, refrigerated truck com-
ponents, frames for light rail cars, interior
structure for passenger trains and subway
Fig. 23.25 Pedestrian bridge in Pennsylvania. trains69, load carrying beds for small trucks
(Courtesy of Creative Pultrusions Inc.) and frangible airport approach masts.
 References 521

Figure 23.27 shows a bus interior, fabricated

from pultruded sections and Fig. 23.28 dis-
plays a frangible airport approach mast.

23.9.5 MISCELLANEOUS
Another market which is using pultruded
products is the oil and gas industry for off
shore oil well platforms. Application here is
for floor gratings, hand rails, stairs and storage
buildings and living quarters on the platforms.
Another emerging application is in the con-
stant cross section blade for the Darius design
of windmills for alternate power.
This is just a brief review of the current appli- Fig. 23.27 Bus interior showing pultruded compos-
cations. The future of the pultrusion process and ites. (Courtesy of Creative Pultrusions Inc.)
its applications is only limited by the scope of
human imagination. The market will continu-
ously increase and it is predicted that by the
year 2000 the total volume of pultruded prod-
ucts will have tripled over the 1995 levels.

REFERENCES
1. Goldsworthy, W. Brandt, US Patent, 2 871 911
Apparatus for Producing Elongated Articles
from Fiber-reinforced Plastic Material; Issued
2/12/59.
2. Birsa, R and Taft, P, New Materials Approach
for Providing Transverse Strength in Pultruded
Shapes, RP /C Reinforced Plastics/Composites
'84; Composites go to the Market; Papers pre-
sented at Technical Sessions of the 39th Ann.
Conf., New York. Jan 16-19, 1984, Session I-A,
p. 4, 627, SPI Reinforced Plastics/Composites
Institute.
3. Taft, P. and Birsa, R, Transverse Strength for
Pultruded Parts, Plast. Engng., 1984, 40, (5),634.
4. Kirkland, c., Technology Outlook: Reinforced
Plastics Processing in the 80s, Plast. Technol.,
1981, 27(3), 53-8.
5. Iddon, K.J., Blundell, c., Pultrusion Process and
Its Application, Fabrication Techniques for
Advanced Reinforced Plastics Symposium,
Salford, April 22-23, 1980, p. 58-68, Confer.
627-88.
6. Hollaway, L., Pultrusion, In Developments in
Plastic Technology-3, (eds Whelan, A., Craft, J. L.),
Barking: Elsevier Applied Science, 1986, p. 1-46, Fig. 23.28 Frangible airport approach mast.
012. (Courtesy of Creative Pultrusions Inc.)
522 Pultrusion

7. Martin, J., Pultrusion, Plastics Products Design 19. Beever, W.H. and O'Connor, J.E., Pultruded
Handbook. Part B. Processes and Design for Thermoplastic Composite Structures, Int.
Processes, (ed Miller, E.), New York: Marcel SAMPE Symp. Proc., 32, 1309, 1987.
Dekker Inc., 1983, p. 37-74. 20. Beever, W.H. and O'Connor, J.E., Polyphenylene
8. Martin, J. and Sumerak, J.E., Pultruded Sulphide Pultruded Type Composite Structure,
Composites - The Case Against Aluminum 42nd Ann. Conf., Plastics/Composites Inst.,
Extrusions, Pultrusion Technology, Inc., RP /C 1987.
Reinforced Plastics/Composites '84; 21. Wood, AS., Pultrusion is Poised for New
Composites go to the Market; Papers presented Growth and It Won't be All Thermosets, Mod.
at Technical Sessions of the 39th Annual Plast. Int., 1976, 6(6),47-9.
Conference, New York. Jan 16-19, 1984, Session 22. Goldsworthy, W. Brandt, Thermoplastic
1-D, p. 5, Confer. 627, SPI Reinforced Plastics/ Composites: The New Structurals, Plast. World,
Composites Institute. 1984,42(9),56-8.
9. Laguan, 0., Pultrusion: Economic Aspects, 23. Kidd, AC., Wmding and Profile Production -
Applications and Design, Rev. Plast. Mod., 1985, Tape, Filament, Pultrusion-Extrusion,
50(349),61-6 (Spanish). Reinforced Plastics, In Proc. Electrical Symp.,
10. Spencer, RAP., Developments in Pultrusion, In Bristol, Feb 1975, Paper 5, p. 18 Preprint 627-61
Developments in GRP Technology - 1, (ed. Harris, 24. Smith, A, Pull Wmding Techniques Improve
B.), Barking: Applied Science, 1983, p. 1-36, Pultruded Products, Pop. Plast., 1988, 33(4),
6272. 42-3.
11. Martin, J.D., Pultrusion: The Other Process, 25. Shaw Stewart, D.E., Pullwinding Conf. Proc.,
Plast. Engng., 1979,35(3),53-7. 2nd Int. Conf. on Automatic Composites, Paper
12. Beck, D.E., New Processes and Prospects in 15, Noordwijkerhout, The Netherlands, 26-28
Pultrusion, Goldsworthy Engineering Inc., Sept. 1988.
Composite Solutions to Material Challenges: 26. Goldsworthy, W. Brandt, US Patent, 3674601
38th Ann. Conf Preprint, Houston, Tex., Augmented Curing of Reinforced Plastic Stock;
February 7-11, 1983, Session 6-B, p. 4, Confer. Issued 7/4/72.
627, SPI, Reinforced Plastics/Composites 27. Goldsworthy, W. Brandt, US Patent, 3 793 108
Institute. Augmented Curing of Reinforced Plastic Stock;
13. Roubinet, P., Curved Pultrusion, Composites Issued 2/19/74.
Plast. Renf Fibers. Verre Text, 24(4), July/Aug 28. Parry, T.V. and Wroksky, A.S., Effect of
1984, 69-73 (French). Hydrostatic Pressure on the Tensile Properties
14. Goldsworthy, W. Brandt, New Technology for of Pultruded CFRP, J. Mater. Sci., 1985, 20(6),
Continuous Reinforced Plastics Processing: Its 2141-7.
Called IPulforming' and It Permits Extrusion of 29. Bibbo, M.A. and Gutowski, T.G., Analysis of the
Variable Cross Section Parts and Curves, Mod. Pulling Force in Pultrusion, Antec 86. Plastics -
Plast. Int., 1979,9(9), Sept, 100-1. Value Through Technology. Proc. 44th Ann.
15. Goldsworthy, W. Brandt, Pulforming - The Techn. Conf., Boston, April 28-May I, 1986, p.
Changing Shape of Composites. 1430-2.012. SPE.
16. Ewald, G.w., Curved Pulforming - A New 30. Anon, Pultruded Fibre-Reinforcements, Plast.
Manufacturing Process for Composite News (Aust.), Nov. 1979,20.
Automobile Springs, Working Together for 31. Hill, J.E., Goan, J.c. and Prescott, R, Properties
Strength, 36th Ann. Conf., Washington, D.C., of Pultruded Composite Containing High
February 16-20, 1981, Session 16-C, p. 1-6, Modulus Graphite Fibers, SAMPE Qtly, 1973, 4,
Confer. 012 SPI Reinforced Plastics/Composites (2),21-7.
Institute. 32. Spencer, RAP., Advances in Pultrusion of
17. Goldsworthy, W. Brandt, Pulforming Makes Carbon Fibre Composites, Carbon Fibres, 2nd
Curved Pultrusions, Brit. Plast. Rubb., Nov 1985, Int. Conf., London, Feb 1974, Paper 21, 140-7,
p.36. Confer. 51FIC.
18. Nepasicky, J and Kannebley, G., Advantages 33. Evans, D.J., Designing with Pultrusions: From
and Limitations of Vertical and Horizontal the Idea to the Application, Composite
Pultrusion Processing. Examples of Typical Solutions to Material Challenges: 38th Ann.
Applications. Conf Preprint, Houston, Tex., February 7-11,
 References 523

1983, Session 6-A, p. 5, Confer. 627, SPI Process Variables, Mod. Plast., March 1985.
Reinforced Plastics / Composites Institute. 45. Sumerak, J.E. and Martin, J.D., Pultrusion
34. Browning, J., Synthetic Surface Veils for GRP Process Variables and their Effect Upon
Laminates, Int. Reinf Plast. Ind., 1986, 5(5) 14/6. Manufacturing Capability, RP /C Reinforced
35. Werner, RI., Improvements in Means of Plastics/Composites '84; Composites go to the
Evaluating Weathering Characteristics of Market; Papers presented at Technical Sessions
Pultrusions, Rising to the Challenge: 35th Ann. of the 39th Ann. Conf., New York. Jan 16-19
Conf. New Orleans, LA, Feb 1980, Section 4-E, 1984, Session 1-B, p. 7, Confer. 627, SPI
p. 6, Confer. 627. SPI Reinforced Plastics/ Reinforced Plastics/Composites Institute.
Composites Institute. 46. Sumerak, J.E., Understanding Pultrusion
36. Heritt, RW., New, High Performance, Fast Process Variables, Mod. Plast., 1985, 62(3), 58-64.
Curing Epoxy Resin System for Composites, 47. Sumerak, J.E., Understanding Pultrusion
Composite Solutions to Material Challenges: Process Variables for the First Time, Conference,
38th Ann. Conf. Preprint, Houston, Tex., Atlanta, GA, January 28-February 1, 1985,
February 7-11,1983, Session 19-D, p. 7. 627,SPI, Paper No 2-B, p. 8, Confer. 627, SPI Reinforced
Reinforced Plastics/Composites Institute. Plastics/Composites Institute.
37. Kershaw, J.A., New Epoxy Resin Systems for 48. Armstrong, RE, Calcium Carbonate,
Pultrusion, Shell Development Co., Composite Encyclopedia of Chemical Technology, Vol. 4, 2nd
Solutions to Material Challenges: 38th Ann. ed., Wiley Interscience, New York, 1964.
Conf Preprint, Houston, Tex., February 7-11, 49. Anderson, R and Riddel, R, Effects of Pigments
1983, Session 6-C, p. 4, 627, SPI Reinforced on Pultrusion Physical Properties and
Plastics/Composites Institute. Performance, Molded Fiberglass Co.; Morrison
38. McQuarrie, T.S., New Generation Resins for Molder Fiberglass Co., Composite Solutions to
Pultrusion, 33rd Ann. Conf., Washington, D.C., Material Challenges: 38th Ann. Conf Preprint,
Feb 1978, Section 8-E, p. 5, Confer. 627 SPI Houston, Tex., February 7-11, 1983, Session
Reinforced Plastics/Composites Institute. 6-H, p. 4, Confer. 627, SPI Reinforced Plastics/
39. Howard, RD. and Sayers, D.R, Development of Composites Institute.
New Methacrylate Resins for Use in Pultrusion, 50. Rufolo, A., Design Manual for Jointing of Glass
RP /C Reinforced Plastics/Composites '85. 40 Reinforced Plastics, US Naval Material
Years of Innovative Technology; Proc. 40th Ann. Laboratory Report, Navship 250-634-1, August
Conf., Atlanta, GA, January 28-February 1, 1951.
1985, Paper 2-A, p. 5, 627, SPI Reinforced 51. Hodgkinson, J.M., de Beer, D.L. and Matthews,
Plastics/Composites Institute. EL., The Strength of Bolted Joints in Kevlar R.P.,
40. Boinot, E and Daspet, Y., Phenolic Resins: A Proc. Composite Design for Space Application,
Boon to the Building and Transport Industries, The Netherlands, (esa SP-243) 15-18, October
Composites Plast. Renf Fibres Verre Text, 26, 3, 1985.
May/June 1986, 97-100. 52. Hollaway, L. and Baker, S., The Development of
41. Lo Scalzo, E., Phenolic Resins in Advanced Nodal Joints Suitable for Double Layer Skeletal
Composites, Mat. Flast. Elast., 1988, 3, p. 124-9 System made from Fibre/Matrix Composites,
(Italian). Part 7, Paper 21, Proc. 3rd Intern. Conf. on Space
42. Hunger, G.A., Pultruding Epoxy Resin, Structures (ed. Nooshin, H.), Barking: Elsevier
Reprinted from 43rd Ann. Conf. and Focus '88. Applied Science, 1984.
Proc. 43rd Ann. Conf. SPI Reinforced 53. Green, A.K. and Phillips, L.N., Crimp-Bonded
Plastics/Composites Institute End-Fittings for Use on Pultruded Composite
43. Kiernan, D., Tessier, N. and Schott, N., Sections, Composites, 1982, 13(3), 219-24.
Modification of Epoxy Resins for Improved 54. Hart-Smith, L.J., Adhesively Bonded Joints for
Pultrusion Processing, US Army Materials & Fibrous Composite Structures, Douglas paper
Mechanics Research Center, RP /C Reinforced 7740, Long Beach, California, 1986.
Plastics/Composites '85, 40 Years of Innovative 55. Anderson, R, Use of Pultruded Reinforced
Technology; Proc. 40th Ann. Conf., Atlanta, GA, Plastics in Energy Generations and Energy
January 28-February I, 1985, Paper 2-G, p 6 627, Related Applications, Working Together for
SPI Reinforced Plastics/Composites Institute. Strength, 36th Ann. Conf., Washington, D.C.,
44. Sumerak, J.E., Understanding Pultrusion February 16-20, 1981, Session 22-B, p. 1-3,
524 Pultrusion

Confer. 012. SPI Reinforced Plastics/Composites 62. Starr, T.F., Structural Applications for Pultruded
Institute. Profiles, TECHNOLEX, Composite Structures 2;
56. Morara, F. and Riva, G., GRP Conduits and Proc. 2nd Intern. Conf. Composite Structures,
Poles, Agrosistemi, Macplas, 1985, 10(67), Paisley, September 14-16, 1983, p. 192-216,627.
136--9. 63. Tickle, J.D., Halliday, G.A., Lazzarou, J. and
57. Pultrusions for Cable Rack, Brockhouse Group, Riseborough, B., Designing Structures With
Eur. Plast. News, 1982, 9(4), p. 36. Pultruded Fibre Glass Reinforced Plastic
58. Mallick, P.K., Djiauw, L.K. and Fesko, D.G., Structural Profiles as Compared to Standard
Design and Evaluation of a Pultruded Hybrid Steel Profiles, 33rd Ann. Conf., Washington,
Beam, Working Together for Strength, 36th Ann. D.C, Feb 1978, Section 8F, p. 8, Confer. 627.SPI
Conf., Washington, D.C, Febraury 16--20, 1981, Reinforced Plastics/Composites Inst.
Session 17-C, p. 1-5, Confer. 012. SPI Reinforced 64. Owens-Coming Fiberglas Europe SA, Fiberglas
Plastics/Composites Institute. in Action: FRP Lighting Poles, Burssels, 1977,
59. Head, P.R, GRP Walkway Membranes for PubIn, 13-Ch. 4-5, p/ 4 12 ins. 16/2/77
Bridge Access and Protection, 13th Reinforced 6272-6R.
Plastics Congress, 1982, Brighton, November 65. Mutch, w., Composite Utility Pole, Plast. World
8-11,1982, Paper 20, 97-91, BFP PubIn. 293, BPF, 1987, 45, (8), 43.
Reinforced Plastics Group. 66. Kliger, H.S., Yates, D.N. and Davis, G.CR,
60. Head, P.R, Pultruded Box Beams, Fibreforce Driveshafts: The Next Step for Composites?,
Composites, Ltd.; Maunsell Structural Plastics; Automot. Engng, 1980, 88(3), 100--3.
UK Dept. of Transport; Windfoll Ltd. 67. Roubinet, P. and Delacroix, B., Industrial
61. Anderson, RA. and Thomas, C, Development Development of Composite Leaf Springs,
of Large Hollow Rectangular Tubes for Composites Plast. Renf Fibres. Verre Text., 26,(3),
Structural and Electrical Markets - A Unique May /June 1986, p. 79-83 (French).
Application for Pultrusion, Rising to the 68. de Goncourt, 1. and Sayers, K.H., Composite
Challenge: 25th Ann. Conf., New Orleans, LA, Spring Systems, Composites Plast., Renf Fibres
Feb 1980, Section 4-A, p. 5, Confer. 627. SPI Verre Text, 1988, 28(3), 145-50 (French).
Reinforced Plastics/Composites Institute. 69. BTR Permali RP Ltd, Pultrusion Protects
Passengers, Europ. Plast. News, 15(3), 1988, 46.
PROCESSING THERMOPLASTIC 24
COMPOSITES
James L. Throne

24.1 INTRODUCTION
Thermoplastic polymers are seldom converted
into products without the time-dependent Product Requirements
strength
application of temperature, pressure, shear or Impact
Electrical
other types of mechanical manipulation. The Environmental
mechanical manipulation of plastics is called Rigidity
Temperature
'polymer processing'. Many polymer
processes and combinations of polymer
processes are used in modern commercial
manufacturing. The selection of a process to
produce a thermoplastic polymer product
from pellets, powder, or other granular forms
begins with general characteristics of the prod- Fillers
Fiber Type
uct itself. The two primary concerns to be met Fiber Length
in the manufacture of any polymer product Other Adducts

are:
Rheological
• Will the finished part meet all required, Concerns

specified and desired design criteria?

Compound Grades
• Can the product be produced at the mini-
mum cost for the projected market size? Economics

The first concern focuses primarily on the Final Polymer

ability of the polymer to meet mechanical and Recipe

environmental challenges throughout its func- Fig. 24.1 A schematic for choosing the proper poly-
tionallifetime. This is shown as the left branch mer and an attendant processl .
of the Fig. 24.1 schematic l . The second concern
deals with the economic ability to process the
acceptable polymers into the useful product, fillers, reinforcements or foam cells, usually
and this is shown as the right branch of Fig. have one or more adducts or additives that
24.l. alter the basic characteristics of the polymer.
Commercial pol~ers are rarely pure. Even Table 24.1 gives a short list of some of the
'neat' polymers or polymers that contain no adducts used with thermoplastics2 • Some of
these, such as coupling agents, are vital in
achieving the desired final solid mechanical
Handbook of Composites. Edited by S.T. Peters. Published performance of other adducts, such as fillers
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 and reinforcements. Typical fillers used in
526 Processing thermoplastic composites

Table 24.1 Adducts in thermoplastic polymers2 cussed in detail shortly. As is apparent, fillers
and fibers increase the polymer processing dif-
Antioxidants ficulty. In certain instances, as with continuous
Antistatic agents reinforcing fibers, conventional thermoplastic
Colorants and pigments processing cannot be used.
Coupling agents Part geometry is one way of classifying suit-
Flame retardants
Fillers
able polymer processes, Table 24.56 • This
Foaming agents classification is further amplified in Table
Heat stabilizers 24.67• Again, not all these processes are suit-
Mold release agents able for processing all polymers with all
Odor suppressors "'combinations of fillers or reinforcements. As
Plasticizers noted in Fig. 24.1, the cost of the polymer is
Processing aids only one aspect of the overall economics of
Emulsifiers
Lubricants
product manufacture. Table 24.7 gives a rela-
Reinforcing fibers tive comparison of the process costs for filled
Ultraviolet stabilizers and reinforced polymers8•
VIscosity depressants Extrusion and injection molding are the pri-
mary methods for producing foamed, filled
and discontinuous-fiber reinforced thermo-
plastics. As an example of the growth in
thermoplastic polymers are given in Table molding thermoplastic composites, in the
24.23. Typical fibrous reinforcements used in 1950s esentially all injection molded thermo-
thermoplastic polymers are given in Table plastics were neat or unfilled and
24.34. Filled, reinforced and foamed thermo- unreinforced. By the early 1990s, filled,
plastics offer great breadth of solid mechanical foamed and reinforced polymers accounted
properties. In many cases, they offer substan- for,more than 25% (wt) of all injection molded
tial processing challenges, as well. Nearly all parts. In certain instances, blow molding and
thermoplastic processes shape the polymer in rotational molding are possible.
its fluid state (The most notable exception to Thermoforming or rubbery sheet deformation
this is thermoforming, where forming occurs is now being applied to continuous fiber rein-
when the polymer is in a rubbery state. forced polymers. These processes are
However, thermofOrming depends on the pro- described below, with the objective of compar-
duction of sheet that is produced by ing general operating conditions of neat
calendering or extruding the polymer in its polymers with thermoplastic composites. The
fluid state.) Fillers and reinforcing fibers technical details of these processes are given
increase the viscosity of the polymer, making elsewhere9-14. Two axioms apply:
it more difficult to shape. As expected, pro-
• Axiom I: If the neat polymer is processed in
cessing difficulty increases with increased
conventional polymer processing equip-
filler or reinforcement lo.ading.
ment, composite versions of that polymer
There are more than twenty major types of
are usually processed in adapted or modi-
polymer processes5• Not all these processes are
fied versions of that equipment.
suitable for thermoplastics and not all thermo-
• Axiom II: Processing is always more diffi-
plastic processes are suitable for filled,
cult with composite versions of processable
reinforced or foamed thermoplastics. Table
24.4 lists most of the thermoplastic processes neat polymers.
that are used with neat, filled or reinforced An important corollary also applies:
polymers. Several of these processes are dis-
 Rheology, fiber flow and fiber orientation 527

Table 24.2 Fillers for thermoplastic polymers3

Silica products Metallic oxides

Minerals Zinc oxide
Sand Alumina
Quartz Magnesia
Novaculite Titania
Tripoli Beryllium oxide
Diatomaceous earth Aluminum trihydrate
Synthetic amorphous silica
Wet process silica Other inorganic compounds
Fumed coloidal silica Barium sulfate
Silica aerogel Silicon carbide
Molybdenum disulfide
Silicates Barium ferrite
Minerals Mica
Kaolin (China clay)
Metal powders
Mica
Nepheline silicate Aluminum
Talc Bronze
Wollastonite Lead
Asbestos Stainless steel
Synthetic products Zinc
Calcium silicate
Carbon
AlUll'tinum silicate
Carbon black
Glass Channel black
Glass flakes Furnace black
Hollow glass spheres Ground petroleum coke
Cellular glass nodules Pyrolized products
Glass granules or cullet Intercalated/ exfoliated graphite
Cellulosic fillers
Calcium carbonate
Chalk Wood flour
Limestone Shell flour
Precipitated calcium carbonate Comminuted polymers

• Corollary I: Adding foam, fillers or rein- economic importance of polymer processing

forcements to neat polymers will never and the technical challenge of predicting
improve their processability. molten polymer response to applied load have
resulted in an incredibly rich literaturel~21.
Neat polymer melts are considered as vis-
24.2 RHEOLOGY, FIBER FLOW AND FIBER
coelastic non-Newtonian fluids. Viscosity is
ORIENTATION
the measure of fluid resistance to applied load.
Rheology is the study of polymer flow. Shear The viscosities of oil and water are material
flow and elongational flow dominate polymer constants, independent of shear rate. Fluids of
processing. The great length of polymer chains this type are called Newtonian fluids. In
results in extensive entanglements and com- steady-state shearing flow, polymers typically
plicates the study of neat polymer flow. The exhibit shear-dependent viscosities, as with
528 Processing thermoplastic composites

Table 24.3 Fibers for reinforcing thermoplastics differences. Shear rate-dependent viscosity
and normal stress differences represent poly-
Cellulose fibers mer material functions and are not material
a-Cellulose constants.
Pulp preforms
Cotton flock Jr---~----r---~----~--~----,
Jute
Sisal
Rayon
Synthetic organic fibers
Polyamide (nylon, PA)
Polyester (PET)
Polyacrylonitrile (PAN)
Polyvinyl alcohol (PVOH)
Carbon fiber
Asbestos fiber
Fibrous glass
Filaments Fig. 24.2 Shear-rate dependent viscosity of low-
Chopped strand density polyethylene at 180°C (356°F) with titanium
Reinforcing mat dioxide filler in volume %. (Adapted and redrawn
Glass yarn from Ref. 22 by permission of the Academic Press.)
Glass ribbon
Whiskers
Aluminum oxide (Corundum)
Titanium dioxide The inclusion of particulates further compli-
Boron cates the rheological behavior of polymers
Boron nitride (Fig. 24.2). The effect of filler loading on
Boron carbide steady-state shear viscosity of polymers is
Metallic fibers approximated by:
Aluminum
Stainless steel (24.1)
Copper
Tungsten where Tw is the shear stress at the wall, yis the
shear rate, and Y, K and n are empirical con-
stants. This Hershel-Bulkley modep,24 is the
power-law equivalent of the Bingham model
for Newtonian plastic fluids. As anticipated,
low-density polyethylene in Fig. 24.222. When increasing particle surface area to volume
the shearing force on a Newtonian fluid is increases the viscosity of the polymer, even at
released, the fluid resistance instantaneously the same loading level and particle size distri-
ceases. When the shearing force on a poly- bution (Fig. 24.325 ). Increasing particulate
meric fluid is released, the fluid exhibits a loading levels usually decreases polymer vis-
measure of time-dependent reorganization, coelasticity as measured by the first normal
the extent of which depends on the extent and stress difference (Fig. 24.426).
duration of the applied forces. Viscoelastic flu- It has been suggested that'the shear-rate
ids exhibit fading memory of deformation dependent viscosity of particulate and fiber-
history. This is manifested by normal stress filled polymers can be predicted from a
 Rheology, fiber flow and fiber orientation 529

Table 24.4 Effect of adduct type on thermoplastic polymer processability

Degree of difficulty (0 = Easy, 9 = Difficult, X = Not done)

Polymer process Neat Filled" Short-fiber Long-fiber Continuous Continuous
reinforced reinforced organic fiber inorganic
reinforced fiber
reinforced
Extrusion o 3 5 8 b b
Sheet
Single screw
Extrusion o 3 5 8 b b
Profile
Single screw
Extrusion o 2 4 5 b b
Profile
Twin screw
Extrusion 1 3 5 7 b b
Compounding
Twin screw
Extrusion 2 5 8 X X X
Foam
Pultrusion X X X X 9 8
Injection molding 1 3 6 8 X X
Injection molding 2 4 7 X X X
Foam
Blow molding 2 8 9 X X X
Thermoforming 1 3 8 8 9 9
Compression molding X X X 6 5 5
Rotational molding 2 5 8 X X X

a Short aspect-ratio adducts.

b Pultrusion replaces conventional extrusion for continuous filament thermoplastic polymers.

temperature-invariant master curve: polymer temperature needed for the viscosity

MF~, determination (K), Ts = T + 50 (K), and T is
IOglOMF~, = the glass transition temperature (K)27. the
experimental results for many polymers and
8.86 (T2 - Ts) 8.86 (Tl - Ts) filler and fiber types show good agreement
(24.2) with the master curve.
101.6 + (T2 - Ts) 101.6 + (Tl - Ts) Blow molding and thermoforming are poly-
where MFI is the melt flow index determined mer processes that employ melt or rubbery
using a melt flow indexer die of LID = 3.8, phase stretching. Elongational viscosity shows
according to ASTM D1238. Tl is the ASTM rec- an increase in value with increasing filler load-
ommended test temperature (K), T2 is the ing in a manner similar to shear viscosity, Fig.
530 Processing thermoplastic composites

Table 24.5 Part geometry as means of classifying

processes for filled and reinforced thermoplastic
polymers6

Linear forming
Extrusion (sheet) IIJ
104

"
Extrusion (profile)
Pultrusion a.
Ii
u
c:::
Formation of a solid body by injecting I!!
into a cavity GI

Unfoamed injection molding ~

-
III
Foamed injection molding IIJ
I!!
Formation of a hollow object en
103
Blow molding ti
...0
E

-
Rotational molding
z
Sheet forming ...III
11lermoforming i!

24.528. As a first approximation, the elonga-

102~~~~~~~~~~~~~~
tional viscosity for filled polymers is 10-2 10-1
determined in a fashion similar to that for Shear Rate. S-1
shear viscosities, using equation (24.1)29. For
many neat polymers, the elongational viscos- Fig. 24.4 Shear-rate dependent first normal stress
ity is proportional to three times the shear difference for polystyrene at IS0°C (356°F) with car-
viscosity. For filled polymers the proportional- bon black filler in volume %. (Adapted and
ity is substantially greater than three and is redrawn from Ref. 26 by permission of Academic
usually shear rate-dependenf30. Press.)
Fiber flow in a shear field is far more com-
plex than particulate flow. Independent fiber As fiber concentration increases, fiber interac-
closed-orbit rotation as a consequence of tion inhibits independent fiber motion. At
shear is dominant in very dilute suspensions . moderately high fiber concentration or for

• • • • •
• • • •
No ......
• a

Fig. 24.3 Shear-rate dependent vis-

cosity of polypropylene at 200°C
(392°F) with two types of fillers
having particle sizes of 44 Jlm or
103 ~~I.-"""'-'-""""'u.L._....L.....-,-....&..&~u...._~...L....I..L""""'.L.I.I less. (Adapted and redrawn from
10~ 10-2 1CS" dJ Ref. 25 by permission of Academic
Shear Rate. S-1 Press.)
 Thermal properties offilled and reinforced thermoplastics 531

fraction, these overshoots can be sustained for

4
several minutes.
3
24.3 THERMAL PROPERTIES OF FILLED AND
REINFORCED THERMOPLASTICS
Most polymer processes depend on energy
interchange between the environment and the
interior of the melt or solid polymer. Fillers,
reinforcing fibers and cells dramatically affect
the thermal properties of polymers. Typically,
polymers have lower thermal conductivities
than inorganic fillers and fibers and higher
thermal conductivities than gases that are
3
used as foaming agents. Thermal conductivity,
2
heat capacity, density and thermal diffusivity
are the most important.

24.3.1 THERMAL CONDUCTIVITY

0.1 D.2 o.a 1M 05 0.7 1.0

Elongation Rate. 8-1 Both the solid and fluid thermal conductivities
of a filled polymer depend on the relative filler
content and the shape of the filler. The
Fig. 24.5 Elongation-dependent elongational vis-
cosity for polypropylene at 20QoC (392°F) with Halpin-Tsai equation modified by Nielsen32
calcium carbonate in volume %. (Adapted and yields useful values:
redrawn from Ref. 28 by permission of Academic
kc 1 + ABrp
Press). (24.3)
~ = 1-Barp
long fibers, fiber interaction acts to momen- where A =Is, - 1, and Is, is the Einstein coeffi-
tarily align fiber segments into bundles. The cient, Table 24.833 • rp is the filler volume
nature of the polymer flow field then deter- fraction, P is the maximum packing fraction,
mines whether these momentarily aligned Table 24.934, kc is the thermal conductivity of
fiber bundles remain oriented as the product the composite, kp is the thermal conductivity of
is produced. Increased shear implies the neat polymer, and kf is the thermal con-
increased fiber interaction. Fibers not oriented ductivity of the filler. B and a are given as:
in the flow direction are subjected to k/k -1
increased bending stresses around the fiber B = p (24.4)
bundles. The result is fiber length degrada- k/kp +A
tion. In addition to increased normal stress
difference with increasing fiber loading, fiber a =1+ (1 - P)rp
p2
(24.5)
reinforced polymers exhibit substantial shear
stress and normal stress difference overshoot The Einstein coefficient, Is" is a measure of the
at flow inception. These overshoots are attrib- shape of the filler particle. Is, =2.5 for regular
uted to the interactions of fibers moving from shapes such as spheres and becomes large for
an isotropic random state to the more ordered fibrous or acicular particles. The packing frac-
shear flow state31 • Depending on flow tion, P, for uniformly sized particles varies in
strength and fiber aspect ratio and volume value from 0.52 for random packing to 0.91 for
532 Processing thermoplastic composites

Table 24.6 Classification of polymer processes by size and shape for filled and reinforced thermoplastic
polymers7

Processes 1 2 3 4 5 6 7 8 9 10 11
Blow molding Hollow, Platen x x x x
thin
wall
Calendering Sheet Width of x x
roll,
Compression molding Platen x x x X X X
Sheet extrusion Sheet Width X X
of roll,
Die
Profile extrusion Linear Die X
Injection molding Platen X X X X X X X X
Injection molding Platen X X X X X X X
(foam)
Pultrusion Linear Die X
Rotational Hollow X X X X X
Molding
Thermoforming Thin wall Platen X X
1 = Shape limitation
2 =Factor limiting maximum size
3 = Complex shapes
4 = Controlled wall thickness
5 = Open hollow shapes
6 = Closed hollow shapes
7 =Very small items
8 =Plane area greater than 1 m 2
9 =Inserts
10 = Molded-in holes hexagonal packing. For fillers with random
11 =Threads sizes, P is typically about 0.85 to 0.9.
Thermal conductivity is a tensor quantity,
with unique values in each of the three princi-
pal directions. Because filled polymers are
Table 24.7 Ranking of polymer processes according relatively isotropic, the principal values of
to unit cosfl thermal conductivity are usually equal. For
continuous carbon graphite filament compos-
Process Cost ites, the thermal conductivity in the fiber
direction is usually much greater than that in
Calendering
Injection molding Very low the transverse directions. The packing frac-
tions are determined from Table 24.9 for the
Blow molding fiber and crossfiber directions. For random
Foam injection molding
mats, the typical packing fraction value is
Profile extrusion Low
Rotational molding about 0.5 in the cross-fiber direction.
Sheet extrusion
Thermoforming 24.3.2 HEAT CAPACITY
Compression molding Medium
The polymeric heat capacity or specific heat is
Pultrusion the isobaric change of enthalpy with tempera-
Machining High ture:
 Thermal properties offilled and reinforced thermoplastics 533

Table 24.8 Values of the Einstein coefficient, kEF for various types of fillers 33

Filler type kEa

Spheres, one size, maximum packing 2.50
Spheres, random close packing 2.50
Spheres, random loose placing 2.50
Rods or ellipsoids, random packingb
Aspect ratio = 2 2.58
=4 3.08
=6 3.80
=10 5.93
Mixed sizes, irregular shapes, minimum surface area 4.00
Mixed sizes, plates flakes c 5+
Agglomerates of spheresd 2.5/p.
Agglomerates, generally" f k/p p
a Correction factors for the mechanical case of Poisson's ratio, v, of the matrix:
v Factor
0.50 l.00
0.40 0.90
0.35 0.87
0.30 0.84
0.20 0.80
b At high rates of shear and low if! a' for the rheological case, these shapes tend to orient and to reduce Is,.
C These values may be approximated from oil absorption data or more precisely determmed by intrinsic viscosity.
d 4>. is the volume fraction of the agglomerate that is spherical.
4> is the volume fraction of particles in the agglomerate and Is, is the appropriate value for the particle shape.
These values are generally less than the maximum packing fraction for the particular particle shape.

Table 24.9 Effect of filler particle shape and packing type on maximum
packing fraction, P, for uniformly sized particles34

Particle shape Packing type Maximum packing fraction, P

Sphere Hexagonal close packing 0.7405
Sphere Face-centered cubic 0.7405
Sphere Body-centered cubic 0.60
Sphere Simple cubic 0.5236
Sphere Random close packing 0.637
Sphere Random loose packing 0.601
Fiber Parallel hexagonal packing 0.907
Fiber Parallel cubic packing 0.785
Fiber Parallel random packing 0.82
Fiber Random orientation 0.52

t
are constant. The enthalpies of most amor-
Cp = (~~ (24.6) phous polymers such as polystyrenics, acrylics
and polyimides are nearly linearly dependent
The specific heats of fillers and reinforcements on temperature and so their specific heats are
534 Processing thermoplastic composites

nearly constant. Semicrystalline polymers adduct loading. Thermal diffusivity also

such as polyolefins, liquid crystal polymers increases with increasing void fraction in cel-
and polyketones have enthalpies that exhibit lular composites35 • As a result, the rates of heat
the effects of melting and so their specific addition and of heat removal increase with
heats are temperature-dependent. The specific increasing adduct loading.
heat of a mixture of materials is given as:
24.4 EXTRUSION
cp,c =L x.I Cp,t. (24.7)
Extrusion is a means of producing continuous,
where x., is the weight fraction and cp,t. is the linear products in a steady state fashion. Both
specific heat of the ith species. single-screw and twin-screw extrusion
processes are used to produce sheet and pro-
files from composite thermoplastics.
24.3.3 COMPOSITE DENSITY
Single-screw extruder length-to-diameter
The density of a mixture of materials is given (LID) ratios range from 15:1 to 48:1 with 18:1
as: to 30:1 being usual. Twin-screw extruder LIDs
range from 12:1 to 30:1. Figure 24.6 is a
(24.8) schematic of a typical single-screw extruder36.
The process is generally characterized as hav-
where Pc is the density of the mixture, Xi is the ing four sequential segments37,38:
weight fraction and Pi is the density of the ith
species. • Solids conveying, where the polymer pel-
lets, flake, powder or granules are conveyed
from the extruder hopper to the conveying
24.3.4 THERMAL DIFFUSIVITY flights of the extruder. The particulates are
Thermal conductivity is the key material prop- compacted and begin to heat by shearing
contact with the metal screw and barrel.
erty in steady state heat transfer. The heat flux,
Table 24.10 tabulates the polymer character-
q, is proportional to the thermal driving force
istics that are important during solids
as:
conveying. Methods of determining some
(24.9) of these properties are discussed in Sections
24.2 and 24.3.
where T is temperature and x is distance • Plasticating or melting, where the gap
between the barrel and the root of the screw
through the composite.
is reduced, the polymer melts and the com-
The proportionality is thermal conductivity.
pressed solid cake breaks up. The degree of
In transient heat transfer, thermal diffusivity is
compression usually depends on the mor-
the proportionality, as:
phological nature of the neat polymer.
(24.10) Crystalline polymers need greater compres-
sion ratios than amorphous polymers.
where t is time and (X. is thermal diffusivity, Polymers with particulate adducts need a
lower compression ratio than neat poly-
given as:
mers. Table 24.10 also includes the polymer
(X. = -k (24.11) properties that are important in plastica-
pCp tion.
For most filled and reinforced polymers, ther- • Melt pumping, where the fully fluid
polymer is pressure- and temperature-
mal diffusivity increases with increasing
 Extrusion 535

Table 24.10 Polymer properties important in extrusion

Solids conveying
Frictional coefficients of particulate polymers with various metallic surfaces
Temperature-dependent thermal properties of particulate polymers:
Thermal diffusivity
Thermal conductivity
Heat capacity
Bulk density
Polymer density
Temperature dependent modulus of polymer
Temperature dependent yield strength of polymer
Plasticating
Temperature dependent and shear rate dependent melt viscosity,
Density of polymer melt
Density of polymer cake in solid bed
Polymer melt thermal conductivity
Melt pumping
Melt viscosity dependency on
Temperature
Shear rate
Pressure
Polymer melt thermal conductivity
Viscous dissipation
Thermal and shear degradation potential
Extruder die
Polymer shear sensitivity
Degradation potential- temperature limitation
Melt fracture potential
Extrudate swell
Rheological characteristics
Normal stress difference
Temperature dependent elongational viscosity

conditioned for the extruder die. The pri- - if the extruder die is annular, the result-
mary polymer property is shear viscosity. ing product is a hollow pipe or tube. This
Other important properties are given in tube or pipe is also a parison for extru-
Table 24.10. sion blow molding;
• Extruder die, where the polymer melt is - if the extruder die is irregular, the result-
shaped and presented to the take-up ing product is called a profile. Some of
equipment39 • The extruder die shape the major polymer properties that are
depends on the product being produced. important in profile extrusion are given
For example: in Table 24.10.
- if the extruder die is slot-like, the result- Extrusion dies are dissipative. That is, the
ing product is a planar sheet; polymer exhibits pressure drop through the
- if the extruder die is cylindrical, the die equal to that provided at the tip of the
resulting product is a rod; extruder screw.
536 Processing thermoplastic composites

Drive

Motor

Fig.24.6 Schematic of conventional single screw extruder without extrusion die. (Redrawn from Ref. 36 by
permission of Carl Hanser Verlag.)

Twin-screw extruders are used extensively Fiber length degradation is much less influ-
in producing linear composite products. Figure enced by screw speed than mixing time. As a
24.7 is an illustration of a cylindrical intermesh- result, twin-screw compounding extruders
ing twin-screw design40,41. Twin-screw provide less fiber length degradation than do
extruders are classified according to the relative tandem compounding single screw extruders.
screw rotational directions, whether the screws In fiber-reinforced extruded products, the
intermesh and the relative screw speeds, Table average fiber orientation for discontinuous
24.11. Comparative twin-screw performance is fibers is up to about 20° from the axis4B. The
given in Table 24.12. Despite some important orientation is the result of converging flow
shortcomings, twin-screw extruders are desired from the extruder screw tip to the die end.
for low shear and controlled feed rates, impor- (Converging flow is one of the standard die
tant aspects of compounding composite design critia for neat polymer extrusion.
thermoplastics as well as extruding them into Accelerating flow allows molecular alignment
uniformly consistent products. in the axis direction and tends to minimize
Increasing time in the shear field results in a extrudate swell shear dependency49.)
linear decrease in fiber length (Fig. 24.841). Extrudate swell decreases with increasing

Compression

Fig. 24.7 Schematic of standard configuration of cylindrical twin screws40 • (Redrawn by pennission of Carl
Hanser Verlag.)
 Extrusion 537

• ,.L.-......L---L1O-......I--..J-_~-:l:..--'--~..
ToO. Volume Percent. ...

Fig.24.9 Effect of volume fraction of titanium diox-

• • ...
300 Post-Injection Molding
• ide on extrudate swell ratio for high-density
polyethylene at 180°C (356°F). Capillary LID = 28
and shear rate = 10 S-l. (Redrawn from Ref. 50 by
permission of Academic Press.)

Mixing Length. Diameters

Fig. 24.8 Effect of residence time in shear field on

glass fiber length in injection molding and extru-
sion47• (By permission of Society of Plastics
Engineers. )

Table 24.11 Classification of twin-screw extruders

Intermeshing screws'
Corotating screws
Low speed extrusion for profiles, foams, filled polymers, short-fiber profiles
High speed extrusion for compounding and devolatilization
Counter-rotating screws
Conical extrusion for profiles
Cylindrical extrusion for profiles
Non-intermeshing screwsb
Counter-rotating separated screws
With blades for kneading, compounding
High speed for in situ polymerization
Corotating screws
Not used in practice
Counter-rotating tangential screws
High speed for compounding, devolatilizing
Low speed for plasticating fluffy, bulky regrind
a Intermeshing screws are also classified as fully or closely intermeshing or partially intermeshing42-4'3.
b Non-intermeshing screws are also classified as separated non-intermeshing and tangential non-intermeshing
screws42.
538 Processing thermoplastic composites

filler loading (Fig. 24.950) and with increasing of hoop reinforcement is related to the channel
fiber orientation in the hoop direction. width expansion52 •
Orientation in the extrusion direction is
desired for profiles that are designed for
24.5 INJECTION MOLDING FILLED AND
strength in the bending direction. For pipe and
REINFORCED THERMOPLASTICS
tubing, on the other hand, the strength in the
hoop direction is half that in the axial direction Injection molding is a means of producing dis-
for an isotropic polymer. Reinforcement and crete products on a cyclic basis53,54. The
hence fiber orientation is desired in the hoop injection molding machine consists of two
or cross-extrusion direction. This is achieved major parts: the plasticating and pumping sec-
by using a diverging die section following the tion, and the clamping mechanism.
converging section (Fig. 24.1051 ). The amount
24.5.1 THE PLASTICATING AND PUMPING
Annular o.e SECTION, FIG. 24.1155
A single Archimedean screw similar to an
extrusion screw acts to convey and compress
Extruder
the solid thermoplastic, plasticate and melt the
polymer and melt conveyor pump the poly-
mer melt through a non-return valve or check
ring to an accumulation region ahead of the
screw. As the polymer accumulates between
the screw tip and the nozzle, it pushes the
Converging SectkJn
screw backward away from the nozzle, effec-
tively shortening it. When a suitable amount
Fig. 24.10 Expanding mandrel extrusion die with of polymer is accumulated, the mold is closed,
converging section for extrusion of highly fiber- the nozzle abuts the sprue, the screw advances
reinforced polymers. (Redrawn from Ref. 50 by and the melt is pushed into the mold. Injection
permission of Society of Plastics Engineers.) molding screws have LID ratios of 15:1 to 30:1.

Plaslicaling Screw
Nozzle

Non-Relurn Valve
Frame Electromechanical Drive

Fig. 24.11 Schematic of plasticating and injection portion of conventional reciprocating screw injection
molding machine. The nozzle inserts into the sprue of the mold mounted on the press shown in Fig. 24.12.
(Redrawn from Ref. 55 by permission of Carl Hanser Verlag.)
 Injection molding filled and reinforced thermoplastics 539

Table 24.12 Advantages and disadvantages of twin-screw extruders44

Advantages
Controlled compaction of powders in feed zone
Powder feeding independent of friction with screw or barrel
Starved feed decouples feed rate, screw speed and extent of viscous shear heating
Rapid but gentle heating and plasticating of thermally sensitive polymers
Kneading action provides superior thermal and melt mixing and homogenization
Control of pressure build-up by proper element selection
High outputs at low speeds, minimizing shear heating
Ability to custom design processing sections
Gas injection easy to locate
Addition of adducts, fillers, reinforcing elements relatively straight forward
Comparatively little wear on extruder elements when processing aggressive fibers, fillers
Comparatively little fiber attrition
Disadvantages
Equipment cost per unit output very high
Output limited when compared with tandem extruders
Screw wear harder to predict
More difficult to mix gases into polymer melts in direct gas injection
Maintenance tends to be specialized, expensive
Although flexibility in changing mixing elements touted, no real way of determining a priori what
mixing elements are best for optimum throughput"
Pressure build-up entering the die is less effective
The thrust bearing remains the primary mechanical weakness
• Recently, computer models have been developed to aid in understanding polymer flow in certain elements such as for-
ward pumping screw elements, backward pumping screw elements and kneading disc elements45 • Owing to the
complex geometry, twin-screw extruder configuration and element design has not achieved the sophistication of sin-
gle screws46 •

The length ratio of solids conveying to plasti- methods are common means although new
cating to melt pumping is 50:25:25. Injection electric drives are considered to need less
molding screw compression ratios are usually maintenance. In addition, means for ejecting
the same as those for extrusion. the part from the mold cavity are usually
attached to the moveable platen. The polymer
is transferred from the accumulator section of
24.5.2 THE CLAMPING SECTION
the screw into the mold cavity by ram-
Figure 24.1256 shows a fully hydraulic clamp, advancement of the screw at relatively high
one method of holding the mold halves closed shear rates of 100--10 000 s-1, with transfer
against the pressure of the injecting melt. The times of seconds. Since polymers are com-
clamp consists of a stationary platen and a pressible57 at injection pressures and exhibit
moveable platen. The platens are guided open decreasing density with increasing tempera-
and closed along tie bars. The polymer is intro- ture58, pressure is applied to the polymer in the
duced through the sprue, a hole in the mold cavity, runner system and transfer lines
stationary platen. The moveable platen con- after the cavity is filled, until the polymer cools
tains a means of applying pressure to hold the sufficiently to hydraulically seal the mold cav-
mold halves closed. Hydraulic and mechanical ity. This packing pressure is most important in
540 Processing thermoplastic composites

Traversing
Cylinder

Fig. 24.12 Schematic of hydraulic clamping portion of conventional injection molding machine. (Redrawn
from Ref. 56 by permission of Carl Hanser Verlag.)

Table 24.13 Polymer properties important in injection molding

Particulate polymer frictional coefficients

Solid and liquid polymer thermal properties
Thermal diffusivity
Thermal conductivity
Heat capacity
Particulate bulk density
Pressure and temperature dependent melt density
Polymer shear sensitivity
Degradation potential- temperature limitation
Shear and temperature dependent viscosity over the shear rate range of 0-10 000 S-l
Melt fracture potential
Extrudate swell
Rheological characteristics
Normal stress difference
Temperature dependent elongational viscosity
Melt compressibility
Volume expansivity
Isothermal compressibility
Pressure dependent shrinkage
Crystallization kinetics
Heat of crystallization
Rate of crystallization
 Injection molding filled and reinforced thermoplastics 541

the injection molding of all polymers, includ- result exhibit differential shrinkage values that
ing heavily filled and reinforced compounds. increase with increasing loading (Fig. 24.1359).
A tabulation of polymer properties impor- As a first approximation, the flow of a filled
tant in injection molding is given as Table molten thermoplastic can be considered simi-
24.13. Note that many of these properties are lar to the flow of the neat polymer, regardless
important in extrusion, as well.
For filled and reinforced polymers, mold
design is critical. There are several elements to
a mold:
• Sprue and runner system. The sprue
directs the polymer from the injection
molding machine nozzle into the mold
body through the stationary platen wall.
The runner system directs the molten poly-
mer to the appropriate cavities.
• The gate. Polymer flows into a mold cavity
through a constriction called a gate. The ~
size and location of the gate are critical to 0
injection mold part performance, as noted
below. 1--
Diameter .,1
• Mold temperature control. Coolant lines
are normally placed parallel to the machine
platens. Adequate coolant flow to all part
.-.1.Cup
surfaces is important in minimizing part f
distortion and warpage. Warp = Cup/Diameter
• Part removal system. Typically, parts are
molded under substantial pressure of 2.0
10-15 MPa. Ejector pins and rings are used
to press the part from the mold surface after *'
Q)
0) Cross-Flow
the mold has opened. In addition, other ~
devices such as sliding cores and unscrew- .lII:
I: Relative Measure
ing devices are employed to meet certain 'i: of Warp
..r:.
-
design criteria. en to
Crystalline neat polymers exhibit greater over- "0
all shrinkage and more differential shrinkage '0
~
...0
--
0.5
than amorphous neat polymers. Filled poly-
mers exhibit lower overall shrinkage but may
show substantially greater differential shrink- ...
~ 0
age, called warp or cupping, than neat a.. 0 10 20 30 40
polymers. Typically, the value for shrinkage of
a filled or reinforced polymer is less than that Glass Fiber Content, %(wt)
for the neat state of the polymer, and the value
shows substantially less injection-pressure sen-
Fig. 24.13 Differential shrinkage, cupping and
sitivity. Fiber-reinforced polymers exhibit warping of glass fiber reinforced polyacetal
reduced in-flow shrinkage and greater cross- homopolymer (polyoxymethylene). (Redrawn from
flow shrinkage than neat polymers and as a Ref. 59 by permission of Carl Hanser Verlag.)
542 Processing thermoplastic composites

of the loading level. The amount of force mold system (Fig. 24.1463). (It has been mathe-
needed to transfer the polymer from the injec- matically shown that the fountain flow effect
tion molding machine to the mold cavity holds for cylindrical and planar flow fields 61 •
increases with increase in apparent shear vis- Even at very high fiber loading, individual
cosity. Since the transfer pressure is usually fibers or fiber aggregates align themselves
fixed by the machine hydraulic system, the with the flow streamlines62.) As is apparent
transfer rate usually decreases with increasing from Fig. 24.1564, there is no appreciable shear
filler loading. Usually, sprue, runner and gate field in the center core. As a result, fiber
dimensions are increased to accommodate the
lower compressibility of the filled polymer Wall

melt. Solid particles tend to migrate away

from planes of highest shear. Since shear is
m-tRale
greatest near the mold confining surfaces,
parts of highly filled polymers tend to have
resin-rich surface layers. These layers are typi-
cally 10-100 pm in thickness. Similarly, weld
lines or planes perpendicular to meeting flow Center1ine

fronts, also tend to be resin-rich and are usu-

ally weaker than regions on either side of the
weld line60• Increasing levels of filler tend to
minimize the problem of jetting, where a neat
polymer forms a thread rather than a radial
disk as it enters a mold cavity from the gate.
For fibrous reinforcements, where the LID
aspect ratio is greater than about 10:1, orienta- WaR
tion during flow is very important. The fiber Fig. 24.15 Typical shear rate and velocity profiles
rotation attempted in the shear field is compli- for nonisothermal flow during injection mold filling
cated by the fountain-like flow of the of a cavity. (Redrawn from Ref. 64 by permission of
polymeric mass as it advances through the Society of Plastics Engineers.)

Fiber Bundle Orientation

~ ~ldvar'''ina Row-f'ront

Straamli'-

Fig. 24.14 Fountain flow in injection molding, showing fiber bundle behavior along a streamline and
development of frozen layer at cold walls. (Redrawn and reinterpreted from Ref. 63 by permission of
Society of Plastics Engineers.)
 Injection molding filled and reinforced thermoplastics 543

bundles move preferentially from the molded;

upstream organizing region transverse to the • at the very center, a very thin plane of fibers
center axis toward the flow front. In doing so, that are oriented in the flow direction.
they orient as the advancing front is formed.
The general nature of flow into an injection
As shown in the schematic, the interface
mold cavity through a constriction or gate is
between the molten core and the frozen skin
radial. Since there is substantial in-plane
grows rapidly into the melt once the advanc-
stretching, fiber orientation in the gate region
ing front has passed. The oriented fibers are
of long narrow molded parts are quite similar
aligned parallel to the flow direction. As with
to fiber orientation in center-gated axisymmet-
extrusion, the surface at the mold surface is
ric molded parts. If there is no in-plane
resin-rich and, in many cases, essentaUy fiber-
orientation, cores typically have random pla-
free. When the mold is full, the structure
nar orientation. Typically, the thickness of the
shows three, five, seven and even nine layers
transverse fiber orientation in the core
in cross section (Fig. 24.1665,66):
decreases with distance from the gate. For cer-
• the resin-rich surface layer, up to 10 rom tain polymer-fiber combinations and certain
(0.4 in) in thickness; injection speeds and mold temperature, the
• a relatively thin layer of fibers oriented in resin-rich surface layer may not be apparent.
the flow direction, with some of these fibers In fact, fiber prominence on molded part sur-
tipped toward the flow axis; faces is apparent at long flow length-ta-cavity
• a relatively thick layer of fibers oriented thickness ratios, in regions very near the gate
transverse to the flow direction, with a sub- when the flow is turned as it enters the mold
stantial amount of in-plane randomness, cavity, at low neat polymer viscosities, for cold
depending on the nature of the part being polymer melts, and at slow injection rates.

•
Flow Direction

-- - -------- --------- ---

- -- -- --- -- -- ---
---
----
-
- - --- - - - - - - - - - ----- ----
- - --
- -- ---
- - - ---- ----
- - - - - - - - - --
--

..,... -:~.........~~-I,........... :: ... ..:: ~. _t_ _ _ __-=--=.-:--=-=--==-_-==

Transverse OrienIation
In-flow OrientatiOn

Fig. 24.16 Schematic of the development of fiber orientation in injection molding. showing disappearance
of fiber-free surface layer and centerline transverse fiber orientation as flow proceeds into the mold cavity.
(Redrawn and reinterpreted from Ref. 65 by permission of Society of Plastics Engineers.)
544 Processing thermoplastic composites

Furthermore, the center line or plane of in-axis Weld lines are particularly bothersome with
fibers is usually the result of continuing injec- fiber-reinforced polymers. The primary con-
tion as the flow channel freezes closed. The cern is the undesirable fiber orientation at the
resulting high shear orients the fibers in the interface between two advancing flow fronts
flow direction. This layer may not be apparent (Fig. 24.17'°). Typically the relative weld line
in all fiber-reinforced parts. In very thin parts, strength decreases with increasing filler load-
fibers are nearly always oriented parallel to ing and aspect ratio71 (Table 24.14).
the mold surface67• The flow in an injection
mold is mathematically modeled using
Hele-Shaw or creeping flow 68,69.

• •
Flow Direction Flow Direction
Wekj Line

Fig. 24.17 Fiber orientation at the interface of two impinging flow fronts, creating a weak weld line.
(Redrawn from Ref. 70 by permission of Carl Hanser Verlag.)

Table 24.14 Relative weld line strength for neat and reinforced polymers 72,73

Polymer Tensile strength retention for various glass loadings, %

0 10 20 30 40
Polypropylene 86 47 34
SAN 80 40
Polycarbonate 99 86 64
Polysulfone 100 62
Polyphenylene sulfide 83 38 20
Nylon 66 (PA-66) 83-100 87-93 56-64
 Thermoforming and compression molding 545

24.6 THERMOFORMING AND Table 24.15 Polymer properties important in ther-

COMPRESSION MOLDING moforming

For neat polymers, thermoforming and com- Temperature dependent polymer hot strength
pression molding are different disciplines74,75. Elastic modulus at forming temperature
Elongational viscosity at the forming temperature
Thermoforming begins with a formed sheet of Strain-rate hardening at high elongation
plastic that is heated to the rubbery state of the Other properties that are important in extruding
polymer, usually a few degrees above its glass polymer sheet
transition or melting temperature. The rub- Temperature dependent thermal properties of
bery sheet is then pressed with relatively little rubbery solid polymer:
differential force into or onto a cooled single- Thermal diffusivity
sided mold and held there until the polymer Thermal conductivity
Heat capacity
temperature is substantially below the form- Polymer density
ing temperature. (In traditional vacuum
forming, the space between the sheet and the
mold surface is evacuated, thus applying dif- form the composite from the planar state to a
ferential pressure of up to 0.1 MPa. In pressure useful product. Mechanical means such as
forming, air pressure is applied to the free sur- matched dies, or hydraulic forces using super-
face of the sheet, thus applying differential air plastic aluminum or polyimide films, replace
pressure of up to 1 MPa76 .) The desired part is pneumatic forces when the differential form-
then trimmed from the web. Compression ing pressures exceed 1 MPa. The foamed
molding is usually reserved for thermosetting polymer, on the other hand, cannot be heated
polymers, but certain thermoplastic polymers to the same forming temperature as the
such as UHMWPE, PTFE, and certain poly- unfoamed polymer without dramatic cell col-
imides are compression molded. These lapse. As a result, foamed polymers are
polymers are characterized as having very formed at temperatures substantially below
high viscosities even at temperatures hun- forming temperatures for the unfoamed poly-
dreds of degrees above their melting or mers. Mechanical forces, such as matched dies,
processing temperatures 77• A polymer of this are used for forming foamed polymer sheet
type is compressed as a powder into a pre- into useful products. Typically, the ratio of
form, heated in a convection oven to the polymer modulus at forming temperature to
processing temperature, and transferred to a applied pressure, E(T)/P, should be in the
compression molding press where it is com- range of 2-10 with a value of 5 most typical for
pressed between heated mold halves. Typical traditional thermoformingB°. As an example,
molding pressures are 7-35 MPa. The mold is the temperature-dependent flexural modulus
then cooled until the formed part temperature of glass fiber-reinforced polyetherimide is
is substantially below the polymer processing shown in Fig. 24.1881 • At 200°C (392°F), the
temperature78• Table 24.15 gives polymer prop- modulus of neat PEI is 2 MPa. The pressure
erties important in thermoforming. required to thermoform this polymer at this
For filled, foamed and reinforced polymers, temperature is about 0.4 MPa. For 30% (wt)
the boundaries between thermoforming and glass fiber-reinforced PEl the modulus is
compression molding blur. Frequently, the 7 MPa and the required pressure is about
composite sheet forming process is simply 1.4 MPa. Pneumatic pressures at this level are
called stamping79 • Fillers and discontinuous possible, but mechanical forming is preferred.
fibers stiffen the polymer so that even at the In many cases, the elastic modulus of the com-
upper forming temperature of the polymer, posite exibits similar temperature dependency
substantial differential force is required to to the elastic modulus of the polymer matrix,
546 Processing thermoplastic composites

'O~---r---""T""----'---"""'-----'
'400

.:€
'200

1IlOO
1\1
II..
Q ::; 2O%(wt)
iii 8
iii 800 ::I
..i! :;
::I 'g
OC 0 1O'IIIwt)
::;
::; eco
0

~
•
~::I
::I
400
~
~ 200
iL 2

0
Fig. 24.18 Temperature dependent
0 -50 50 100 200 flexural modulus of glass fiber rein-
Temperature. "C forced polyetherimide. (Redrawn
from Ref. 81 by permission of Carl
o 100 200 3DO
Hanser Verlag.)
Temperature, "F

with the isothermal value of the modulus of the rubbery sheet is key to the forming
increasing monotonically with increasing filler process. Figure 24.20 is a schematic of a temper-
or fiber loading (Fig. 24.1982). ature-dependent stress-strain curve for a neat
Thermoforming is a surface-generating polymer that exhibits a yield at low tempera-
process. That is, the total area of the finished tures83• The effect of filler and discontinuous
part plus web is greater than the total area of fiber on the shape of this curve is shown in Fig.
the initial sheet. The dominant method of gen- 24.2184. Typically, the initial slope of the curve,
erating surface is biaxial stretching of the the tensile modulus increases, the yield point
rubbery solid polymer. The shape of the tem- disappears and the elongation at break
perature-dependent tensile stress-strain curve decreases rapidly with increasing fiber or filler
loading. This implies that the forming of com-
12~------r----~~-----'-----~ posite thermoplastics requires high
temperatures, substantial forces and the parts
10

O~O-------1O~----~m~----~~~-----'.~

Glass Fiber Content. 'l6(wt) Elongation

Fig. 24.19 The effect of glass fiber loading on flexural Pig. 24.20 Schematic of temperature dependent
modulus of polysulfone at 25°C (77°P). (Redrawn stress-strain curve for a neat polymer exhibiting a
from Ref. 82 by permission of Carl Hanser Verlag.) yield point.
 Thermoforming and compression molding 547

so produced are restricted to relatively shallow

draws. Typically, polymers containing nonwo-
ven continuous fiber mat and short- and
long-fiber chopped fiber mat are formable with
matched metal dies. Composites containing
woven fiber mat are not as formable85• For very
long glass fiber- and continuous fiber-rein-
forced composites, the extensibility of the sheet
is so restricted that even shallow draw parts
cannot be formed without substantial polymer
migration (the squeezing of the polymer matrix
from the fiber bundle is called 'percolation'),
compression buckling, pleating wrinkling and
fiber breakage when standard matched die
molding techniques are used. (Note that this
problem is not unique to fiber-reinforced ther-
moplastic composites. Paper, organic synthetic
paper and mixed fiber paper are nonwoven
fiber structures that are very difficult to form
into deeply drawn products86•87.) It is well
known that buckling, folding and pleating are
minimizM by keeping the sheet under tension
Elongation throughout the forming process. Further, for
Fig. 24.21 Schematic of the effect of fiber content on composites with limited extensibilities, the
isothermal stress-strain curve for a polymer material that makes up the formed shape must
exhibiting a yield point when unreinforced. come from the region outside the formed

Drive Shaft

Guide Rod

Slide Rod

Forming Plug --;IH<~

Stop

Clamp Frame

Mold Open Mold Closed

Fig. 24.22 Schematic of a mold designed to slip-form continuous fiber-reinforced composite. (Redrawn
from Ref. 88 by permission of Society of Plastics Engineers.)
548 Processing thermoplastic composites

Intraply Shear F
Resin Percolation Through Fiber Bundle

Transverse Squeezing Flow

F

oo!ft.~o
o~_oo
~~oo
crCl...~~
o'fJ8f>o 0

Fig. 24.23 Schematic of fiber-resin matrix interaction during deformation for several types of deforma-
tions. (Redrawn and reinterpreted from Ref. 93 by permission of Society of Plastics Engineers.)

shape. Slip fonning is the principal method for

accomplishing this (Fig. 24.228S-9O). The key is to
supply sufficient tension to the sheet to mini-
mize folding but not an excess amount that will
tear the very hot sheee1,92.
Bulk mechanical deformation of the rein- Heater
forced sheet is not the only concern when
forming reinforced sheet93 • Figure 24.23 shows Mold
several fiber-matrix interactions that take
place locally during composite thermoform-
ing94 • Intraply shear, interlaminar shear or slip
and interlaminar rotation are local shearing
effects that involve the fiber. Local resin flow
between fibers, fiber bundles and plies allow Mold Open
local distortion of the matrix. Matrix percola-
tion and squeeze flow are predominantly resin
effects. Voids and delaminations are micro-
scopic defects that are attributed to local
fiber-matrix interactions during forming.
Compression molding is flow molding or
'squeezing flow' of polymer between two mold
halves (Fig. 24.2495). The flow behavior for a
filled or short-fiber reinforced polymer mimics Mold Closed
that for the neat polymer96• The flow is usually
characterized as planar radial with the wave- Fig. 24.24 Schematic of flow molding, flow form-
front being fountain-like as with injection ing or squeezing flow of nonwoven fiber-reinforced
molding. Fiber orientation is quite similar to thermoplastic resin composite. (Redrawn from Ref.
that observed with center-gated injection mold- 95 by permission of Society of Plastics Engineers.)
 Other processes for filled and reinforced polymers 549

ing&'. Compression molding is also used with halves close on the parison, pinching it
long fibers or continuous fibers. In one exam- between the mold halves. The remaining por-
ple, a fiber preform is inserted in the mold tion of the parison is then inflated against the
cavity and a heated neat polymer preform is mold walls (Fig. 24.25). Parison thickness con-
placed on top. The press is closed to 35 MPa or trol is the key to uniform wall thickness.
more, squeezing the polymer into the fiber pre- Paris on thickness uniformity is governed by
form. The process works best if the polymer is extrudate swell as the polymer exits the die
crystalline and has a very low melt viscosity at and by parison sag owing to the parison
the molding temperature. Compression hanging weight. The former is a function of
molded composite parts are usually more com- the viscoelasticity of the polymer in general
plex than thermoformed composite parts. and the normal stress difference in particular.
Voids, warping, fiber prominence at the part The latter is a function of the mass of the pari-
surface intially against the mold and resin rich- son, the extrusion time and the elongational
ness at the other surface are typical problems viscosity of the polymer. Neat polyethylenes
attributed to the forming process. exhibit extensive extrudate swell. Fillers and
fibers reduce the extrudate swell in propor-
24.7 OTHER PROCESSES FOR FILLED AND
tion to the filler or fiber loading (Fig.
24.9 101 ,102). Filler loading increases parison dif-
REINFORCED POLYMERS
ferential weight. And fillers and fibers cause
As noted above, foamable, filled and short-
fiber reinforced thermoplastics are usually
processed in fashions similar to the neat ther- Mold Half Mold Half
moplastic. Filled and reinforced hollow
structural parts are fabricated by blow mold-
ing and rotational molding. Rotational
molding is also used to produce multilayer
structures having foam cores97 •
Polyolefins account for approximately 90%
(wt) of all non-disposable industrial products
produced by blow molding and nearly all
products produced by rotational molding.
Mica flake reinforced HDPE to 30% (wt) has Mold Open
been accumulator blow molded into flat struc-
360" Pinch-Off
tural parts and ducts for automotive, truck
and agricultural vehicles since the early
1960s98-100. Graphite-filled HDPE is blow
molded into conductive electrical boxes and
doors. Glass fiber-reinforced HDPE is blow
molded into flotation devices, sailboards and
kayaks. Accumulator blow molding machines
are used to extrude a large quantity of plastic
in a short period of time. The filled or short-
fiber reinforced polymer is plasticated at a
constant rate in a conventional screw Mold Closed
extruder. The melt is stored in an accumulator
until the mold is cleared, then extruded in sec- Fig. 24.25 Top view schematic of lay-flat squeezing
onds into a parison or vertical tube. The mold of extruded parison in structural blow molding.
550 Processing thermoplastic composites

Table 24.16 Polymer properties important in blow molding

Shear and temperature dependent viscosity over the shear rate range of 0--10 000 S-1
Melt fracture potential
Nonisothermal melt strength or temperature dependent elongational viscosity
Extrudate swell
Rheological characteristics
Normal stress difference
Temperature dependent elongational viscosity
Polymer strain recovery
Molecular weight dependency
Molecular weight distribution dependency
Particulate polymer frictional coefficients
Solid and liquid polymer thermal properties
Thermal diffusivity
Thermal conductivity
Heat capacity
Particulate bulk density
Pressure and temperature dependent melt density
Polymer shear sensitivity
Degradation potential - temperature limitation
Effect of orientation on gas permeability of polymer
Strain-oriented crystallinity levels and effect on
Barrier properties
Permeability
Tensile strength

Fig. 24.26 Schematic of co-axis

rotational molding for filled or
lightly reinforced polyolefins 103•
 Other processes for filled and reinforced polymers 551

the parison to cool more quickly. As a result, ment of the mold and so rarely yield useful
parison wall thickness control must be products. Fibers with aspect ratios of 1000 or
changed radically when blow molding filled less can be successfully molded if the maxi-
or reinforced polymers. Table 24.16 gives a list mum fiber content is less than about 15% (wt).
of polymer properties important in blow Lower fiber loadings may be necessary for cer-
molding. tain mold geometries and certain types of
Rotational molding is an atmospheric polyethylenes. If the fiber loading is too high,
process in which polymer powder is charged the fibers orient at right angles to the mold sur-
to a metal clam-shell mold. The mold is rotated face, producing an unacceptable setaceous
about the polar or major and equatorial or inner surface104• One method of forming a hol-
minor axes (Fig. 24.26103), while being heated in low composite is to impregnate nonwoven
a forced air convection oven for several min- fiber mat with up to 30% (wt) electrostatically
utes, until the powder sticks to the mold charged polymer powder, then sinter the struc-
surface, melts and densifies into a void-free ture to fuse the powder to the fibers. The hot
hollow object. The mold is then air- and/or structure is then manually pressed against the
water mist-cooled, the part removed, the mold mold surfaces prior to adding additional poly-
recharged and the process repeated. mer powder, closing the mold, rotating it,
Polyethylene is the major rotational molding heating it to the forming temperature, and
polymer, with particle sizes ranging between cooling it in standard fashion 105 • Evacuating
50 pm and 500 11m. Coarse particle fillers hav- the mold through the rotating concentric shafts
ing particle sizes of about 50 11m or more, such is helpful in minimizing voids but is usually
as CaC03, milled glass and glass cullet are suc- quite difficult to achieve successfully. Table
cessfully molded to loadings of 30% (wt). Fine 24.17 gives a list of polymer properties needed
particle fillers such as Ti02, carbon black and for rotational molding.
talc fluidize readily in the tumbling environ-

Table 24.17 Polymer properties important in rotational molding

Particle size distribution

Temperature dependent thermal properties of particulate polymer
Thermal diffusivity
Thermal conductivity
Heat capacity
Bulk density
Polymer density
Molecular characteristics
Molecular weight
Molecular weight distribution
Crystallization kinetics
Heat of crystallization
Rate of crystallization
Crosslinking characteristics
Reaction rate
Level
Zero-shear viscosity at melt temperature
552 Processing thermoplastic composites

Table 24.18 Effect of fillers on processing properties

Processing property Effect offiller

Viscosity Increases
Melt flow Decreases
Compounding Machine dependent, usually complicates the process
Processing temperature Increases
Temperature No effect or cools faster
Injection pressure Increases
Flow in injection mold Mold dependent but usually decreases
Injection mold shrinkage Decreases
Injection mold cycle time No effect
Thermoforming sheet sag Decreases
Thermoforming pressure Increases
Thermoforming depth of draw Decreases
Thermoforming part surface quality Decreases
Extrudate surface quality Decreases
Extrusion die pressure Increases
Melt extrudate swell Decreases, sometimes dramatically
Melt fracture severity Decreases
Foam cell size No effect to decreases

24.8 SUMMARY REFERENCES

Filled polymers tend to process in manners 1. Progelhof, RC. and Throne, J.L., Polymer
Engineering Principles: Properties, Processes, and
similar to their neat polymer counterparts. The Tests for Design, Munich: Carl Hanser, 1993,
general trend is to increase the melt viscosity Figure 5.1.
of the polymer at low shear rates and to 2. Progelhof, RC. and Throne, J.L., Polymer
decrease the relative effect of viscoelasticity of Engineering Principles: Praperties, Processes, and
the polymer. Table 24.18 shows this for most of Tests for Design, Munich: Carl Hanser, 1993,
the processes described above. Short-fiber Table 1.2.
reinforcements show more local flow orienta- 3. Progelhof, RC. and Throne, J.L., Polymer
Engineering Principles: Praperties, Processes, and
tion than fillers but, by and large, process Tests for Design, Munich: Carl Hanser, 1993,
parameters are not dramatically influenced by Table 1.4.
their presence. Long-fiber reinforcements on 4. Progelhof, RC. and Throne, J.L., Polymer
the other hand show substantial local flow ori- Engineering Principles: Properties, Processes, and
entation in all polymer melt processes. This Tests for Design, Munich: Carl Hanser, 1993,
orientation is controlled to a limited extent by Table 1.3.
the design of the dies and molds. Table 24.18 5. Charrier, J.-M, Polymeric Materials and
Processing. Plastics, Elastomers and Composites,
summarizes many of the important processing Munich: Carl Hanser, 1991.
variables for filled and discontinuous fiber- 6. Progelhof, RC. and Throne, J.L., Polymer
reinforced polymers. Flow of continuous Engineering Principles: Praperties, Processes, and
fiber-reinforced polymers is restricted to local Tests for Design, Munich: Carl Hanser, 1993,
squeezing flow around the fiber bundles and Figure 5.2.
so thermoforming, compression molding, 7. Progelhof, RC. and Throne, J.L., Polymer
stamping and diaphragm forming are the Engineering Principles: Praperties, Processes, and
Tests for Design, Munich: Carl Hanser, 1993,
major ways of fOrming these thermoplastic Table 5.3.
composites into useful products.
 References 553

8. Progelhof, Re. and Throne, J.L., Polymer 27. Shenoy, AY., Saini, D.R, and Nadkarni, V.M,
Engineering Principles: Properties, Processes, and Rheograms of filled polymer melts from melt-
Tests for Design, Munich: Carl Hanser, 1993, flow index, Polym. Comp., 1983,4,53-63.
Table 5.5. 28. Han, CD., Multiphase Flow in Polymer Processing,
9. Progelhof, Re. and Throne, J.L., Polymer London: Academic Press, 1981, p. 113.
Engineering Principles: Properties, Processes, and 29. Tanaka, H. and White, J.1., Experimental inves-
Tests for Design, Munich: Carl Hanser, 1993, tigations of shear and elongational flow
Chapter 5. properties of polystyrene melts reinforced with
10. Throne, J.L., Plastics Process Engineering, New calcium carbonate, titanium dioxide, and car-
York: Marcel Dekker, 1979. bon black, Polym. Eng. Sci., 1980,20,949-956.
11. Middleman, S., Fundamentals of Polymer 30. Chan, Y., White, J.L. and Oyanagi, Y., A funda-
Processing, New York: McGraw-Hill Book Co., mental study of the rheological properties of
1977. glass-fiber-reinforced polyethylene and poly-
12. Tadmor, Z. and Gogos, e.G., Principles of styrene melts, J. Rheol., 1978, 22, 507-524.
Polymer Processing, New York: Wiley- 31. Kamal, M.R. and Mutel, A.T., The prediction of
Interscience, 1979. flow and orientation behavior of short fiber
13. Crawford, RJ., Plastics Engineering, 2nd Edn, reinforced melts in simple flow systems, Polym.
Oxford: Pergamon Press, 1987. Compos., 1989, 10, 337-343.
14. McCrum, N.G., Buckley, e.P. and Bucknall, 32. Progelhof, Re. and Throne, J.L., Polymer
e.B., Principles of Polymer Engineering, Oxford: Engineering Principles: Properties, Processes, and
Oxford University Press, 1988. Tests for Design, Munich: Carl Hanser, 1993, p.
15. Bird, RB., Armstrong, Re. and Hassager, 0., 165.
Dynamics of Polymeric Liquids. Volume 1: Fluid 33. Progelhof, Re. and Throne, J.L., Polymer
Mechanics, New York: John Wiley & Sons, 1977. Engineering Principles: Properties, Processes, and
16. Larson, RG., Constitutive Equations for Polymer Tests for Design, Munich: Carl Hanser, 1993, p.
Melts and Solutions, Boston: Butterworths, 1988. 166.
17. Han, e.D., Rheology in Polymer Processing, New 34. Progelhof, Re. and Throne, J.L., Polymer
York: Academic Press, 1976. Engineering Principles: Properties, Processes, and
18. Ferry, J.D., Viscoelastic Properties of Polymers, Tests for Design, Munich: Carl Hanser, 1993, p.
New York: John Wiley & Sons, Inc., 1961. 161.
19. Lenk, RS., Plastics Rheology: Mechanical 35. Throne, J.L., Thermoplastic Foams, New York:
Behaviour of Solid and Liquid Polymers, New Chapman & Hall, 1994, Fig. 9.108.
York: Wiley Interscience, 1968. 36. Progelhof, Re. and Throne, J.L., Polymer
20. Middleman, S., The Flow of High Polymers: Engineering Principles: Properties, Processes, and
Continuum and Molecular Rheology, New York: Tests for Design, Munich: Carl Hanser, 1993, Fig.
Wiley Interscience, 1968. 5.3.
21. Brydson, J., Flow Properties of Polymer Melts, 37. Rauwendaal, e., Polymer Extrusion, Munich:
New York: Van Nostrand Reinhold, 1970. Carl Hanser, 1986.
22. Han, e.D., MuItiphase Flow in Polymer 38. White, J.L., Twin Screw Extrusion: Technology and
Processing, Academic Press, 1981, Fig. 3.14, p. Principles, Munich: Carl Hanser, 1990.
101. 39. Michaeli, W., Extrusion Dies: Technology and
23. Herschel, W.H. and Bulkley, R, Measurement Principles, Munich: Carl Hanser, 1984.
of consistency as applied to rubber-benzene 40. Throne, J.L., Thermoplastic Foams, Hinckley,
solutions, Proc. Amer. Soc. Test. Mater., 1926,261 OH: Sherwood Publishing, 1996, Section 5.2.
621-674. 41. White, J.L., Twin Screw Extrusion: Technology and
24. Hershel, w.H. and Bulkley, R, Kolloid-Z., 1926, Principles, Munich: Carl Hanser, 1990, Chapter
39,291-299. 7.
25. Han, e.D., Multiphase Flow in Polymer 42. White, J.L., Twin Screw Extrusion: Technology and
Processing, London: Academic Press, 1981, Fig. Principles, Munich: Carl Hanser, 1990, p. 10.
3.22, p. 106. 43. Rauwendaal, Polymer Extrusion, Munich: Carl
26. Han, C.D., Multiphase Flow in Polymer Hanser, 1986, p. 462.
Processing, London: Academic Press, 1981, Fig. 44. Throne, J.L., Thermoplastic Foams, Hinckley,
3.24, p. 108. OH: Sherwood Publishing, 1996, Table 5.4.
554 Processing thermoplastic composites

45. White, J.L., Twin Screw Extrusion: Technology 1994, p. 55.

and Principles, Munich: Carl Hanser, 1990, 61. Bay, RS. and Tucker, c.L., Fiber orientation in
Chapter 11. simple injection moldings. Part I: theory and
46. Rauwendaal, Polymer Extrusion, Munich: Carl numerical methods, Polym. Compos., 1992, 13,
Hanser, 1986, p. 403. 317-331.
47. Wall, D., The processing of fiber reinforced 62. Givler, RC., Crochet, M.J. and Pipes, RB.,
thermoplastics using co-rotating twin screw Numerical prediction of fiber orientation in
extruders, Polym. Compos., 1989, 10, 98-102, dilute suspensions, J. Compos. Mater., 1983, 17,
Fig. 2. 330-343.
48. Goettler, L.A., Mechanical property enhance- 63. Kenig, 5., Fiber orientation development in
ment in short-fiber composites through the molding of polymer composites, Polym.
control of fiber orientation during fabrication, Compos., 1986, 7, 50-55, Fig. 4.
Polym. Compos., 1984,5,60-71. 64. Kenig, 5., Fiber orientation development in
49. Michaeli, W., Extrusion Dies: Design and molding of polymer composites, Polym.
Engineering Computations, Munich: Carl Compos., 1986, 7, 50-55, Fig. 5.
Hanser, 1984, 110-115. 65. Kenig, 5., Fiber orientation development in
50. Han, C.D., Multiphase Flow in Polymer molding of polymer composites, Polym.
Processing, London: Academic Press, 1981, p. Compos., 1986, 7, 50-55, Fig. 1.
109, Fig. 3.26. 66. Darlington, MA and Smith, A.c., Some fea-
51. Goettler, L.A., Mechanical property enhance- tures of the injection molding of short fiber
ment in short-fiber composites through the reinforced thermoplastics in center sprue-gated
control of fiber orientation during fabrication, cavities, Polym. Compos., 1987, 8, 16-21, Fig. 4A.
Polym. Compos., 1984, 5, 137-154, Fig. 12. 67. Bay, RS. and Tucker ill, c.L., Fiber orientation
52. Goettler, L.A., The extrusion and performance in simple injection moldings. Part II: experi-
of plasticized poly(vinyl chloride) hose rein- mental results, Polym. Compos., 1992, 13,
forced with short cellulose fibers, Polym. 332-341.
Compos., 1983,4, 249-255, Fig. 1. 68. Advani, S.G. and Tucker ill, c.L., The use of
53. Rubin, 1.1., Injection Molding: Theory and Practice, tensors to describe and predict fiber orientation
New York: Wiley Interscience, 1972, p.2. in short fiber composites, ,. Rheol., 1987, 31,
54. Progelhof, RC. and Throne, J.L., Polymer 751-784.
Engineering Principles: Properties, Processes, and 69. Advani, S.G. and Tucker ill, c.L., Closure
Tests for Design, Munich: Carl Hanser, 1993, approximations for three-dimensional struc-
ChapterS. ture tensors, ,. Rheol., 1990, 34, 367-386.
55. Progelhof, R.c. and Throne, J.L., Polymer 70. Malloy, R.A., Plastic Part Design for Injection
Engineering Principles: Properties, Processes, and Molding: An Introduction, Munich: Carl Hanser,
Tests for Design, Munich: Carl Hanser, 1993, 1994, Fig. 2.46.
Fig. 5.19. 71. Malloy, RA., Plastic Part Design for Injection
56. Progelhof, R.c. and Throne, J.L., Polymer Molding: An Introduction, Munich: Carl Hanser,
Engineering Principles: Properties, Processes, and 1994, p. 55.
Tests for Design, Munich: Carl Hanser, 1993, Fig. 72. Malloy, RA., Plastic Part Design for Injection
5.21. Molding: An Introduction, Munich: Carl Hanser,
57. Throne, J.L., Plastics Process Engineering, New 1994, p. 54.
York: Marcel Dekker, 1979, Chapter 14, 73. Brahmbhatt, S. and Malloy, R, An evaluation of
699-800. weld line strengths for long glass fiber rein-
58. Progelhof, RC. and Throne, J.L., Polymer forced styrene/maleic anhydride copolymer,
Engineering Principles: Properties, Processes, SPE ANTEC Tech. Papers, 1992, 38, 2563-2567.
and Tests for Design, Munich: Carl Hanser, 74. Progelhof, RC. and Throne, J.L., Polymer
1993, p. 97. Engineering Principles: Properties, Processes, and
59. Malloy, RA., Plastic Part Design for Injection Tests for Design, Munich: Carl Hanser, 1993.
Molding: An Introduction, Munich: Carl Hanser 75. Throne, J.L., Thermoforming, Munich: Carl
1994, Fig. 2.77. Hanser, 1987.
60. Malloy, RA., Plastic Part Design for Injection 76. Throne, J.L., Thermoforming, Munich: Carl
Molding: An Introduction, Munich: Carl Hanser, Hanser, 1987, pp. 34-35.
 References 555

77. Anon., 1900 ultrahigh molecular weight poly- 93. Q'Bradaigh, CM. and Pipes, RB., Issues in
mer compression molding techniques, Bulletin diaphragm forming of continuous fiber rein-
HPE-102, Himont U.S.A., Inc., Wtlmington DE, forced thermoplastic composites, Polym.
undated. Compos., 1991, 12, 24&-256.
78. Narkis, M. and Rosenzweig, N., Eds., Polymer 94. Q'Bradaigh, CM. and Pipes, RB., Issues in
Powder Technology, Chichester: John Wiley & diaphragm forming of continuous fiber rein-
Sons, 1994. forced thermoplastic composites, Polym.
79. Fong, L., XU, J. and Lee, L.J., Preforming analy- Compos., 1991, 12, 24&-256, Fig. 2a-2b.
sis of thermoformable glass fiber mats - 95. Bigg, D.M., Hiscock, D.E, Preston, J.R. and
deformation modes and reinforcement charac- Bradbury, E.J., Thermoplastic matrix sheet
terization, Polym. Compos., 1995, 15, 134-146. composites, Polym. Compos., 1988, 9, 222-228,
80. Throne, J.L., Thermoforming, Munich: Carl Fig. 6.
Hanser, 1987, Chapter 4. 96. Progelhof, RC and Throne, T.L., Polymer
81. Domininghaus, H., Plastics for Engineers: Engineering Principles: Properties, Processes, and
Materials, Properties, Applications, Munich: Carl Tests for Design, Munich: Carl Hanser, 1993, pp.
Hanser, 1993, Fig. 523, p. 564. 471-475.
82. Domininghaus, H., Plastics for Engineers: 97. Shutov, E, Integral/Structural Polymer Foams,
Materials, Properties, Applications, Munich: Carl Berlin: Springer Verlag, 1986, Chapter 10.
Hanser, 1993, Fig. 457, p. 506. 98. Peters, D.L., Kowalski, RC and Hughes, J.K.,
83. Domininghaus, H., Plastics for Engineers: Blow molded reinforced HDPE for structural
Materials, Properties, Applications, Munich: Carl applications, SAE Tech. Paper No. 830077,
Hanser, 1993, Fig. 454, p. 506. 1983.
84. Domininghaus, H., Plastics for Engineers: 99. Rosato, D.V. and Rosato, D.V., Eds., Blow
Materials, Properties, Applications, Munich: Carl Molding Handbook: Technology, Performance,
Hanser, 1993, Fig. 457. Markets, Economics. The Complete Blow Molding
85. Bigg, D.M., Hiscock, D.E, Preston, J.R and Operation, Munich: Carl Hanser, 1989, p. 481.
Bradbury, E.J., Thermoplastic Matrix Sheet 100. Rathgeber, J., Evolution of an imaginative tech-
Composites, Polym. Compos., 1988, 9, 222-228. nology - double wall blow molding, in Blow
86. Wolpert, V.M., Synthetic Polymers and the Paper Molding Handbook: Technology, Performance,
Industry, San Francisco: Miller Freeman, 1977. Markets, Economics. The Complete Blow Molding
87. d' A. Clark, T., Pulp Technology and Treatment for Operation, Munich: Carl Hanser, 1989 (D.V.
Paper, San Francisco: Miller Freeman, 1978. Rosato and D.V. Rosato, Eds), pp. 828-829.
88. Cakmak, M. and Dutta, A., Instrumented ther- 101. Han, CD., Multiphase Flow in Polymer
moforming of advanced thermoplastic Processing, London: Academic Press, 1981,
composites. nl: Relative performance of vari- Figure 3.26, p. 109.
ous prepregs in forming double curvature 102. Minagawa, N. and White, J.L., The influence of
parts, Polym. Compos., 1991, 12, 354-369, Fig. 3. titanium dioxide on the rheological and extru-
89. Throne, J.L., Thermoforming, Munich: Carl sion properties of polymer melts, J. Appl.
Hanser, 1987, p. 28, Fig. 1.22 Polym. Sci., 1976, 20, 501-523.
90. Berins, M.L., SPI Plastics Engineering Handbook 103. Throne, J.L., Rotational molding, in Polymer
of the Society of the Plastics Industry, Inc., 5th Powder Technology, (M. Narkis and N.
Edn, New York: Van Nostrand Reinhold, 1991, Rosenzweig, Eds.), Chichester: John Wiley,
Figure 13-11b, p. 392. 1994, Fig. 3.
91. Dutta, A, Niemeyer, M. and Cakmak, M., 104. Ramazzoti, D.J., Rotational molding, in Plastics
Thermoforming of advanced thermoplastic Engineering Handbook of the Society of the Plastics
composites. I: Single curvature parts, Polym. Industry, Inc., (J. Frados, Ed.), 4th Edn, New
Compos., 1991, 12, 257-272. York: Van Nostrand Reinhold, 1976, p. 353.
92. Cakmak, M. and Dutta, A., Instrumented ther- 105. Crawford, RJ., Rotational moulding of plastics,
moforming of advanced thermoplastic Prog. Rubb. Plast. Technol., 1990, 6(1}, 1-29.
composites. II: Dynamics of double curvature
part formation and structure development
from PEEK/carbon fiber prepreg tapes, Polym.
Compos., 1991, 12,338-353.
TOOLING FOR COMPOSITES 25
Jerry L. Cadden and Paul F. Sadesky

25.1 INTRODUCTION lized for tooling, but no one material solves all
of the problems, particularly when factors
The manufacture of composite details and
such as cost, longevity and tolerances are con-
assemblies requires that some kind of accurate
sidered. The primary objective of any tool for
repeatable tool surface be provided, indexed
composite fabrication is to make an accurate
to an engineering database or reference model
repeatable part, within the confines of the
and be capable of withstanding repeated expo-
process parameters defined by the composite
sures to the cure cycle environment of high
material supplier and the detail performance
temperatures and pressures. Once the specific
characteristics of the end use customer. Design
manufacturing process has been selected (i.e.
of the initial tool becomes the most pressing
vacuum bag lay-up or resin transfer molding),
initial issue of tooling for composites.
decisions regarding tolerances, heat up rates,
coefficients of thermal expansion, tool
longevity etc. influence the construction of the 25.2 TOOL DESIGN BASICS
tool from an engineering design and material
selection standpoint. Individual composite 25.2.1 COEFFICIENT OF THERMAL EXPANSION
parts or details will require a variety of sup-
port tooling beyond the initial cure tool, such One of the most critical parameters in the
as master model reference patterns, trim or design of tooling for composites is the differ-
router tools, precision hole location drill tools, ence between the coefficient of thermal
assembly fixtures, ply locating templates and expansion (eTE) of the tool being designed
other shop aids. Planning must ensure that a and of the composite detail being fabricated.
point of reference is established that will con- During the cure cycle of the composite lay-up
trol all tooling in anyone part family. This will on a tool, the lay-up expands during the heat
guarantee that critical dimensional tolerances up cycle. The specific rate of expansion is
are maintained within the relationship directly related to the type and combination
between different tools supporting the fabrica- of resin or matrix and fibers or reinforcement
tion of one composite detail or assembly. In used. The tool will also expand and contract
addition, coordination between various com- at a specific rate determined by the material
posite details will ensure that and construction techniques utilized. If the
interchangeability or replacement is main- eTE values for the laminate and the tool dif-
tained throughout entire structures. There is fer significantly, stresses may result in the
an extensive list of materials which can be uti- laminate causing the occurrence of dimen-
sional, strength and part stability problems.
The greater the difference between the eTE of
Handbook of Composites. Edited by S.T. Peters. Published the composite detail and the tool, the more
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 pronounced the effect will be.
 Tool design basics 557

One of the effects that occurs as a function of the fabrication of composite details. One
these dimensional differences is called spring- method is careful selection of the appropriate
back. Composite details, when cured, hold the tooling material. Each of the commonly used
specific molded shape, as defined by the tool, tooling materials available has a specific CTE
as a result of the cured combination of resin value (Table 25.1). When selecting the appro-
and reinforcement. The springback, or more priate tooling material based on the issue of
accurately defined as a warpage condition, CTE compatibility only, first determine the
occurs when the composite detail is cured into eTE of the composite detail being fabricated.
a tool, that at a specific temperature has one The specific expansion rate will be determined
definite dimensional tolerance and then upon by the combination of the resin and reinforce-
cooling to ambient temperature, contracts to its ment utilized along with the particular fiber
original ambient dimensions. The composite orientation that is incorporated into the lami-
detail, based on the resin chemistry, cures dur- nate. For example, the CTE of the more
ing the specific period when tool expansion is common unidirectional carbon fibers used in
at it greatest. Warpage occurs when stresses are most composite epoxy laminates is approx. 4.5
induced to the composite as the tool begins to x 10-<> fOe (2.5 x 10-<> fOF ). The strength of the
return to the ambient dimensions, because the reinforcement material lies along the direction
composite detail is being forced to conform to of the fiber, not perpendicular to it. If a lami-
the new dimensional range against the dynam- nate is balanced, quasi-isotropic, with
ics of the state it reached during cure. This individual laminate layers or plies equally dis-
condition will increasingly become greater as tributing loads throughout the laminate, the
the temperature difference between ambient CTE of the laminate will be consistently equal
and cure temperature increases and the dimen- in all directions. If one direction is dominated
sional size of the tool increases. A common by more material plies than any other direc-
method of minimizing the effects of springback tion, the eTE value will vary, with the
or warpage of quasi-isotropic composite dominant direction having the lower CTE.
details during and after cure cycling is to deter- Once the CTE of the laminate is determined,
mine the CTE of the composite part being using the appropriate chart select the tooling
fabricated and the CTE of the tooling material material with the closest match to the laminate
selected. During the design of the tooling, care- value.
fully match as closely as possible the The other method of accounting for eTE
appropriate tooling material CTE to that of the variations between the detail being fabricated
composite detail. and the associate tool, besides material selec-
Other conditions that might lead to a tion, is the use of shrink factors in the
warpage of the laminate include an unbal- calculation of dimensions prior to tool fabri-
anced laminate orientation where the number cation. If requirements dictate that when
of layers or plies of material are more domi- fabricating the tool, a material with an incom-
nant in one direction than another. This patible CTE to the detail being manufactured
condition is separate from any function of the must be used, steps may be taken to mini-
tool and must be considered during the design mize the effect of this variation. During the
of composite detail. design phase of the tool, accurate estimates of
the actual tool size at its greatest expansion
point or at the highest temperature during
25.2.2 USING CTE IN THE DESIGN OF
the cure cycle must be made. The percentage
TOOLING FOR COMPOSITE
difference between this calculation and the
Two methods are commonly used to minimize dimensions at ambient must be applied to the
the effect of CTE when designing tooling for base design as a 'shrink factor' reducing the
558 Tooling for composites

Table 25.1 Properties of commonly used tooling materials

Material Coefficient of thermal expansion, Density, Thermal conductivity,

x l(J-6tC g/cm3 cal cm/s cm2 °c
(x l(J-6rF) (lblJf3) (BTU in/h Jt2 oF)
Aluminum 23.7 2.7 0.48
(13.6) (170) (1400)
Eleetroformed nickel 13.3 8.6 0.19
(7.4) (540) (564)
Carbon/ epoxy prepreg 4.5 1.4-1.5 16 x 10-4
(2.5) (87-94) (4.6)
Glass/epoxy 13.1 1.6-2.0 6 x 10-4
(7.2) (100-125) (1.7)
Invar 36 3.4 8.0 0.016
(1.9) (504) (48)
Invar42 5.4 8.1 0.016
(3.0) (507) (48)
Monolithic graphite 2.16 1.76 0.28
(1.2) (110) (840)
High carbon cast steel 9.7 7.3 0.12
(5.4) (456) (360)
Steel 12-14 7.8 0.10
6.7-7.8 (490) (300)
VVood(nruahogany) 22 0.7 0.90
(12) (44) 2600
Urethane board stock 50 0.77 n.a.
(28) (48)

size of the tool by that same percentage. Por sions. Since the composite detail was cured
example, if a specific tool at 17J70C (3500 P) has while the tool was at the larger dimension, if
a growth factor equaling 1.27 mm (0.050 in) of the detail is confined to the tool surface or
growth over ambient dimensions, this same restricted from movement due to complexity
factor would be applied as a reduction of the in the tool surface, the composite detail could
overall tool dimension while at room temper- become entrapped, resulting in dimensional
ature. This method of applying a shrink abnormality in the laminate and possible
factor to allow for variations of CTE between damage to the composite detail or the tooling
tools and details must be approached cau- surface. A variety of other factors should also
tiously when complex shaped surfaces are be considered such as tool durability, tool
involved. This is due to the potential for the usage rates, thermal conductivity of the mate-
composite details to become entrapped rial and machinability or fabrication cost.
within the geometry of the tool as the tool Each of these factors must be weighed indi-
returns to the ambient temperature dimen- vidually before final selection is made.
 Tool design basics 559

25.2.3 MATERIAL CHOICES IN THE DESIGN OF machining the final tool surface directly from
TOOLING FOR COMPOSITE a computer model can be accomplished using
a material capable of the final cure tempera-
The use of CTE is not the sole determinant for
tures. Monolithic graphite and a variety of
selection of a tooling material. For example, as
epoxy and polyester tooling boards would
defined in Table 25.1 monolithic graphite
allow this with monolithic graphite offering
which as a tooling material has a CTE value of
the highest quality and lowest CTE at a com-
2.2 x 10-6fOC (1.2 x 1()-6 fOF) has good machin-
parable cost to the board stocks available.
ability at a relatively low cost compared to
These costs would be similar to laminated
other tooling materials, but exhibits poor
tooling without the intermediate and time
durability when utilized in the high usage of
consuming steps necessary to complete the
production environments. Invar® 36, however,
laminated tool. However, in a high usage pro-
displays a similarly low CTE value, displays a
duction program, these materials can be
much higher level of damage resistance but
damaged more easily than the composite lam-
has poor machinability and also has a high
inate or metallic tools currently in use.
acquisition cost. Careful selection of the
Continued advances by the suppliers of com-
appropriate material for tool use must include
posite tooling prepregs have drastically
review of the following criteria:
increased the ability of composite laminate
• anticipated tool usage (expected life of tool); tools to support a high number of cycles at
• cost available for tool fabrication; cure temperature.
• materials available for tool construction; Another factor directly affecting the
• available methods of tool manufacturing; longevity of tooling fabricated from composite
• level of dimensional tolerances required materials is the effect of proper employee
from composite detail. training in the care required of such tooling.
Improper handling techniques will drastically
shorten the expected life of composite tooling.
Anticipated tool usage
Employee caused damage such as cutting on
The life expectancy of any tool fabricated for the surface of the tool, using sharp instru-
the lay-up and cure of composite details is ments to facilitate the removal of details after
dependent on a variety of factors. Material cure and improper application of release
selection, shop handling procedures and cure agents are the most readily identified causes of
cycle times all affect the ability of the tool to shop induced damage. Special handling pro-
withstand long usage. Certain materials dis- cedures and employee indoctrination can
play characteristics that allow longer tool life, minimize this type of damage. While all tool-
however the advantage and disadvantage of ing, both metallic and non-metallic, is
each material must be analyzed prior to selec- susceptible to damage, tooling fabricated from
tion. In addition, each of the tooling materials composite materials is especially sensitive to
presently used is sensitive to damage specific surface damage caused by employee careless-
to that material. If short term usage is antici- ness.
pated, temporary tooling such as wet lay-up
epoxy or polyester tooling (dependent on
Cost available for tool fabrication
detail cure temperature) may be acceptable,
however master models and intermediate Cost of tool fabrication is difficult to quantify
transfer tooling would be necessary to main- since material procurement and labor cost
tain the correct surface tolerances. To vary widely throughout the industry.
minimize the costs associated with master However, comparisons can be made between
models and intermediate transfer tooling, the different tooling materials and methods
560 Tooling for composites

of manufacturing to assist the user in deter- up fiberglass or graphite tooling. Equipment

mining the optimum approach based on a obligations can be as minimal as a calibrated
cost/performance comparison. Comparisons surface table and measuring instruments for
must include all necessary steps to deliver a the accurate setup and weighing of mixed
completed tool family. There is a temptation resins and a general knowledge of the systems
to compare tooling cost for just the lay-up involved. Use of composite prepreg materials
mold in vacuum bag processing and ignore requires additional employee skills levels plus
the procedures and tooling necessary to the additional equipment expenditures of an
obtain that final surface. In direct compari- oven or autoclave to cure the materials accord-
son, a machined lay-up tool will appear to be ing to manufacturers' instructions. Increasingly
substantially more expensive when com- stringent facility requirements involving the
pared to a lay-up tool fabricated from a installation of a controlled environment, such
composite prepreg. However, when the asso- as a 'clean room', have been recommended by
ciated costs of reference patterns, additional material suppliers. As the complexity of the
splashes to obtain the correct surface and tooling rises, so does the requirement for ade-
possible autoclave or oven processing time, quate employee training and engineering
are factored in, the tool machined directly support.
from numerical control (NC) data, which will To support the fabrication of complex,
eliminate the need for all the intermediate highly accurate machined tooling, machine
steps mentioned, will become more compara- tools with sufficient work envelopes are
ble in price. required in the correct axis of motion to sup-
port the complexity of tooling surfaces.
Three-axis machine tools will support basic
Materials available for tool construction
simple contours, with 4, 5 and 6 axis machines
Because of the advent of computer-aided capable of machining the much more complex
design and manufacturing systems compound contour surfaces. A sufficient com-
(CAD/CAM) to support tool fabrication, there puter-aided design and manufacturing
are more materials now available. Baseline (CAD/CAM) system is required to supply
methods such as plaster master models and software commands in order to drive machine
plastic-faced plaster transfer tooling will tool directional and cutting speeds. In addi-
always have a specific application within tool- tion, if board stock materials are chosen for
ing, but as the use of CAD/CAM increases reference patterns, the skill and equipment
beyond aerospace into all regimens (e.g. sport- necessary to assemble, bond and seal the
ing goods, medical and transportation), such material after machining are required.
tooling practices that depend more on the skill
level of shop personnel than on the accuracy of
Level of dimensional tolerances required
a machine tool are rapidly being replaced.
from composite detail
Based on the type of manufacturing method
Available methods of tool fabrication
and the type of material selected, different lev-
Methods of manufacturing vary, dependent on els of dimensional tolerances are possible.
equipment and personnel resources available. Initially, the designer determines the level of
Plaster type master models, intermediate trans- dimensional tolerance for the composite detail
fer tooling and wet lay-up type molds take being fabricated. Compliance to this tolerance
minimum facility requirements. Basic shop skill is critical in meeting structural demands and
levels require a familiarity with resins and conformance to any form, fit or function
geometry in order to support simplistic wet lay- requirements.
 Tool design basics 561

In the early days of advanced composites, result in dimensional changes occurring dur-
applications were limited to aerospace, which ing this process.
invoked strict dimensional requirements for The highest level of tolerances available are
both tooling and detail parts. These require- obtained by machining the tooling surface
ments continue within aerospace today. With directly from the computer model. Certain
the expansion of composite usage into other master model materials such as plaster and
areas, such as sporting goods and automotive some board stock materials have a limited
applications, the range of acceptable dimen- temperature exposure level which inhibits the
sional variations has increased but visual ability to pull composite laminates directly off
requirements are much more stringent than their surface. Intermediate tooling must be
before. The variation of CTE between differ- fabricated either to obtain the correct surface
ent materials for tooling has a major effect on level from the master or to be capable of with-
dimensional tolerances. In addition, some standing elevated temperatures within an
materials display sensitivity to environmental autoclave or oven, above that of the original
conditions that have an adverse effect on model. By machining directly from the engi-
dimensional stability. Tooling, such as refer- neering database, the need for intermediate
ence patterns or master models manufactured surface splashes to obtain the correct surface is
from urethane board stocks or plaster models, eliminated. Each time a splash or the original
are hygroscopic and may absorb moisture surface model is duplicated, a 'stackup' or
from the atmosphere. This condition will, at a accumulation of the tolerances for each model
minimum, cause dimensional changes related is combined, resulting in a much greater range
to the level of moisture absorbed. Also, this of tolerances in the final tool. For example, if a
condition could be excaberated if the moisture plaster master is fabricated to ±O.25 mm
contamination is extensive and the model is (±O.010 in) tolerance and each of two addi-
taken to elevated temperatures. At higher tional splashes have the same tolerance range,
temperatures, the moisture will expand and prior to fabrication of the final lay-up mold,
may result in possible significant changes in the beginning tolerance range is now ±O.76
structural integrity. Urethane or epoxy board mm (±O.030 in). If a tool is machined directly
stock materials have greater resistance to from NC data, the tolerance stackup is elimi-
moisture. However, if machined and used for nated and only the range of the individual
reference patterns they are both still suscepti- machine tool applies. Machine tools, depen-
ble to contamination, which could result in dent on the condition and environment of
dimensional changes and possible failures in machining, are capable of providing ±O.128
the bond joints between the block surfaces. mm (±O.005 in) accuracy or greater.
Additional steps must be taken to protect all
models manufactured from these materials to
25.2.4 DESIGNING TOOLS FOR RESISTANCE TO
prevent these types of contamination. Sealers
FAILURE
must be applied and the items must be segre-
gated from potential sources of Because of the abusive environment experi-
contamination. Monolithic graphite offers enced by tooling during the fabrication of
advantages over these materials because it is composite details, life expectancy of tooling
inert and resists contamination from the envi- will always be short of anticipation. Repetitive
ronment. Lay-up molds manufactured from cycling from ambient to over 177°C (350°F),
both ferrous and nonferrous materials must inadequate care and handling procedures,
be protected from oxidation. Failure to main- incorrect fabrication techniques have allIed to
tain a nonoxidized lay-up surface will require a variety of defects resulting in premature
restoration of the tooling surface which could temporary or permanent failure of the tool.
562 Tooling for composites

Failure modes common to composite lay-up laminating the initial tool, apply additional
tools fabricated by both wet lay-up and plies of pregreg in the area of the bushings to
prepreg methods generally involve fiber sepa- increase support in those areas to resist move-
ration. This is due to a variation in CTE ment during part removal.
between the resin matrix and the fiber. Another possible solution to the problem of
Generally, the neat resin systems used in most delamination between layers of prepreg tooling
composite tooling systems have a CTE of 65 x is the application of glass transition tempera-
lO-6;oC (36 x 1()-6;OF) The graphite fibers used ture values (T ) to extend the life expectancy of
in most prepreg tooling systems have a CTE of a tool. Most tooling resin systems are formu-
around 4.5 x 10-6;oC (2.5 X 10-6 ;OF). During lated with a T value at or slightly above the
exposure to cure cycles where temperatures .
maxunum use gtemperature 0 f teh · system.
resm
will vary from ambient to 177°C (350°F) and As a function of the resin chemistry, glass tran-
above, the difference in CTE between the sition temperatures decay or reduce with each
fibers and the resin will eventually cause dis- exposure to the cure temperature that the sys-
bonds between laminate layers resulting in tem was designed for. This decay, in
leaks internally within the tool. In addition, incremental steps will continue until well
the expansion of the resin is somewhat con- below the cure cycle temperature that the tool
trolled by the fiber reinforcement in the x and was intended to be cycled at. When this point is
y axis. Because no reinforcement exists in the z reached, the resin will begin to break down
plane linking the individual layers together, with a mechanical failure of the bond between
the difference in CTE between the resin and the resin matrix and the fiber reinforcement.
reinforcement becomes more pronounced. The solution to this problem is to use a
Failures between the individual plies increase resin system with the highest possible Tg
because of the lack of reinforcement restrain- value available. For example, if the tool is
ing the resin from the repeated expansion and intended to be cycted repeatedly at 177°C
contraction. (350°F), a T value of the resin system in the
Furthermore, when laminating layers of 220°C (425°~) range will allow more cycles. It
either prepreg or wet lay-up tooling, by cut- is common among some aerospace compa-
ting each of the plies into pieces 304-457 mm nies to now fabricate composite tooling for
(12-18 in) square, no continuous fiber path epoxy laminates from a bismaileimide or a
will pass through the tool. By discontinuing cyanate ester resin system with T values
these pathways, leaks occurring along the higher than 260°C (500°F). This all~ws the
fiber path will be minimized. In addition, inevitable decay of the T value to span a
because each layer or ply consists of sections greater difference allowinggthe life of the tool
without any continuous fiber path, stresses to be extended. The same principle may be
within the laminate will be lower, minimizing applied for any prepreg system from poly-
warpage during use. A majority of failures in esters to the higher temperature resin
composite tooling may be directly tracked to systems. One consideration in using this
leakage around tooling holes or plumbing fit- method requires the selection of a master
tings. Tooling hole fittings are exposed to model material capable of exposure to the
repeated shocks during the removal of cured elevated temperatures that the higher tem-
composite details. If steel bushings are used, perature systems require during cure. Plaster
the difference in CTE will possibly lead to and most of the board stocks available are not
cracks in the tool surface which will become capable of these higher temperatures and
potential leak paths. One solution is to install intermediate splashes or surfaces would have
Invar 36 bushings in laminated tools which are to be provided. Monolithic graphite does pro-
closer in CTE to the parent tool. Also, when vide a surface capable of exposure to higher
 Master models 563

temperatures in addition to having the lowest struck with a hard object. Carriers designed to
CTE available, allowing tools fabricated from transport the tool to and from work station
these resin systems to be taken directly off the must also function as a protective barrier to
model surface without need for intermediate prevent the tool from striking walls or beams
surfaces. within the shop environment. Support tool-
Employee-induced damage of tooling can ing, where applicable, must be designed to be
playa much greater part in the reduction of as lightweight as possible to prevent injury to
expected tool life. Correct indoctrination into the employee and damage to the tool surface
the importance of the tool to the fabrication of when handled.
accurate details must be stressed. Most of the
employee-induced damage will occur either
25.3 MASTER MODELS
during the lay-up procedures, the removal of
cured composite details from the tool surface, A master model is considered to be just that -
or preparation of the tool surface prior to the a master source identified with holes, scribe
next lay-up. During lay-up of details, damage lines, trim lines or any other feature of the
will be the result of employees using knives or part that requires duplicating to other tools.
other sharp objects during the trimming of the The master model is the physical representa-
composite material. If proper care is not tion of the design or a point of reference to
taken, the employee will not only cut the which all support tooling, both for fabrication
material but also cut into the tooling surface. and inspection, would be indexed. Because
While not as detrimental to a metallic tool, this surface will provide the reference pattern
damage of this kind may be catastrophic to a for all subsequent operations beyond initial
tool fabricated from a composite material. The fabrication, such as assembly fixtures index-
cut will allow a breach in the vacuum ing a variety of details from different
integrity in addition to allowing resin to pen- locations, extreme care must always be taken
etrate beyond the surface of the tool. Also, in protection of the master model. Master
when laminates are removed from tooling models may be fabricated from a variety of
after completion of the cure cycle, damage materials. Common materials include plaster,
occurs when personnel use sharp equipment machined urethane or epoxy board stock,
to force the completed detail from the tool. monolithic graphite or most ferrous and non-
The greatest care must be given when ferrous metals. Each material offers distinct
attempting to remove the detail, to prevent advantages and disadvantages. To determine
inadvertent damage if the detail fails to which material is the most feasible, the entire
cleanly release from the tool surface. Damage, tooling family philosophy must be reviewed.
not only to the tool, but also to the detail may Master models are generally stored indefi-
result. To prevent this damage from occur- nitely so that they may be referred to over the
ring, proper steps must be taken. Employees life cycle of the manufactured parts. In cases
must be indoctrinated in the proper tech- where cost and! or time schedules are impor-
niques of tool maintenance and lay-up tant, temporary models are produced and
procedures and must be provided with then destroyed once they have been used.
acceptable tooling aids to assist in the safe However, because of the hygroscopic nature
removal of cured details from the tool surface. of plaster, care must always be taken to pro-
Soft wood or plastic wedges must replace tect the master model from the environment
hammers and hard-faced chisels for detail to maintain accuracy. Adequate storage con-
removal and tool surface preparation. Tools tainers, allowing for complete protection,
must be designed with adequate laminate must be utilized throughout the life
thickness to prevent damage to the tool if expectancy of the model.
564 Tooling for composites

25.3.1 PLASTER MASTERS 25.3.2 TEMPLATE METHOD

One of the oldest methods of producing a mas- There are several techniques of building a
ter is from plaster. Plaster is made from the plaster master determined by the shape of the
mineral gypsum (CaSOJ which is finely part. If the part is not symmetrical and does
ground and calcined (dehydrated) to produce not have a constant cross section or the size is
a fine powder with uniform properties. With large, the master model is made from a series
the addition of water to form a workable of templates secured to a flat base to form a
slurry, a reaction occurs which produces heat three dimensional full scale model of the part.
and the inert gypsum on drying. Plaster is Space between the templates is relative to the
manufactured in various textures or grades degree of abruptness of the contour. For nor-
which support the level of detail required on mal gentle contours a space of 15.24-20.32 mm
the model. Coarse grades are used to build up (6-8 in) is common. Templates are usually
the master model surface and then followed made from 0.317 mm (0.125 in) thick alu-
by the fine grades which allow precise details minum to prevent corrosion. For temporary
such as trim lines or other identifications to be masters, steel is sometimes used, but, because
scribed into the surface. Depending on the of the amount of moisture used in the mixing
grade being used, plaster has a setting expan- and application of the plaster, steel templates
sion of approximately 0.080% and a thermal may rust (Fig. 25.1).
expansion in the dried state of a maximum of If electronic data is available, the templates
0.027;oC (0.0156;oP). can be NC machined or cut with a water or

...
REFERENCE LINE
REFERENCE LINE

Fig. 25.1 Skeletal structure for template plaster master.

 Master models 565

laser jet directly from the flat pattern generated steel blade, the plaster is 'faired' or swept flush
by the data. Section cuts taken at specific sta- between the templates to form a smooth accu-
tion lines from two-dimensional blueprints can rate surface. Because of the propensity of
be used to saw out a template. Except for the plaster to absorb moisture, it should be sealed
NC machining method, deburring is generally after the surface has had adequate time to cure.
required to remove spurs or sharp edges from Commercially available lacquers can be used
the templates prior to use. Holes are drilled to seal the surface and provide a suitable pro-
into the templates for threaded rod spacers and tection within the shop environment.
screen support rods. For larger models, air pas-
sages are cut into the bottom of the template to
25.3.3 FOLLOW BOARD METHOD
allow for even curing of the plaster. Once suffi-
cient templates have been prepared, bluing is A method widely used when a constant cross
applied to a flat ridged steel table and scribed section is to be built is the follow board. A flat
with an awl to denote the location of each tem- surface is required with an accurate side surface
plate. Flatness of the table is critical and should to act as a guide rail. A template of the contour
be within 0.127 mm (0.005 in). Tooling balls is prepared from a rigid 3 mm (0.125 in) mini-
which indicate the x, y and z direction are mum sheet of aluminum or steel and attached
sometimes placed on the table comers as refer- to a wooden guide support. Plaster is mixed
ence points for the system. Tooling balls can and built up on the surface to within 3 mm
vary in size but a common size is 12.7 mm (0.125 in) of the final contour. Partial drying is
(0.5 in) diameter on a 6.35 mm (0.250 in) diam- recommended before the final plaster mix is
eter x 12.7 mm (0.5 in) long pin. The pins, each applied. This will prevent shrinking and crack-
with a 'ball' on top are placed into location ing of the plaster surface which would affect
holes and optically sighted relative to the posi- accuracy. Using the template and guide sup-
tion of the each ball location. Location can also port, the plaster contour is formed by pushing
be treated relative to a position on the master the template evenly over the surface (Fig. 25.2).
such as a station line. Each template is attached
90° to the base table with angles and held to
within 0.127 mm (0.005 in) of the reference line
at the base, the face square to the base to within
0.076 mm (0.003 in) in 304.8 mm (12 in) and
within 0.127 mm (0.005 in) of the base reference
line. Threaded rods are secured with sheet nuts
on each side of the template to provide rigidity
to the template face. Wire mesh is placed
between the templates and secured to the
threaded rod with wire hooks approximately
101.6 mm (4 in) below the top surface of the
template. This is used to hold the plaster in
FOllOW BOARD IIETlIOO
place. Plaster is mixed with hemp and placed FOIl PAOOUCINQ PI..AIInR ~

against the screen to approximately 9.5 mm Fig. 25.2 Follow board method for producing plas-
(0.375 in) below the template surface. A second ter master.
layer without hemp is added to this surface to
approximately 12.7 mm (0.5 in) A sawtooth
25.3.4 SWEEP METHOD
scraper is used to build a striated surface and
allowed to dry. A final mix is made with the A third method called a sweep is best utilized
fine grade of plaster and using a flat spring when a symmetrical surface such as a cone or
566 Tooling for composites

hemispherical shape is involved. As with the operation significantly. The system, known as
follow board method, a flat surface from which Automated Tool Manufacture for Composite
a frame can be constructed of the shape to be Structures (ATMCS), is an expert system with
produced is required. For large shapes, inter- macros which dramatically speeds up the tool
mittent templates should be placed within the design process. ATMCS takes the composite
framework to allow support for the sweep and detail surface model into either IBM Catia or
provide adequate support for the sweep to fair EDS Unigraphics II and creates the tool
against. The sweep itself is usually made from required around the part model.
sheet metal 3 mm (0.125 in) minimum thick- The system, acting through a series of
ness and supported by a wooden guide or inquiries made to the tool designer, selects the
other mechanical guides that can ride the sur- optimal configuration, material, manufactur-
face of the flat surface table. Smaller shapes, of ing process and design. The design is then
course, do not require this extent of rigging. created around the part model, with signifi-
Plaster can be reinforced with saturated hemp cant savings in time. Although the system was
fibers, mixed into the slurry and applied to developed initially for the aircraft industry
form rough shapes and to form strengthening and is presently used for basic open-faced lay-
ribs on the back surfaces of casts. All master up molds, it could be expanded for many
models fabricated from plaster require, in different types of tools and processes such as
addition to sealing with commercial grade lac- resin transfer molding and injection molding
quer, suitable storage containers if the model is in other industries.
required to be stored for any period of time
outside the shop environment.
25.4 COMPOSITE TOOLS
Composite tools are usually made from epoxy
25.3.5 NC MACHINING
resin matrix and either E-glass or carbon fibers
Because of the widespread use of CAD (com- as reinforcement. Depending on the life cycle
puter aided design) systems, older methods required, tools can be made from prepreg or
which utilized two-dimensional prints to by 'wet' lay-up procedures. Prepregs generally
build master models are now used less fre- require curing within an autoclave because of
quently. With CAD systems, a great deal of the elevated pressure specified by the manu-
accuracy can be transferred into the master facturer. Because of the increased compaction
model via the NC machining operation. Table available when curing in an autoclave, tooling
25.1 lists various materials widely used today fabricated from prepregs are capable of a
for NC machined master models. From a CAD greater number of cure cycles than the wet lay-
model of the part, a tool manufacturer must up method. In addition to greater compaction,
design a tool from the surface data supplied. autoclave curing offers better control of resin
Advances with CAD/CAM systems seek to content and uniformity of reinforcement.
minimize the operator input to the system and However, for shop aids such as trim tools,
transfer design responsibility to the computer. room temperature curing epoxy systems are
One example of this technology is demon- recommended.
strated in a system developed by a
multicompany team lead by the Northrop
25.4.1 LAY-UP MOLDS
Grumman Corporation for the US Air Force
Manufacturing Technology Directorate at Lay-up molds are used to form the shape of
Wright Patterson Air Force Base. This system, the part to be produced and have the part
while not totally removing the tool designer periphery scribed on the surface as well as
from the design process, does streamline the the location of any required cross hairs and
 Composite tools 567

tooling holes. Tools can be made directly is always a good idea at a minimum of 6.2 kPa
from a NC machined master model or from a (25 in Hg). A loss of 500 Pa (2 in Hg) within 5
plastic faced plaster splash taken from a mas- min with the pump nonoperating is acceptable.
ter model not capable of elevated Apply masking tape around the tool periphery
temperatures and pressures. The choice of for later application of the sealant tape. It is
glass or carbon fiber / epoxy for the mold is absolutely necessary, regardless of prior his-
generally governed by the complexity and tory of the master surface, that it can be
CTE of the part to be fabricated. Lay-up released with a suitable hard wax or other
molds must be capable of maintaining a vac- release agent. The prepreg manufacturer may
uum tight environment while being subjected recommend a specific release agent for his
to high temperatures and pressures. prepreg system and it is advisable to follow
those instructions due to the possibilities of
chemical reaction occurring between the resin
25.4.2 PREPREG METHOD
system of the tool and the release agent used
There are a considerable number of prepregs (Table 25.2). After the cleaning and releasing
available as epoxy 'B' staged glass or carbon processes have been completed, release coated
reinforced cloth. Prepregs can be obtained in tooling pins should be placed into the holes of
rolls or as precut squares or rectangles. The the master. These are generally index and
weave style can vary depending on the locating holes that have bushings and are used
amount of drape to be encountered but gener- to position or align one tool to another, or to a
ally plain or satin weaves are readily available. production part. Bushings can be installed dur-
The resins are tailored for tack, out time and ing lay-up of the tool or potted in after final
glass transition temperatures at a minimum cure of the tool.
and are around 40% by volume of the prepreg. In general, a face or gel coat layer is not
(Tooling prepreg manufacturers have very used by most manufacturers today. The reason
detailed procedures that they recommend for is associated with the difference in CTE of a
their specific system. These comments are not neat resin on the face and the CTE of the
meant to supersede the recommendation of a underlying reinforced prepreg which over the
manufacturer, but rather to place emphasis on life span of a tool can cause cracking and craz-
important steps that should not be overlooked ing of the face and subsequent loss of vacuum
for tool fabrication.) integrity. If a gel coat is used it should be of
Within the last several years, an innovation minimum thickness to minimize these effects
for tooling prepregs is the capability for low over time.
temperature curing 61°C (145°P) in an auto- A lightweight (style 7781) cloth is the first
clave, followed by a free standing post cure at layer applied to prevent mark through to the
177°C (350 0 P) after removal of the tool from surface from heavier cloth. Carefully lay each
the master. This has allowed the use of plastic ply onto the surface and work out wrinkles or
faced plaster and urethane based tooling air bubbles and maintain the warp direction of
boards for direct lay-up of composite tools. each ply in the 0° direction. An overlap
Monolithic graphite with a low CTE and capa- between the plies should be 3-6 mm
bility to withstand 315°C (600 0 P) under (O.125-{).250 in). Some manufacturers recom-
autoclave pressures can be a good choice. mend a debulking step at this point to ensure
The first step prior to prepreg application on no air entrapment at the interface and a
the master surface is to ensure that the prepreg smooth surface on the tool. Debulking is
and the master surface are absolutely clean and accomplished by application of a peel ply net
free of debris and that the surface is smooth to the edge of the laminate and working out
and without pin holes. A quick vacuum check wrinkles and air bubbles. A resin dam (sealant
568 Tooling for composites

Table 25.2 Cures for common mold release problems

Problem Cause Solution

Non-adherance of paint or Build up of mold release, layers Use manufacturer's
adhesive to part of incompatible mold releases, or recommendations for layers and
too much release in formulation cure schedules. Avoid adding
incompatible layers such as
silicone, wax and flurocarbon.
Check with manufacturer for
possible revision or custom
formulation to allow multiple
releases and adequate paint
adhesion
Poor release with small particles Porosity in mold Thoroughly clean mold with an
of visible on mold surface appropriate solvent and then
add a mold sealer before the
release coat
No paint or adhesive sticks to Check for the presence of silicones Remove silicones from plant
part in the area. Silicone mold releases where painting or adhesive
can be transferred over long bonding is performed
distances
Multiple release not possible Poor release in some areas, Apply one or two additional
particularly in severe contours coats of release to severe contour
areas
Failure to obtain any release even Inadequate cleaning of the mold Strip out the part and thoroughly
though release was applied before application of release has clean the mold, then apply and
interfered with the ability of the fully cure release agent. Also
release to bond to the mold or the check shelf life of release
release has been improperly
cured. Shelf life of release may
have expired
Generally poor release with Mold surface not properly cleaned Remove all release and
patches of white transferring to which results in poor bonding of thoroughly clean mold with an
the part release to mold appropriate solvent before
reapplication of release. Follow
manufacturer's
recommendations for cure cycle

tape) can be placed around the perimeter to permit a higher resin content on the tool sur-
prevent resin flow (Fig. 25.3). Next, lay-up one face. Over this layer, one ply of polyester
ply of Teflon® beyond the resin dam and attach breather cloth or 7500 style glass cloth is
to the resin dam. Using the manufacturer's applied. A nylon vacuum bag is placed over
recommendation, holes should be placed into the entire stack and a vacuum of at least
the Teflon ply to allow for resin bleed. Pre-per- 625 mm (25 in Hg) is applied for at least one
forated film can be obtained for this purpose hour. Removal of the bag, breather, separator
and provides greater control over the size and film and peel ply should be done very carefully
spacing of the holes. If only one lightweight to avoid lift up or shifting of the prepreg lay-
ply has been applied, no holes are required to ers. The orientation for each additional ply
 Composite tools 569

Vacuum Line
Vacuum Bag Sealant

Non-perforated
Film
Peel Ply

Laminae
T Release Film
Fig. 25.3 Laminate pre-bleed stackup.

should be such that a balanced system is main- after every 4-5 plies. Final build up of the lam-
tained to minimize stress build up in the inate should be at least 0.013 mm (0.375 in) or
laminate. Prepreg manufacturers will clearly whatever is recommended by the prepreg
stipulate the lay-up sequence. After the second manufacturer. Final vacuum bagging is per-
or third ply has been laid down, knurled bush- formed in the same manner as for debulking
ings should be placed onto the tooling pins with a layer of peel ply, perforated Teflon,
and pressed down to seat them. Subsequent polyester breather and vacuum bag. Prepreg
plies will be placed over the bushings to inte- manufacturer will provide a detailed heat up
grated them into the laminate. After the rate and cure temperature for their system and
seventh or eighth ply, the pins can be removed this should be carefully followed. Most sys-
so that later plies can cover the bushing com- tems can be initially cured at up to 63°C
pletely to prevent vacuum leaks. In some (145°F) and 586--689 kPa (85-100 psi) of pres-
instances a pad or build up of plies over the sure for 14 h.
bushing is recommended. An alternate After the autoclave cycle, carefully remove
method is to pot the bushings into the lami- the bag and films from the laminate to avoid
nate after the final cure. To do this a tapered lift up from the master surface. Attachment of
wax or rubber plug should be placed over the the egg crate structure (support or back-up
pin to allow space for the potting compound structure, Fig. 25.4) to the laminate is very
after the final cure. Each ply should be care- important to minimize any potential residual
fully worked into comers and radii making stresses built into the laminate or stresses from
sure all entrapped air is removed. Wrinkles the egg crate itself. Leave the laminate on the
should also be carefully worked out before master surface and attach board structure of
another ply is placed over it. If a persistent the same material to the surface of the lami-
wrinkle or air bubble can not be rolled out, nate. If the laminate is glass/epoxy, the egg
then carefully slit the pockets with a sharp crate material can be made from glass/epoxy
knife and work it down into the surface. In the or aluminum honeycomb sandwiched
excess area of the tool, thermocouples can be between glass. The point is to avoid stresses
strategically located between the plies for caused by the difference in CTE between the
recording during the autoclave run. As a rule egg crate and laminate by using similar mate-
of thumb, debulking should be performed rials.
570 Tooling for composites

3 LAYERS OF GLASS CLOTH

IMPREGNATED WITH
LAMINATINI~G~R:E~S:IN~---:;;~,,;-~
.--.<l_

AIR CIRCULATION HOLE

Fig. 25.4 Support structure details.

The egg crate should be cut to the contour of oven cure at 177°C (350 0 P) (i.e. metal or mono-
the laminate with a standoff of 3.17 mm lithic graphite) the post cure can be performed
(0.125 in). This prevents hot spots during pro- without removal of the tool from the master.
duction part curing and also mark off into the However, if the master material will not toler-
tool laminate. The egg crate should be con- ate this temperature, careful removal of the
structed so that it will lie flat on the surface. tool from the master must be done prior to the
The intersections of the board stock are held post cure.
together with cloth and resin with at least Separation of the tool from the master
three strips per junction. To ensure minimal should be done carefully to avoid damage to
stress to the laminate from the egg crate struc- the master or the tool itself. Tooling pins
ture, it is advisable to remove the structure should be removed prior to separation. Tools
from the laminate and cure the strips holding with severe contours may require plastic
the structure together at 177°C (350 0 P). wedges to be inserted around the tool periph-
After curing, the structure can be placed ery until it releases. Once the tool is separated,
back onto the laminate and 'tied' into the lam- the surface should be inspected for pinholes or
inate with at least three strips of cloth and resin roughness. Pinholes can be filled with resin
around the periphery of the egg crate. Shims and the roughness can be smoothed out with a
can be used to provide for the standoff. If the fine grit sandpaper. Edges of the tool can be
type of master used for the lay-up permits an sawed to even up the periphery and then
 Composite tools 571

sanded lightly with 220 grit sandpaper to occur in comers or the bottom of contours. Too
remove any loose fibers. Care must be taken much resin will result in cracking and crazing
that no fibers are lifted by sanding along the later in the tool life cycle. Also, ensure that all
length of the fibers. air bubbles have been worked out by repeat-
Once the tool has been cleaned up, the edly applying the brush back and forth across
required check for vacuum integrity is accom- the surface. To ensure that all air has come to
plished by placing a layer of polyester the surface while brushing, pause occasionally
breather cloth on the surface and a vacuum and allow the air to rise to the surface where it
bag over it. The acceptance criterion is gener- can be brushed out. Air that remains
ally that there be no loss greater than 500 Pa entrapped either on the tool surface or within
(2 in Hg) in 5 min at a minimum of 6.2 kPa the layers of cloth could result in possible blis-
(25 in Hg) at the start of the test. If possible, ters and delamination later during tool usage.
depending on the complexity of the tool, place The resin supplier will provide mix ratios for
the tool back onto the master and check for resin and hardener as well as pot life and gel
any warp or out of contour problems. times. Tooling cloth generally comes in rolls
and is either a satin or plain weave with the
warp direction noted. Sufficient squares or rec-
25.4.3 WET LAY-UP MElHOD
tangles can be precut from the roll prior to
The wet lay-up of composite tools can be for laminating. Sections over 609 rom (24 in)
room or elevated temperature use. The differ- become too cumbersome to work on the tool
ence is in the resin selection. Procedurally, the face, therefore smaller sections are advised.
process is the same except for the cure cycles. The first 2-3 plies should be from light weight
The master or tooling aid should be cleaned of cloth such as 7500 glass or 2534 carbon which
all defects and debris such as scratches and will prevent mark through to the surface. The
loose fibers. Solvent clean the surface to gel coat should be advanced with time prior to
remove any residual resin or oil. Check for application of the first ply. If enough tack is
vacuum integrity using a criteria of a maxi- not present, the ply will sink too deep into the
mum loss of 500 Pa (2 in Hg) at a minimum of gel coat and be visible on the tool surface. One
6.2 kPa (25 in Hg) starting vacuum. simple test is to place a finger onto the surface
Place masking tape around the periphery and release. If the gel coat has not advanced
for later use for sealant tape. Regardless of the adequately, the fingerprint will disappear. If
prior history of the tool, it should be released the fingerprint remains, the gel coat has
with a suitable hard wax or release agent. advanced far enough to withstand the
Tooling pins should be released and placed repeated pressing of subsequent layers of
into the holes provided on the tooling aid. cloth. Using the mix ratios provided by the
For wet lay-ups, two resins are used, one for supplier, mix enough resin to cover the surface
the gel or face coat and one for laminating. The of the gel coat in the time allotted by the pot
gel coat is generally the same as the laminating life or around 30 min. Approximately 40 g
resin but with additives to thicken it to make it (O.088lb) per 0.009 m2 (1 £t2) of tool surface for
adhere to the contour of the master or tooling each ply should be adequate. Application of
aid. Resin manufactures can supply both room the first several plies should be done carefully
temperature and high temperature systems. to avoid pushing through the gel coat surface.
Apply the gel coat to the surface using a To ensure complete wetting of the ply, a
short bristled brush or squeegee. Work the short bristled brush or squeegee can be used to
coating as evenly as possible over the surface carefully work the ply into the laminating resin.
at a thickness of approximately 12 mm All wrinkles and air entrapped areas should be
(0.030 in). Do not allow excessive build up to worked out before another ply is added and if
572 Tooling for composites

necessary, use a sharp knife to slice through the FEP, bleeder and the vacuum bag. Depending
ply wrinkle in order to work it down. Bushings on the resin system and the tooling aid mater-
should be placed over the tooling pins at this ial used, a precure is recommended and
point and worked into the resin to seat them. A should be supplied by the manufacturer.
heavier cloth such as 7587 glass or 2548 carbon Fabricate an egg crate structure using 9.25 mm
can be used for the remaining plies. Since the (0.375 in) thick board stock of similar material
cloth was precut into squares or rectangles and to the laminate to avoid stresses caused by the
the warp direction was maintained, each ply difference in CTE between the egg crate and
should be placed at 45° to the previous ply. laminate. If the laminate is glass/epoxy, the
Overlaps of 6.35 mm (0.250 in) between plies egg crate material can be made from
should be maintained but a seam should never glass/epoxy or aluminum honeycomb sand-
be placed over a seam from a previous ply. wiched between a glass laminate. The egg
After each ply, add additional resin to cover the crate should be cut to the contour of the lami-
surface. Place the next ply and work the resin nate with a standoff of 3.17 mm (0.125 in). The
up through it by pressing the cloth with the standoff prevents heat differences or hot spots
brush. If there is not sufficient resin to com- on the tool surface during production part cur-
pletely wet out or saturate the ply, additional ing and also prevents mark off from the back
resin should be used. Saturation of the cloth up structure pressing upward into the tool
with resin on a table separate from the lay-up surface laminate. The egg crate should be con-
surface and then transferring the saturated ply structed so that it will lie flat on a surface. The
to the tool can cause air entrapment and bridg- intersections of the board stock are held
ing of the ply. After the fourth ply, or prior to together with cloth and resin with at least
the resin curing, apply a peel ply to the surface three strips per junction. To ensure minimal
for a compaction cycle. The peel ply when stress to the laminate from the egg crate struc-
removed before other operations will eliminate ture, it is advisable to remove the structure
the need for sanding the surface prior to bond- from the laminate and cure the strips holding
ing. This is followed by a Teflon film with the structure together at 177°C (350°F). This
perforations every 250-300 mm (10-12 in). will allow the tool surface to be tied into a sta-
Cover this with a heavy glass or bleeder cloth to bilized support structure and minimize
bleed off excess resin during compaction. warpage during subsequent cure cycles.
Finally, place a nylon vacuum bag over the sur- After the support structure is cured, the
face using sealant tape to attach to the surface structure can be placed back onto the laminate
and apply a vacuum of at least 6.2 kPa and attached or 'tied' to the tool laminate with
(25 in Hg). Hold this vacuum for 10-12 h or at least three strips of cloth and resin around
overnight or until the peel ply can be removed the periphery of the egg crate. Shims can be
without disturbing the laminate layers. used to provide for the standoff to prevent
Following this cycle, the bag and peel ply can warpage. If the type of master used for the lay-
be removed along with the tooling pins and up permits an oven cure at 177°C (350°F), then
lamination can commence as previously the post cure can be performed without
described. Debulking should be performed removal of the tool from the master. However,
after every 6 plies or before the resin begins to if the master material will not tolerate this tem-
cure. Final laminate thickness should be 9.5 mm perature, careful removal of the tool from the
(0.375 in). It is probably a good idea to build up master must be made prior to the post cure.
the bushed hole area with additional plies to Allow the tool to stand at ambient temperature
ensure vacuum integrity. for a minimum of 24 h prior to post cure.
After the final ply has been applied, the After the final 177°C (350°F) post cure,
compaction step is repeated with the peel ply, inspect the surface for pinholes and repair any
 Composite tools 573

blemishes with gel coat resin. A final vacuum the same thickness and place any bushings
check at 635 mm Hg (25 in Hg) with a loss of through the second layer and flush to the first
no more than 51 mm (2 in) is acceptable. layer. Place a ply of 7500 glass cloth onto this
layer and work in to impregnate the cloth.
Allow to cure to the fingerprint test. Mix a
25.4.4 PLASTIC FACED PLASTER
third batch of the resin but add about 10-15%
Plastic faced plasters (PFPs) are tooling aids by weight of wet plaster to the mix and apply
that minimize the wear and tear on masters by to a thickness of 2.5 mm (0.1 in). Do not wait
duplicating the master surface with a suitable for curing but proceed with a layer of plaster
unit that can be used for a variety of purposes. approx. 25.4 mm (1 in) thick. Allow this to
PFPs allow for tooling to be directly fabricated partially dry and then finish the tool by
from the master surface without exposing the adding plaster and hemp to the surface to a
master model to adverse environmental con- thickness that will allow support for the size
ditions, such as autoclave temperatures or of the tool [50-76 mm (2-3 in) for a 914 mm x
pressures. If taken directly from a master sur- 914 mm (4 ft x 4 ft) tool]. Support structure
face, the PFP is the reverse of the master can be built in for small tools using plaster
contour. An intermediate plaster splash is and hemp to make strengthening ribs on the
required to get back to the master contour back surface. For large tools, steel pipe or tub-
with a PFP. If the surface required is directly ing can be tied into the back structure with
from the master model, the PFP will be taken plaster and hemp ropes. Approximately 24 h
directly from the master surface. If the surface is required to dry and cure the system and
is above or below the master surface, appro- depends on the thickness and size. Drying in
priate steps must be taken by either taking an oven up to 60°C (140°F) will provide a sta-
additional splashes with or without layers of ble system for use. PFPs can be used in an
tooling wax to achieve the appropriate dimen- autoclave (with vacuum integrity) up to
sion. approximately 105°C (220°F), however a limit
Prepare the master surface by cleaning thor- of one or two runs is all that can be expected.
oughly and removing blemishes, debris and PFPs provide tooling aids for a variety of
pinholes by filling with a compatible filler other room temperature shop applications.
compound. Release the surface and any tool-
ing pins with a hard wax or release agent. If
25.4.5 DRILL TEMPLATES
the master model is plaster, a hard wax can be
used with a minimum of three applications, Drill templates or fixtures are used primarily
dried adequately and buffed between each to drill and locate precision holes in the pro-
application per the manufacturers directions. duction composite part. While their use is
From the resin supplier, request a water- limited to hole location and drilling, their
proof or hydrophobic resin system which will function may also be combined with other
cure in the presence of water from the plaster. support tooling, such as a trim/router fixture
Apply the resin evenly 0.76 mm (0.030 in) to minimize tooling expenditures. Drill fix-
thick to the master surface with a short bris- tures are fabricated using a room temperature
tled brush and work out air bubbles as they cured fiberglass/epoxy laminating and face
appear. Ensure that no bristles are pulled coat system. Because the tool is used in the
from the brush to contaminate the resin shop environment in ambient conditions, no
Allow the resin to cure to a point that a fin- vacuum integrity or elevated temperature
gerprint may be imprinted lightly into the requirements are needed. Location of the holes
resin and will remain for a period of time can be obtained from the master model sur-
after touching. Then apply a second coat of face. In addition, to facilitate concurrent tool
574 Tooling for composites

manufacturing, a Mylar® film sheet (0.010 in) trim and routing operations are always con-
or thicker may be used as a transfer medium ducted at ambient temperatures, CTE is not
by relocating the position of the holes to the considered in the design of this type of tooling.
surface of the Mylar and then using the Mylar Procedures for laminating the room tempera-
as a temporary master surface while fabricat- ture cured system are similar to those for drill
ing the fixture. Prior to fabrication, it must be fixtures. Periodic debulking is not required
determined if the fixture is to mount on the and the tool is not required to maintain any
outside surface of the part (OML) or the inter- vacuum requirements. Thickness can vary
nal surface of the part (IML). Since master depending on final use but is usually about
models normally represent the OML surface, 9.5 mm (0.375 in). If the tool is to be used for
most support tooling such as drill and trim fix- routing, a set back or offset will have to be
tures may be fabricated directly off this determined as defined by the type of routing
surface. Occasionally, the surface required will equipment used. This set back must be identi-
be a specific distance above or below the sur- fied on the surface of the tool to alert
face of the master model. If the surface is personnel to which equipment is acceptable
above, tooling sheet wax at the specific dimen- for use with the tool. Failure to use the correct
sion required must be placed on the master set back will result in an under trimmed or
surface prior to fabrication of the tooling aid. over trimmed condition. A witness or verifica-
Sheet wax is commercially available in numer- tion line is usually scribed on trim fixtures as a
ous thickness to accommodate most reference to which edges may be checked for
requirements. If the surface required lies damage. This allows shop personnel to
inside the master model surface, a 'splash' quickly verify the accuracy of the trim fixture
consisting of plaster and hemp reinforcement with minimal inspection equipment. With drill
must first be pulled from the master surface fixture tooling, a determination must be made
and then the splash surface can be waxed to as to what surface the trim fixture is applied. If
the specific dimension inside the master sur- the tool is to represent the same surface of the
face. Tool pins are placed in the tooling aid. master model, then the tool may be directly
After the face coat and two layers of glass have taken from that surface. However, if the sur-
been applied, the drill bushing is seated onto face required is internal or external to the
the surface with additional lamination over master, appropriate steps including plaster
the bushing to provide an integral lock to the splashes or waxing must be completed to
tool. Template thickness can vary depending obtain the correct surface.
on use, but 9.5 mm (0.375 in) in thickness is
typical.
25.4.7 PLY LOCATING TEMPLATES

25.4.6 TRIM AND ROUTER TEMPLATES

Ply locating templates are used during the lay-
up of the production part and designate
These shop aids are used to trim and rout locations for the plies and indexing of detail
cured composite parts to a specific dimen- parts. In addition, these templates may also
sional tolerance. Accuracy is required for these show individual ply orientation and designate
tools in order for the composite detail to fit special features of the part such as splice areas
precisely with adjacent details. Trim and or hardware attachment points. Occasionally,
router templates can be fabricated directly honeycomb core placed within composite pro-
from the master model, composite tool or a duction details must be potted with a syntactic
tooling aid such as a PFP (plastic faced plas- core material to prevent core collapse when
ter). They are generally fabricated using room hardware is attached. Reference locations of
temperature cured epoxy/glass systems. Since all attachment hardware may be transferred
 Bibliography 575

from the master model to produce a core pot- lighten the template to assist in handling.
ting template. Core may then be accurately Lightening holes can be placed by removal of
potted, laminated and then drilled after curing sections of the template not serving a specific
by referencing the location from the potting function. However, as material is removed to
template. Similar templates may be produced reduce weight, stiffeners must be added to
to locate individual details or other assemblies prevent warpage that may affect dimensional
that have to be located during detail construc- stability.
tion. Slits or eyebrows are cut into the laminate
to locate the edge of the production part ply
BIBLIOGRAPHY
and color coded and identified accordingly.
All templates are fabricated from room tem- Mallik, P.K., Fiber Reinforced Composites, New York:
perature glass/epoxy cured systems and are Marcel Dekker, 1968.
designed to be light in weight with a thickness Fiberite Manufacturing Procedures, Toolrite Tooling
Materials System.
of 3.1-3.8 mm (0.125-0.150 in). United States Gypsum, Tooling Techniques.
Because some templates may be quite large, Morena, J.J., Advanced Composite Mold Making, New
provisions must be made where possible to York: Van Nostrand Reinhold.
CONSOLIDATION TECHNIQUES AND CURE 26
CONTROL
Zhong Cai and Timothy Gutowski

26.1 INTRODUCTION r-----------...,- \lQcuum Bog

Consolidation is an important step that occurs ----------t.--Air
-'-'-'-'-'-'-'------ --I--Porous Teflon
Breather

in almost every process used to make an !fI1111I11111I1111111I111I1I1111I1I11/' -+--Metof Plate

------------------------- Nonporous Teflon
advanced composite article. Consolidation is e:.::.:.,~~w·1·~i4!:MWAA1'fm:ffiMm:miij - - 1 - - Bleeder
usually brought about by the application of -·-·-·-·-·-=:l==~wfleffon
pressure at a boundary which squeezes air and ~~~~~~~~~=$=pre
_ PeelP'!!9
Ply
Nonporous Teflon
resin out of the composite thereby changing mUfPfoTT.rTi"T7'7'7'T'T"~T"T"rl'lr4l;~-Dom
ILLLL..'-L.I.J...L.L.LL.J...f..L..I'.J..J.'.J..J..LL..LLLA- Metol Tool Plate
both its microstructure and dimensions.
Improper consolidation can lead to voids,
residual stresses, warping and other unwanted Fig. 26.1 Schematic of the prepreg lay-up used in
effects which could ultimately lead to the rejec- autoclave cure (Springer, 1986).
tion of the part. A comprehensive discussion of
consolidation in composites would include lay-up used in autoclave cure. In a traditional
many complex phenomena. Simultaneously, lay-up process, prepregs with different fiber
there is heat, momentum and mass transfer, orientation and architecture are placed in cer-
accompanied by the chemical curing reaction tain order forming a near-net-shape composite
of the resin and the deformation and motion of structure. During the lay-up operation,
fibers. whether it is done manually or using a robot,
Consolidation techniques have been used in the trapping of air pockets within the structure
the fabrication of both thermoset and thermo- is unavoidable. Thus a consolidation step after
plastic composite parts, but are more crucial the lay-up operation is necessary. Prepregs are
steps in thermoset composite processing. The usually provided with relatively low fiber vol-
traditional composite manufacturing process ume fraction. With the consolidation step, the
for aerospace industry products usually starts fiber volume fraction of the composite product
with the B-stage impregnated prepregs con- can be increased and excessive resin can be
sisting of fiber preforms and staged resin removed.
matrix. Usually the resin content is relatively The basic mechanisms involved in a consol-
high. In order to achieve the required compos- idation process are the fiber deformation and
ite material properties which are dominated resin flow, which are coupled with thermal
by the fibers, consolidation is used as an effects and the resin cure reaction. A similar
important processing step. consolidation process can also be seen in soil
Figure 26.1 shows a setup of the prepreg mechanics such as the settlement of a founda-
tion. However, the deformation behavior of
fibrous materials is substantially different
Handbook of Composites. Edited by S.T. Peters. Published from that of granular structures and resin flow
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 behavior is strongly affected by thermal effects
 Introduction 577

and chemical reaction of the resin. Thus, the surface roughness. Metals are widely used as
study of the consolidation process of fibrous tooling materials for composite processing.
composite materials involves many disciplines. However, their heavy weight and high cost of
To effectively control a consolidation machining become disadvantages when com-
process, the selection of the equipment and plex geometry is involved. Composite tooling
tooling materials is crucial. Major process materials have been used as alternatives in
parameters for a consolidation process include various consolidation processes.
pressure and temperature and both are func- Another tooling component for the consoli-
tions of time and are usually set as operation dation process is the bleeder, which is usually
cycles. Thus the system setup should be able a nonstructurallayer of porous cloth or paper
to effectively control the pressure and temper- which allows the escape or bleed out of exces-
ature profile and transfer heat and pressure to sive gas and resin during the consolidation
composite parts. Figure 26.2 illustrates the process. Sometimes the process is called
process variables applied during autoclave migration. The bleeder cloth or paper is
consolidation and cure. removed after the curing process and is not
part of the final composite.
TEMPERATURE Breather material is used to provide a vac-
PRESSURE uum path over the surface of the part. Typical
materials are glass and mat. They can be
stretched over the part contours to ensure an
effective vacuum path and sometime also to
I':==::~:::':-BLEEDEF(::::::=:~~
------ --------- ---
----- provide a cushion effect to matched metal tools.
Bagging and sealing are crucial to the qual-
ity of the composite parts. General
requirements for the bag are: (1) the bag must
apply curing pressure uniformly; (2) the bag
must not leak under molding conditions; and
t t t t t (3) a good vacuum path must be provided in
bagging. Silicone rubber vacuum bags are
Fig. 26.2 Illustration of the process variables (tem- widely used because of their long service life.
perature, pressure) applied during autoclave Moreover, they are repairable and self-healing
consolidation and cure (Springer, 1986). with respect to pinholes. The initial cost of fab-
rication is relatively higher. Nylon is an
One of the commonly used facilities is an auto- alternative bag material for up to 193°C
clave, which is a closed pressure vessel with (380°F) and is usually discarded after use.
means for heating and applying pressure and The commonly used form of resin matrix
vacuum to its contents. The dimensions of the prepreg has a resin content beyond 40% and
composite parts are limited by the size of auto- requires a significant amount of resin bleedout
claves. Thus, for large size composite during cure to achieve a cured laminate resin
structures, alternative processing techniques content of 28-32%. Low resin content prepregs
have been used, such as vacuum bag molding. have been developed which can be used with-
In addition to the equipment, tooling mate- out resin bleedout processes. Since there is no
rial has direct influence on the composite part bleedout process, less resin and less bleeder
surface quality, dimensional accuracy and material are needed for a consolidation and
residual stress. The main considerations for cure process. However, the removal of
tooling material include strength, stiffness, entrapped air becomes a more critical aspect of
thermal expansion coefficient, hardness and process control.
578 Consolidation techniques and cure control

26.2 CONSOLIDATION MODELS portion of the load, then Gutowski's and

Kardos' models are applicable.
As composite applications were expanded In the following discussions, both
rapidly in the late 1970s and early 1980s, stud- Springer's model and Gutowski's model will
ies on the process science of composite be presented. Kardos' model is equivalent to
materials became very active, especially in the Gutowski's model but different process vari-
areas of consolidation and cure (Lindt, 1982, ables are used in the modeling.
1986; Springer, 1982, 1986; Loos and Springer,
1983a,b; Halpin, Kardos and Dudukovic, 1983;
26.2.1 RESIN FLOW
Loos and Freeman, 1985; Gutowski et al.,
1987a,b; Gutowski and Cai, 1988; Dave, The problem of resin flow in composite pro-
Kardos and Dudukovic, 1987a,b; Tang, Lee cessing can be treated as flow through fibrous
and Springer, 1987; Batch and Macosko, 1988; porous media. In general it can be handled by
Kim et al., 1988 and 1989; Connor et al., 1993). Darcy's law which states that the flow rate is
The purpose of these studies has been to find proportional to the pressure gradient applied
out the most suitable process parameters and and is related to the porous medium perme-
then through scientific process modeling to ability and fluid viscosity. The general form of
achieve the optimized composite product Darcy's law in a one-dimensional case is:
quality.
Pioneering work in composite consolida- q = - E EE. (26.1)
tion and cure modeling was led by Springer Il dx
(Springer, 1982; Loos and Springer, 1983a). where q is the average resin flow rate, K is the
Laminated composite structure was consid- preform permeability with the units of length
ered with bleeder layers placed on top of the squared, Il is the resin viscosity and dp / dx is
composite laminates. When pressure was the imposed pressure gradient. In the case of
applied transversely to the laminate plate, resin flow into the bleeder, the consolidation
excessive resin material which was in a fluid pressure is established between the advancing
state was squeezed out from the laminate. flow front and the tool or mold surface.
Similar consolidation models have also The main issue involved in using Darcy's
been developed by Gutowski's group and law in the consolidation process is that neither
Kardos' group (Gutowski, Morigaki and Cai the fiber preform permeability nor the resin vis-
1987a; Gutowski et al., 1987b; Dave, Kardos cosity is constant over the process. The preform
and Dudukovic, 1987a,b). In these models, permeability is a function of the porosity or
both the fiber material deformation which is fiber volume fraction, fiber diameter, fiber ori-
highly nonlinear and the outgoing flow of the entation and fiber architecture. Among them
resin are considered. the fiber volume fraction changes substantially
As can be seen from the discussion pre- during a consolidation process. Resin viscosity
sented later, both approaches are valid within is related to temperature, the cure status and
the ranges of parameters considered. cure time and changes dramatically in the
Experimental verification results show good process. Usually at the start of a cure process,
agreement with these model predictions resin is in the semi-solid state. With the rise of
(Gutowski et al., 1987b; Kim et al., 1988; Cai temperature, it becomes fluid. As the degree of
and Gutowski, 1989). When resin content is cure increases, it gels and becomes solid.
relatively high and fiber-to-fiber contact is The permeability of fibrous preforms has
insignificant, Springer's model can be applied. been studied both analytically and experimen-
On the other hand, if fiber volume fraction is tally (Williams, Morris and Ennis, 1974;
relatively high and fibers carry a substantial Gutowski et aI., 1987b; Lam and Kardos, 1988,
 Consolidation models 579

1989; Van Den Brekel and De Long, 1989)

using the well-known Kozeny-Carman equa- D 20 ply sample (0.002 in/min ,corn oil!

-., sk
• 20 ply sample (0.005 in/min, corn oill
tion. The estimation formula using the fiber
structural variables can be written as: ... 1:1 8 ply somple (0.002 in/min, silicone fluid)

r; --e
tit
0
(1- vy
~
8 ply sample (0.01 in/min. silicone fluid)

K = - 4k
I I)
V2 (26.2)
o f N Corman - Kozeny Eq. ku=O.7
•
where r f is the fiber radius, Vf is the fiber vol- 2>C 5 \\
ume fraction so that (1 - Vf) is the porosity and >C 60
ko is an empirical constant, called the Kozeny M
t/) 4 ~x
constant, which is usually determined experi- >- 0
~
:scu
3
mentally. For different textile architecture and
orientation, the value of ko will be different. 2 61.,e •
E
Reported experimental data show that for an ~
cu
Cl..
~ll
aligned fiber bundle, ko = 0.5-0.7 for the longi- .90-01>_ - -"
tudinal flow and ko = 11.0 for the transverse
flow. For ± 45° cross plies, ko = 2.70. For woven
0.40 0.50 0.60 0.70 O.so
Fiber Volume Fraction (Vf)
type textile preforms, ko = 5.5. It should be
pointed out that many experimental results
Fig. 26.3 Comparison of measured axial permeabil-
have been reported and the variation of the ity values for aligned AS-4 fibers with
Kozeny constant in some cases is significant. Carman-Kozeny equation (Gutowski et al., 1987b).
Also in the transverse direction, a modified
Kozeny-Carman equation has been proposed
to account for the stop-flow phenomenon 10',.------------------,
when fiber volume fraction reaches the maxi- 3501-6

mum packing efficiency (Gutowski et al., 425K

1987b). Figure 26.3 shows a comparison of 10' 400K

measured axial permeability values for aligned

fibers with the Kozeny-Carman equation.
Resin viscosity can be expressed as an
empirical function of temperature and degree
of cure (Lee, Loos and Springer, 1982). The
expression can be written as:
fi = fi~ exp (U/RT + lea) (26.3)
20 40 60 eo 100 120 140
where fi~ is a constant, U is the activation TIME {mIni
energy for viscosity, a is the degree of cure
and Ie is a constant which is independent of Fig. 26.4 Measured viscosity of 3501-6 resin as a
temperature. function of time (Lee et al., 1982).
Experimental study has been performed for
the Hercules 3501-6 epoxy resin which is at a constant temperature. Thus the value of Ie
widely used in composite fabrication. Figure is found to be 14.1 ± 1.2. The values of p,~ and
26.4 shows the viscosity measurement as func- U are found to be p,~ =7.93 X 10-14 Pa s, U =9.08
tions of temperature and time. To match the x 1()4 J mol-to
model predictions and experimental data, the The degree of cure a and the rate of degree
constant Ie is found by fitting a linear least of cure da / dt were determined from the
square curve to the fi versus a data generated results of isothermal scanning experiments.
580 Consolidation techniques and cure control

Efforts were made to describe the da / dt ver- When the permeability and the resin viscos-
sus a data with a modified Arrhenius type ity are known, with the imposed applied
equation. The proposed empirical equations pressure condition, the rate of the outgoing
are resin flow can be calculated using the Darcy
equation. In general, flow may be multi-direc-
da/dt = (KI + Kp) (I-a) (B -a) (26.4) tional. Thus 2-D or 3-D flow equations have to
a::;0.3 be solved. In practice, resin flow in one partic-
ular direction may be dominant, and the
da/dt = K3 (I-a) (26.5) analysis can be handled as 1-D permeable
flow.
a>0.3
where
26.2.2 FIBER DEFORMATION
~ =Al exp (-M,/Rn The main contribution from Gutowski's model
is the description of fiber deformation behav-
~ =A2 exp (-flE/Rn ior. Instead of treating fibers as separate layers,
a network concept is introduced. In other
K3 =A3 exp (-M,3/ Rn words, fiber-to-fiber contact is assumed within
a fiber assembly, even in the case of aligned
At' A2 and A3 are the pre-exponential factors,
fiber bundles. Thus a fiber filament span
M,I' M,2 and flE3 are the activation energies, R
between the neighboring contact points
is the universal gas constant, and T is the
becomes a small bending beam. During a con-
absolute temperature. The constants in the
solidation process when fibers are pushed
expression are found as:
closer, more and more fiber-ta-fiber contacts
B = 0.47 take place, and the span length reduces. Thus
the bending stiffness of these small fiber
Al = 2.101 X 109 min-I beams increases rapidly, resulting in nonlinear
elastic deformation response. The nonlinear
elastic response of a fiber assembly under a
compressive load has been also studied in the
textile field, and an empirical formula was
proposed (van Wyk, 1946).
flEI = 8.07 X 104 J mol-1 A proposed fiber deformation model for
aligned fiber bundles considers the deforma-
M,2 = 7.78 x 104 J mol-I tion status variable, the fiber volume fraction
Vf1 as a function of the consolidation pressure
M,3 = 5.66 x 104 J mol-I (Gutowski, 1985). The expression is
-'/(V/Vo) -1
As can be seen from the discussion, all the con- a f = As [-./(V/V f ) _ 1]4 (26.6)
stants involved in the model are determined
experimentally through a specified process. where Va is the maximum obtainable fiber vol-
Similar treatment can be used for other types of ume fraction for a given fiber network
resin systems, and experimental investigation configuration, and Va is the fiber volume frac-
results have been reported, including Hercules tion below which the fiber network carries no
HBRF-55 Resin (Bhi et al., 1987) and Fiberite 976 load. The empirical constant A. is obtained
Resin (Dusi et al., 1987). A similar process model from curve fitting on available measurement
has also been discussed by Roylance (1988). data. A typical fiber deformation curve for
 Consolidation models 581

well aligned graphite fibers is shown in Fig. 26.2.3 CONSOLIDATION MODELS

26.5 with the comparison of measured data
As discussed above, in Springer's model, it is
points.
assumed that there is no fiber-to-fiber contact.
Thus a dynamic fluid pressure exists between
:=VI 800 the consolidated layers. The consolidation
Theory~
time, which is crucial to the cure process, is
..9- 700 related to the permeability of the fibrous pre-
b
.. 600
VI
forms, resin viscosity, and the applied
VI consolidation pressure. In Gutowski's model,
- 400
~500
c J)
Doto Point ____
the fiber reinforcement and the fluid state
resin are considered as a system. Both fiber

-
Q)
300 network deformation and fluid resin flow are

-
>
solved together. Both models are presented
~ 200 here with a laminated composite structure as
w 100 the example.
OL---~~~~L-~L-~ The example for Springer's model is the
0.4 0.5 0.6 0.7 0.8 0.9 laminate consolidation with flow in the lami-
Fiber Volume Fraction (Vf) nate transverse direction, or z direction. A
bleeder ply is assumed to be placed on top of
Fig. 26.5 Typical fiber deformation curve for well- the composite. Figure 26.6 shows the setup for
aligned XA-S and AS-4 graphite fibers (Gutowski et the model. At any instant of time the liquid
al., 1987b). velocities in the bleeder Vb and in the compos-
ite Vo are given by Darcy's law. For a constant
viscosity liquid, the integrated forms are:
Ke (Po-p)
The proposed relationship between the com-
qe = IJ (26.7)
F he
pressive fiber stress af and fiber volume fraction
V f provides a tool to estimate the finished con- Kb (Pu - Pb)
solidation status of the composite products. If qb = Y ~ (26.8)
the time window for the consolidation is long
enough, and excessive resin is completely where Pu and Pb are the pressures at the com-
squeezed out from the structure, the consolida- posite-bleeder interface and in the bleeder
tion pressure is then balanced by the fiber respectively, Po is the consolidation pressure
stress. However, because of the dramatic and is related to the applied force or pressure,
change of the resin viscosity and preform per- ~ is the instantaneous thickness of the liquid
meability during a consolidation process, resin in the bleeder, and he is the thickness of the
flow may not be complete. Thus, developed resin starved layer, or the thickness of the lay-
consolidation simulation models are needed for ers through which resin flow takes place, and
the process analysis and improvement. Ke and Kb are the permeability of the compos-
During the compression of fibrous pre- ite layer and bleeder respectively. If the
forms, structural relaxation has been observed compacted composite layer thickness is hI'
(Gutowski, 1987b). Thus the deformation to then
some extent is not elastic but viscoelastic. This he =nh1 (26.9)
issue has been addressed by using a Maxwell
type model (Kim, McCarthy and Fanucci, where n is the number of layers or plies
1991). already compacted.
582 Consolidation techniques and cure control

To, Po

Interface:
Pressure: Pu
.L Temperature: Tu
hb~~~~~~~~~~~~~
-L l...:::::!I'------Resin Flow
hc,--:

Fig. 26.6 Illustration of the consolidation model proposed by Springer (1982).

The equation of continuity gives the rate of layers. The final status of the composite is
change of volume of the composite as: dependent on the compaction of each individ-
uallayer.
_ d(hA) = Aq = Aq (26.10) As a comparison, Gutowski's consolidation
dt c b
model combines the flow of resin through
where A is the surface area of the composite porous media and the fiber deformation
laminate, and h is the total thickness of the behavior. Similar treatment has been pre-
composite laminate. The second equation sented in studies of other fields including soil
expresses the fact that at any instant of time, mechanics (Biot, 1941, 1955, 1956; Gibson and
the flow out of the composite is equal to the Hussey, 1967). In general, consolidation occurs
flow into the bleeder. The pressure Po is related in only one direction, but flow may take place
to the applied force as: in all three directions. Thus an element is
deformable in the z direction. A new variable ~
F
Po = A + Pa (26.11) is used to represent the deformation, and ~ = z
+ w where w is the local displacement of the
where F is the applied force and Pa is the fiber network. The laminate setup for the
atmospheric pressure. By combining these model is illustrated in Fig. 26.7. If the initial
equations, the consolidation equation fiber volume fraction for the composite is ~
becomes: and the fiber volume fraction at any instant is
d(hA) _
-~ - phI
Kll(n + ~hI
KA) (26.12)
VfI the fiber continuity condition states

Vo =: ~ (26.13)
Therefore for each individual layer, the consol-
idation time can be calculated. The total Resin flow continuity condition requires:
consolidation time is the summation for these
 Consolidation models 583

Here it is assumed that the inertial effects in

the process are small. Therefore the applied
pressure is balanced by a combination of the
average resin pressure and the fiber stress. In
(26.14) other words, any load which is carried by the
fibers is then unavailable for pressurizing the
With the application of Darcy's law, a consoli- resin.
dation equation using the fluid pressure Pr and Since both the permeability and the fiber
fiber volume fraction V f as variables can be stress are expressed as functions of fiber vol-
written as ume fraction Vi' with the given initial and
boundary conditions, the variables V f and Pr as
a (Vo Kx apr) a (Vo Ky aPr)
ax V f ----;; ax + dy V f ----;; ay + a function of time and location can be solved.
In general numerical calculation procedures
have to be developed for solving the partial
~(Vf Kz aPr) + V£ aVf) = 0 (26.15) differential equations. In some simplified
az Vo fl az V/ at cases, analytical solutions are possible.
This equation gives a relationship between the
spatial and time-varying nature of the pres-
Example problem 1: One-dimensional flow
sure in the resin and the fiber volume fraction
in compression molding
of the composite. The equilibrium statement
for the consolidation is: A simplified example of composite consolida-
tion is the compression molding of a flat
prfi = Jfr dA + ar4 (26.16) rectangular laminate. The composite part is
pressed between two solid dies. Therefore
only in-plane flow is possible. In other words,
flow components are in the x and y directions
only. If the initial fiber volume fraction is uni-
form, the equation of the resin flow and fiber
deformation becomes:
a2Pr aZpr fl av£
Kx dX2 + Ky dy2 + V f at = 0 (26.17)

Here it is also assumed that there is no signifi-

cant pressure gradient in the z direction, and
the viscosity fl does not vary spatially.
In some cases, K/ a2 » K/b 2 where a and b
are the dimensions of the laminate in x and y
directions respectively. The compression
molding results in primarily one-dimensional
~+~ dz flow in the x direction. Then the equation can
be solved analytically. With the assumed
boundary conditions of Pr = 0 at x = ±a and
ap/ax = 0 at x = 0, the result is a parabolic
pressure distribution as
_ L a2 - x2 ( dVf )
Fig. 26.7 illustration of the consolidation model
proposed by Gutowski et al. (1987a).
Pr - 2Kx Vf dt (26.18)
584 Consolidation techniques and cure control

The solution for the fiber volume fraction V f as Example problem 2: Compression molding
a function of time is: with two-dimensional flow
_ pa2 dVf Here the case of compression molding of a rec-
Po - <1f(Vf) + 3K V dt (26.19) tangular laminate with an isotropic in-plane
x f
permeability is considered. In other words, Kx
This expression shows how the applied pres-
sure Po is carried by the fiber stress a f and the
= Ky = K. This may correspond to a quasi-
isotropic lay-up. The flow equation becomes
average pressure in the resin. The load sharing
Poisson's equation, which can be solved by the
in a composite is directly analogous to how the
separation of variables technique. The solution
load is shared in a parallel spring and damper
for the pressure distribution in a laminate with
set. For example, initially if Vf is less than V0'
zero pressure at the boundaries is:
then there is no deformation in the spring
(fibers) and the entire load is carried by the = 16 a2p. d Vf ~ ~(-l)(n -1)/2
resin. On the other hand, at long times and Pr nJ KV dt f
I. n3
n=l,3, ...
finite viscosity, if the rate of change of Vf is
close to zero, then the pressure in the damper
(resin) goes to zero and the total load must be x [1- cosh (n7ty/2a)]cos mtx (26.20)
cosh (nxb/2a) 2a
carried by the fibers. Figure 26.8 shows an
example of the one-dimensional flow in com-
With the applied load balance condition, the
pression molding with the comparison of
final result is:
computer simulation results.

------------- --
'«{««<U«<Ult'l
"1"

---
<.t<,,,,t,( «tt,(,(,( (
dVf [1- 192!!
dt xs
I. ... n5
b n=I,3,
~ ~tanh nXb]
2a
(26.21)

-
" }'>"),;,,,»»»--- ......
»;;n;;;;;,) >nn,»»·
It can be seen that the result is analogous to the
previous case except for a geometry effect
term which is shown in the bracket.

- p,.. Theory Example problem 3: Bleeder ply molding

o PR• Measured This has been presented with the Springer's
model. In this case, a porous bleeder ply is
placed on top of the composite, and flow is
principally in the z direction. With the intro-
duction of a new variable, the void ratio
e = (1- Vf)/Vf' one may obtain the nonlinear
100
" o one-dimensional consolidation equation. An
o ~6~~~L-~-~=;~~~ equation similar to this was first derived by
o 5 10 15 30 35 40
Gibson et al. (1967) for the consolidation of sat-
Time (min)
urated clays. The expression is:
Fig. 26.8 Example of one-dimensional flow in com-
pression molding and computer simulation results oe
ot
= (e0 + 1)2~(- K. oaf oe)
oz p(l + e) oe oz
(2622)
.
(Gutowski, et al., 1987b).
 Consolidation models 585

The void ratio e or the fiber volume fraction V f

is a function of both time and location. An
equivalent equation using variables Vf and Pr
can be written as

With similar pressure equilibrium conditions,

the distribution and time history of Vf or e can
be solved numerically. Figure 26.9 shows an
example of the bleeder ply molding measure-
PRESSURE
ment setup, and the comparison of the TRANSDUCER
computer simulation results with the mea-
sured data.
It is interesting to see that, with Gutowski's
600
consolidation model, the final status of the
composite in terms of the average fiber volume o data
500 -theory
fraction can be estimated from the proposed
fiber deformation model if the consolidation
~ 400
process is complete. The consolidation time for VI
a.
a particular setup can be solved through
300
numerical simulation. ....
OJ
:::J
As can be seen from the analysis, the total VI

consolidation time for a composite structure is

II)
200
....
OJ

strongly dependent on the dimension in the Cl.

100
resin flow direction. For laminated composite
structures, usually the dimensions in x and y
directions (directions within the laminate 0
0 to 20 30 40 50 60
structure) are much larger than that in the z Time (min)
direction (direction transverse to the laminate
plane). For example, many aerospace struc- Fig. 26.9 Example of bleeder molding and com-
tural parts range from a few inches to several puter simulation results (Gutowski et al., 1987b).
feet in x or y direction, but only have a thick-
ness of a fraction of an inch in the z direction. composites (Kim, Jun and Lee, 1989).
Thus the bleeder ply molding process is pre- It can also be seen from the comparison of
ferred and is widely used in many part the two models that with relatively low fiber
fabrication processes. However, for the so- volume fraction, fibers carry almost no load.
called thick composites, for example with Thus the consolidation process is dominated
lay-up of 64 or 96 plies, the consolidation time by the resin flow through the fiber network.
required increases dramatically in the bleeder Then the difference between the two models is
ply molding cases. With the selected cure cycle very minor. Springer assumes the consolida-
for thin composites, complete consolidation tion is done layer by layer, while Gutowski
may not be achieved for thick composites. treats the fiber network as a whole system.
Thus the final fiber volume fraction of the thick However, in both cases the top layers are con-
composite tends to be relatively lower. This has solidated first. When the fiber volume fraction
been observed in experiments involving thick becomes high, then the predictions from the
586 Consolidation techniques and cure control

two models show significant different results. Specifically, a selected cure cycle must
On the other hand, the numerical schemes of ensure that:
the two models are different. Springer's model
1. the temperature inside the material does
requires only the solutions of a series algebraic
not exceed a preset value at any time during
equations, while in Gutowski's model nonlin-
the cure;
ear partial differential equations have to be
2. at the end of the cure the resin content is
solved. A comparison study has been pre-
uniform and has the desired value;
sented by Smith and Poursartip (1993).
3. the material is cured uniformly and com-
pletely;
26.3 CURE CONTROL 4. the cured composite has the lowest possible
void content;
Fiber reinforced thermosetting resin compos-
5. the cured composite has the desired ther-
ites manufactured in autoclaves are made by
mal and mechanical properties;
forming the uncured fiber-resin mixture into
6. the curing is achieved in the shortest time.
the desired shape and then curing the mater-
ial. Curing requires the application of heat and Figure 26.10 shows schematically the overall
pressure. Heat is used to facilitate and control cure process model structure. In an early study,
the chemical reactions of the resin, and pres- Loos and Springer (1983a) proposed a thermo-
sure is used to consolidate the composite, chemical model. Heat transfer from the
squeeze out the excess resin, and minimize the environment to the composite material deter-
void content. A cure cycle usually means the mines the temperature distribution, the degree
magnitude, duration, and profile of the tem- of cure of the resin, and the resin viscosity
perature and pressure applied during a curing within the composite structure. The tempera-
process. Selection of the cure cycle directly ture inside the composite can be calculated
affects the quality of the finished composite using the law of conservation of energy. By
product, such as fiber content, fiber distribu- neglecting the energy transfer by convection,
tion, and void percentage. the energy equation can be expressed as:

Viscosity ----------I~~ Flow

Reaction kinetics Void

Heat transfer _ _ _ _ _ _-I.~ Residual stress

Fig. 26.10 Schematic of overall cure process model (Dave et al., 1990).
 Cure control 587

resin viscosity, the degree of cure a and the

rate of the cure da/dt can be characterized
using a modified Arrhenius type equation,
with relevant constants in the model deter-
(26.24) mined experimentally (Lee, Loos and
Springer, 1982; Bhi et al., 1987; Dusi et al., 1987;
where p and Cv are the density and specific Roylance, 1988). Figure 26.11 show an example
heat of the composite, k:c' k and kz are the ther- of the rate of heat generation and rate of
T
mal conductivities, and is temperature. In degree of cure of the 3501-6 resin system as
the case of relatively thin composite structure, functions of time and temperature.
conduction heat transfer is mainly in the z It is noted that the density p, specific heat Cyl
direction. Thus terms in the x and y directions heat of reaction H" and thermal conductivity k
can be dropped. The rate of heat generation are all dependent on the instantaneous and
dH/dt is defined as: local resin and fiber contents of each ply, and
dH da
dt = Hr dt (26.25)
0.1.-------,
4QOK

where Hr is the total heat of reaction depend- -

• 0010
Colculo1ed
ing on the resin type. The rate of the cure
reaction is a function of temperature and the
cure status, and can be expressed symbolically
as:

da
ill " a)
= J(T, (26.26)

The degree of cure is then determined as:

It is assumed that for an uncured material, a =

0, and for a completely cured material, a
approaches unity. As discussed earlier the
(26.27)

0.1

~---.j
.,0\ 0 ......
°0!--'-="0.2=--'-""""'0:'-:.4c-'--='0.6 0 0.2 0.4 0.6
DEGREE OF CURE, a

TEMPERATURE (K)
320 350 400 450 500 550 600
..,

"".
z
0
>=
3501 - 6

<I
a:
w Io.smeal/sec

~ +1
w
J:
....

iJ 0 s 10
Fig. 26.11 Rate of heat generation
and rate of degree of cure of the
3501-6 resin (Lee, Loos and
TIME (min) Springer, 1983).
588 Consolidation techniques and cure control

can be handled using rule of mixtures (Loos pressure early in the cure cycle and the initial
and Springer, 1983c) or proposed approximate resin moisture are crucial considerations in
formulas (Springer and Tsai, 1967). producing void-free laminates (Kardos et al.,
The solution to these equations can be 1983, 1988). Since the driving force for diffu-
obtained once the initial and boundary condi- sion rises with temperature, in order to
tions are specified. The initial conditions prevent the potential for pure water void
require that the temperature and degree of growth by moisture diffusion in a laminate at
cure inside the composite be given before the all times and temperatures during the curing
start of the cure. The boundary condition cycle, the resin pressure at any point within
requires that the temperatures on composite the curing laminate must be higher than the
surfaces in contact with the tool be known as a minimum resin pressure required, which is a
function of time during cure. Therefore the function of the relative humidity and temper-
boundary condition is related to the specified ature (Dave et al., 1990). Figure 26.12 shows a
cure cycle and the equipment setup. void stability map for pure water void forma-
The objective for the cure control scheme is tion in epoxy matrices. A similar pressure
to achieve the desired composite quality. Some requirement also holds for small air/water
of the main targets are reasonable temperature voids after an initial growth period. It has also
distribution, complete consolidation, mini- been observed that the void content is reduced
mum thermal stress and minimum void
content.
With a developed numerical scheme, the (1 ATM = 101 kP.)

temperature distribution inside the laminate is

calculated as a function of position and time. A
good cure scheme should realize the two main (RH). = 100% (RH). =50%
targets: (a) the temperature is reasonably uni-
6
form inside the material and (b) the
temperature does not exceed a preselected
maximum at any time. :I
~ 5
For a given cure temperature and cure pres-
iii
sure, the time window for the consolidation is a:
then specified. From the consolidation models,
the compaction status of the consolidated a
a:
composite can be obtained. In Springer's ""z
model, the result is the total number of com- f5 3
pacted plies, while in Gutowski's model the a: VOID GROWTH
result is the V f distribution across the layers. If POSSIBLE
the consolidation cannot be completed with 2
the selected cure cycle, proper modifications
are then made. The compaction issue becomes
crucial to the cure process of the thick com- I ~ 1 --------
posite structure. A multiple stage heating
process may be designed to defer the cure
reaction of the resin and thus prolong the con-
~l
300 400 500
solidation time window. T,K
Voids within the composite material are
harmful to its mechanical performance. Fig. 26.12 Void stability map for pure water void
Experimental study shows that the resin formation in epoxy matrices (Dave et al., 1990).
 Effects of tooling and part shape 589

significantly when the applied pressure is suf- heated to a specified temperature at a speci-
ficiently high to collapse the vapor bubble fied rate under controlled conditions in the
before the gel point is reached. Therefore, after ,autoclave.
the time-temperature cycle is determined, it is Tooling materials may be metal (steel,
possible to obtain a profile of the minimum nickel, nickel alloys, and aluminum),
pressure versus cure time. The boundary pres- graphite-epoxy and elastomer, depending on
sure is then maintained greater than the different composite part shape, size, volume of
minimum pressure throughout the cure cycle. production and curing method. Selection of
During the cooling stage after the cure of the tooling material often reflects a compro-
the composites, residual thermal stress is mise among these considerations. Thermal
related to the difference between the cure tem- behavior of the tooling material is also crucial
perature and ambient temperature, and the in the design and fabrication. Table 26.1 lists
thermal expansion behavior of the composite the coefficient of thermal expansion of differ-
material. For a laminated structure, calcula- ent composite and tooling materials. The
tion of the thermal stress has been discussed values for the composites are dependent on
and formulated by Tsai and Hahn (1980). the ply orientation and fiber volume fraction,
Since the material shows viscoelastic behav- and typical values are shown there.
ior, stress relaxation has been observed over
time. A post-cure process is usually applied to
the structure to relieve the induced thermal
stress. Table 26.1 Coefficient of thermal expansion (CTE)
For large complex-shaped composite struc- for various materials (Borstell and Turner, 1987)
tures, non-autoclave curing methods are used.
Compared with traditional autoclave curing Material CTE (1D-6/K)
methods, the component size restrictions are Structural composite material
eliminated, energy consumption is reduced, Boron-epoxy 3.6-10.8
and capital equipment cost can be cut down. Aramid-epoxy -2.0-5.8
The non-autoclave processes use an oven, inte- Graphite-epoxy 1.8-9.0
grally reinforced tools, and presses. Major Fiberglass-epoxy 7.2-9.0
issues related to non-autoclave curing are the Tooling material
effective compaction of the composite plies, Graphite-epoxy 4.1-9.0
and the elimination of the trapped interlami- Cast ceramic 0.81
nar or intraply air. Tool steel 11.3
Iron (electroformed) 11.9
Nickel (electroformed) 12.6
26.4 EFFECTS OF TOOLING AND PART High-temperature cast epoxy 19.8
SHAPE Aluminum 23.2
Silicone rubber 81-360
Properly designed tools that produce accept-
able parts on a reproducible basis are a must
when fabricating composite structures. The
tool design requires the consideration of as
26.4.1 TOOLING FOR AUTOCLAVE MOLDING
many factors as are studied in the design of
the part itself. The main requirement for the The traditional autoclave molding process
tools is to maintain proper geometric dimen- uses a vacuum bag to impose a pressure dif-
sional stability and surface profile during the ference on the composite lay-up. A typical
compression and thermal cycling processes. bagging system consists of the following steps
On the other hand, the tool must also be (Schwartz, 1983).
590 Consolidation techniques and cure control

1. Cover the lay-up with a perforated parting

Graphite-epoxy laminate~
film or separator. Then lay up a layer or lay-
ers of bleeder material. The requirement of
the bleeder layers should be such as to
ensure adequate bleeding of air and excess
Sealing area
resin out of the part.
2. Place a strip of jute (vent material) just
beyond the edge of the lay-up and put bag-
sealing compound along the outside
perimeter.
3. Cover the lay-up, jute, and sealing com- Mold form/ Caul plate stop
pound with a flexible-film diaphragm and
seal the diaphragm to the mold with the
seal compound. CaP,\ (ReSin reservoir
4. Connect the vacuum lines and slowly apply .---------+~----.
the vacuum pressure while working the
wrinkles and excess air out of the lay-up,
bleeder material, and vacuum bag.
5. Check system for vacuum leaks. Mold half- -Mold half
6. Keep the part under vacuum while it is
waiting to be cured in the oven or auto-
clave.
To prevent surface irregularities on the bag
side (untooled surface) of the parts, a caul
plate may be used. The sole purpose of caul
plates is to improve the visual appearance of
the parts. They do not control part thickness. A
flexible caul plate with a thermally stable rub- \..cap
ber such as silicone or a fluoroelastomer is
often used to accommodate the surface geom- Fig. 26.13 Example of autoclave tooling (Borstell
and Turner, 1987).
etry. Figure 26.13 shows examples of autoclave
tooling setups with caul plates.
The three issues related to the tooling introducing a thermal strain. As the part and
design (Borstell and Turner, 1987) are thermal tool cool down from the gel temperature, the
expansion correction, coordinating the loca- tool usually shrinks more than the part. As an
tion of partial plies and use of caul plates. alternative, graphite-epoxy molds are used in
Because of the low coefficients of thermal some applications. Although some data has
expansion of composites when compared with been published, not all composite materials
metal tooling materials, thermal strain or have been measured. One empirical method is
stress must be considered for a curing process. to cure a representative panel on a plate of the
In the autoclave, the temperature at which the specified tooling material using the specified
resin solidifies is the gel temperature. At that cure cycle. Corrections can be estimated by
specific temperature, the part is the same size comparing the difference between the mold
as the thermally expanded mold. At a temper- and part dimensions. Another recommended
ature above the gel temperature, the tool empirical correction method is to correct steel
expands more than the partially cured part or nickel tools by making the tool 0.999 of the
 Effects of tooling and part shape 591

engineering dimension, and to correct alu- aluminum. During autoclave curing of com-
minum tools by 0.998. For example, a 2540 mm posite parts, the thermal uniformity is
(100 in) dimension is tooled to be 2537 mm excellent with rapid heat-up and cool-down
(99.9 in) for the steel tool. These corrections are rates. It is easy to handle and transport
needed to ensure an acceptable fit of mating because of its light weight. It also offers out-
composite parts. standing durability because the mold surface
Most parts contain partial plies to accom- resists cutting or impact damage and is not
modate local areas of increased stress. Several thermally degraded. When damaged, it is easy
techniques are used to control the location of to repair by welding, soldering, silver-solder-
partial plies, including polyester film tem- ing, or selective plating. It can provide
plates, slotted templates, and rails and complex contours without expensive machin-
banking surfaces. These tools serve as supple- ing. With most resin systems, it shows good
mental guidance to position the partial plies in release properties.
the lay-up process. Figure 26.14 shows the procedures of mak-
Typical cases of applying a caul plate are to ing an electroformed nickel tool. As in some
control the edge of a panel or the flanges of other types of tooling, constructing a model of
channels. The design of the metal caul plates the part surface is the first step in creating an
must take into account the fact that the matrix electroformed mold. The models are the same
resin melts in the autoclave to a very low vis- net dimensions as the required nickel mold.
cosity. The caul plate performs by pushing Compensation may be required when the coef-
excess resin sideways. Thus the rigid metal ficient of thermal expansion of the composite
caul plates must have high rigidity so that part differs greatly from that of the nickel
they do not deflect under autoclave pressure mold. Models are made from plaster, epoxy-
at curing temperature. The thickness of the faced plaster, fiberglass, fiberglass-epoxy,
caul plates can be calculated by use of the wood or other materials. From the model a
equations for unsupported bending beam reverse mandrel 'splash' is generally fabri-
analysis. The deflection of the caul plate can be cated from epoxy-faced fiberglass or plaster.
estimated using the balance condition of resin The mandrel to be used in electroforming is
pressure and applied force (Gutowski and Cai, then copied from the 'splash', although the
1988). The caul plate deflections should be lim- model can be used as the mandrel if it is pre-
ited to half the tolerance permitted in the part. pared correctly. The comers of the mandrel
should be designed to have radii in excess of
0.76 mm (0.030 in) to avoid thin spots in the
26.4.2 ELECTROFORMED NICKEL TOOLING
deposit. Draft and taper should be designed
An electroformed nickel tool consists of a into the mandrel to facilitate its removal from
4.6-6.4 mm (0.18-0.25 in) thick electrode- the electroform. Sharp comers or narrow, deep
posited mold surface that is supported by a grooves should be avoided if possible. The
simple steel substructure. The mold surface is mandrel can be fabricated from epoxy-faced
produced by the electroplating process fiberglass, rubber, or other materials. The sur-
(Sheldon, 1987). face of the mandrel is made conductive by
The electroformed tooling concept offers proper coatings. The back of the mandrel must
numerous advantages. The size of the mold is be reinforced to keep the mandrel from dis-
restricted only by the size of the electroform- torting during the electroforming process.
ing tank. The cost of producing duplicated Electroforming is the process of producing
tools is low. The mold surface is very smooth an article by electrodeposition of a metal onto
and scratch resistant. The coefficient of ther- a conductive mandrel surface. An anode sus-
mal expansion is approximately 40% less than pended in an aqueous electrolyte is connected
592 Consolidation techniques and cure control

Model Splash Fiberglass plating mandrel Mold electroformed

I

- -
Plated mold and tool Mold and structure Plating mandrel
support structure joined removed

Fig. 26.14 Example of electroformed nickel tooling (Sheldon, 1987).

to the positive pole of a DC electric source, and parts. These include low coefficient of thermal
the mandrel (cathode) is connected to its neg- expansion, ease of preparation, low density,
ative pole. The flow of electricity or electrons and thermal stability (Harmon, 1987). Their
results in the oxidation of a nickel anode to disadvantage is that they are less durable than
nickel ions and the reduction of nickel ions to metal tools.
nickel metal at the cathode (mandrel). The Composite tool making starts with a master
typical rate of growth is approximately model, usually built with plaster or hard-
0.01~.025 mm (0.0005-0.001 in) per hour. wood. The master models require proper
When the electroform is removed from the drying, sealing, and coating with mold release.
mandrel, its surface is a mirror image of the Then lay-up can be done directly on the plaster
surface of the mandrel. A natural physical or wood master. Liquid gel coats are required
characteristic of electrodeposition is that elec- to obtain a high fidelity surface on tools cured
tric current will tend to localize the deposit on by the vacuum bag process which does not
all edges and corners, causing an uneven generate enough pressure to ensure a void-free
thickness on the electroform. However, there surface, but may not be required on tools
are a variety of techniques to offset this effect. cured by the autoclave process which does
After the desired mold thickness is provide sufficient positive pressure. Prepregs
obtained, the mold is removed from the tank, with light weight fabrics are used directly
cleaned and the steel back-up structure is against the tool surface, while prep regs with
attached. The nickel mold is then polished to heavier fabrics are used to build up the thick-
the required finish, and ready for use. ness. During the lay-up, care should be taken
to work each ply into all radii and comers and
to remove all entrapped air. Debulking is
26.4.3 GRAPHITE-EPOXY TOOLING
applied after the lay-up, either with a vacuum
Composite tools have definite advantages bag setup or with assistance of an autoclave
over metal molds for large or highly contoured for a pressure debulk, to consolidate the plies
 Effects of tooling and part shape 593

and remove all entrapped air. The curing In thermal expansion molding, two basic
process is done with a vacuum bagging sys- methods are employed: the trapped or fixed-
tem or with an autoclave. With the tool still on volume rubber method and the
the model, the support structure, either a solid variable-volume rubber method. Figure 26.16
laminate or an 'egg-crate' panel is attached to shows the setup for both methods. The fixed-
the tool by means of locally applied fabrics, volume method exploits the large difference
room-temperature curing, and high-tempera- between the coefficient of thermal expansion
ture resistant resins. Once the support of the elastomer and that of metals. The elas-
structure is cured to the laminate shell, it is tomer is confined within a closed metal tool
removed from the master. Care should be
taken to avoid damaging either the tool or the
master. Figure 26.15 illustrates the
graphite-epoxy tooling making process.
Composite tools are being used successfully ____ Rubber tool sized to fill
/ the cavity in the part
throughout the aerospace industry to produce
parts that are structurally reliable, repro-
ducible, and dimensionally accurate.

Teflon separator film Breather cloth

/OuterbOX

Ib) Floating·plate
- - - pressure control

Rubber tool projects above the

- - - part so excess pressure is vented
by forcing the floating plate
to the bag.

Fig. 26.15 Example of graphite-epoxy tooling

(Harmon, 1987).
.........Outer box
26.4.4 ELASTOMERIC TOOLING

Elastomeric tooling or rubber tooling can be

used to generate molding pressure or to act as
a pressure intensifier. In thermal expansion
molding, elastomeric tooling is constrained
within a rigid frame to generate consolidation Fig. 26.16 Example of elastomeric tooling (Foston
pressure by thermal expansion during the cur- and Adams, 1987) (a) fixed volume method;
ing cycle (Foston and Adams, 1987). (b) variable volume method.
594 Consolidation techniques and cure control

cavity. When heated, it expands into the cavity, Impregnated Composites, Proc. 9th Int. Conf.
exerting the pressure required to compact a Composite Mater. (ICCM-9), 1993, 3, 575-583.
composite laminate. The variable-volume Dave, RS., Kardos, J.1. and Dudukovic, M.P., A
Model for Resin Flow During Composite
method offers more flexibility and control than
Processing, Part 1: General Mathematical
the fixed-volume method because a precisely Development, Poly. Composites, 1987, 8(1),
calculated volume of rubber is not normally 29-38.
required. In most applications, the rubber is Dave, RS., Kardos, J.1. and Dudukovic, M.P., A
simply 'set back' to allow for the bulk factor of Model for Resin Flow During Composite
the molding material during assembly of the Processing, Part 2: Numerical Analysis for
tooling details. A floating plate is used for the Unidirectional Graphite/Epoxy Laminates,
Poly. Composites, 1987, 8(2), 123-132.
pressure control. Dave, RS., Mallow, A, Kardos, J.1. and Dudukovic,
Thermal expansion molding with elas- M.P., Science-based Guidelines for the
tomeric tooling has been successfully used on Autoclave Process for Composites
commercial aircraft parts such as rudders and Manufacturing, SAMPE J., 1990,26(3),31-38.
spoilers (Schneider and Carroll, 1987). This Dusi, M.R., Lee, W.I., Ciriscioli, P.R., and Springer,
reduces the number of detail parts fabricated G.S., Cure Kinetics and Viscosity of Fiberite 976
and the need for bonding and mechanical fas- Resin, J. Composite Mater., 1987,21(3), 243-261.
Foston, M. and Adams, R.C, Elastomeric Tooling,
tening on assembly, thereby effecting
in Engineered Materials Handbook, Vol. 1:
significant reductions in production time and Composites, ASM International, 1987, pp.
cost. 590-594.
Gibson, R.E. and Hussey, M.J.1., The Theory of
REFERENCES One-Dimensional Consolidation of Saturated
Clays, Geotechnique, 1967,17,261-273.
Batch, G.L. and Macosko, CW., A Model for Two- Gutowski, T.G., A Resin Flow/Fiber Deformation
Stage Fiber Deformation in Composite Model for Composites, SAMPE Quarterly, 1985,
Processing, Proc. 20th Intern. SAMPE Tech. 16(4),58-64.
Conf., September 1988, pp. 641-650. Gutowski, T.G., Morigaki, T. and Cai, Z., The
Bhi, S.T., Hansen, RS., WIlson, B.A, Calius, E.P., Consolidation of Laminate Composites, J.
and Springer, G.S., Degree of Cure and VIScosity Composite Mater., 1987, 21,172-188.
of Hercules HBRF-55 Resin, Proc. 32nd Intern. Gutowski, T.G., Cai, Z., Bauer, S., Boucher, D.,
SAMPE Symp. Exhib., Vol. 32., 1987, pp. Kingery, J. and Wmeman, S., Consolidation
1114-1118. Experiments for Laminate Composites, J.
Biot, M.A, General Theory of Three-Dimensional Composite Mater., 1987,21,650-669.
Consolidation, J. Appl. Phys., 1941,12,155-164. Gutowski, T.G. and Cai, Z., The Consolidation of
Biot, M.A., Theory of Elasticity and Consolidation CompOSites, in The Manufacturing Science of
for a Porous Anisotropic Solid, J. Appl. Phys., Composites, Proc. Manufacturing International
1955,26(2),182-185. 88, Vol. IY, (ed T.G. Gutowski), 1988, pp.13-25.
Biot, M.A , General Solutions of the Equations of Halpin, rC, Kardos, J.L. and Dudukovic, M.P.,
Elasticity and Consolidation for a Porous Processing Science: An Approach for Prepreg
Material, J. Appl. Mech., 1956, March, 91-96. Composite Systems, Pure Appl. Chern., 1983,55(5).
Borstell, H. and Turner, K.T., Tooling for Autoclave Harmon, B.D., Graphite-Epoxy Tooling, in
Molding, in Engineered Materials Handbook, Vol. Engineered Materials Handbook, Vol. 1: Composites,
1: Composites, ASM International, 1987, pp. ASM International, 1987, pp.586-589.
578-581. Kardos, J.1., Dudukovic, M.P., McKague, E.L. and
Cai, Z. and Gutowski, T.G., Fiber Distribution and Lehman, M.W., Void Formation and Transport
Resin Flow in the Molding Process, Proc. 7th During Composite Laminate Processing: An
Int. Conf. Composite Mater. (ICCM-7), 1989,1, Initial Model Framework, in Composite
76--82. Materials: Quality Assurance and Processing,
Connor, M., Gibson, AG., Toll, S. and Manson, ASTM STP 797, (ed CE. Browning), 1983, pp.
J.AE., A Consolidation Model for Powder 96-109.
 References 595

Kardos, J.L., Dave, R and Dudukovic, M.P., Voids 110-118.

in Composites, in The Manufacturing Science of Loos, AC and Springer, G.S., Curing of
Composites, Proc. Manufacturing International Graphite/Epoxy Composites, Air Force
'88, Vol. IV, (ed T.G. Gutowski), 1988, pp. 41-48. Materials Laboratory Report AFWAL-TR-83-
Kim, T.W., Yoon, K.J., Jun, E.J. and Lee, W.I., 4040, Wright Aeronautical Laboratories, Wright
Compaction Behavior of Composite Laminates Patterson Air Force Base, Dayton, OH, 1983.
During Cure, SAMPE J., 1988, 24 (5), 33-36. Roylance, D., Reaction Kinetics for Thermoset
Kim, T.W, Jun, E.J. and Lee, WI., Compaction Resins, in The Manufacturing Science of
Behavior of Thick Composite Laminates During Composites, Proc. Manufacturing
Cure, Proc. 34th Inter. SAMPE Symp., 1989, 12 InternationaY88, Vol. IV, (ed T.G. Gutowski),
(1),17-19. 1988, pp. 7-11.
Kim, Y.R., McCarthy, S.P. and Fanucci, J.P., Schneider, CW and Carroll, H.E., Elastomeric
Compressibility and Relaxation of Fiber Tooling Application, in Engineered Materials
Reinforcements During Composite Processing, Handbook, Vol. 1: Composites, ASM International,
Polym. Composites, 1991, 12 (1), 13-19. 1987, pp. 595-601. '
Lam, RC and Kardos, J.L., The Permeability of Schwartz, M.M., Composite Materials Handbook,
Aligned and Cross-Plied Fiber Beds During McGraw-Hill, 1983.
Processing of Continuous Fiber Composites, Sheldon, D.L., Electroformed Nickel Tooling, in
Proc. Am. Soc. Composites, Third Technical Engineered Materials Handbook, Vol. 1: Composites,
Conf., Seattle, WAr 1988, pp. 3-11. ASM International, 1987, pp. 582-585.
Lam, RC and Kardos, J.L., The Permeability and Smith, G.D. and Poursartip, A, Comparison of Two
Compressibility of Aligned and Cross-Plied Resin Flow Models for Laminate Processing, J.
Carbon Fiber Beds During Processing of Composite Mater., 1993, 27(17), 1695-1711.
Composites, Proc. 47th Ann. Tech. Conf. Springer, G.S. and Tsai, S.W., Thermal
(ANTEC'89), SPE, New York, 1989, pp. Conductivities of Unidirectional Materials, J.
1408-1412. Composite Mater., 1967, 1, 166-173.
Lee, W.I., Loos, AC, and Springer, G.S., Heat of Springer, G.S., Resin Flow during the Cure of Fiber
Reaction, Degree of Cure, and Viscosity of Reinforced Composites, J. Composite Mater.,
Hercules 3501-6 Resin, J. Composite Mater., 1982,16,400-410.
November 1982, 16, pp. 510-520. Springer, G.S., Modeling of the Cure Process of
Lee, S.Y. and Springer, G.s., Effects of Cure on the Composites, SAMPE J., September/October
Mechanical Properties of Composites, J. 1986, pp. 22-27.
Composite Mater., 1988, 22(1), 15-29. Tang, J.M., Lee, WI. and Springer, G.S., Effects of
Lindt, J.T., Engineering Principles of the Formation Cure Pressure on Resin Flow, Voids, and
of Epoxy Resin Composites, SAMPE Quarterly, Mechanical Properties, J. Composite Mater., 1987,
October, 1982. 21, 421-440.
Lindt, J.T., Consolidation of Circular Cylinders in a Tsal, S.W and Hahn, H.T., Introduction to Composite
Newtonian Fluid, I. Simple Cubic Materials, Technomic Publishing, 1980.
Configuration, J. Rheology, 1986,30. Van Den Brekel, L.D., and De Long, E.J.,
Loos, AC and Freeman, Jr., W.T., Resin Flow Hydrodynamics in Packed Textile Beds, Textile
During Autoclave Cure of Graphite-Epoxy Research J., August, 1989, pp. 433-440.
Composites, High Modulus Fiber Composites in van Wyk, CM., Note on the Compressibility of
Ground Transportation and High Volume Wool, J. Textile Inst., 1946,37, T285-T292.
Applications, ASTM STP 873, (ed D.W Wilson), Williams, J.G., Morris, CE.M. and Ennis, B.C.,
1985, pp. 119-130. Liquid Flow through Aligned Fiber Beds, Polym.
Loos, AC and Springer, G.S., Curing of Epoxy Engng Sci., 1974, 14 (6), 413-419.
Matrix Composites, J. Composite Mater., 1983, 17,
135-169.
Loos, AC and Springer, G.s., Calculation of Cure
Process Variables During Cure of
Graphite/Epoxy Composites, Composite
Materials: Quality Assurance and Processing,
ASTM STP 797, (Ed. CE. Browning), 1983, pp.
COMPOSITE MACHINING 27
Kent E. Kahkonen and Nitin Potdar

27.J. INTRODUCTION 27.2 CONVENTIONAL MILLING

The processes used to manufacture composite When milling graphite-epoxy with polycrys-
structures generally require that trimming and talline diamond (PCD) the chips are formed as
other machining operations be performed small particles of powder dust and fumes. The
prior to assembly. Machining processes are surface roughness is a function of fiber orien-
required to produce accurate surfaces and tation, cutting direction and the angle between
holes to allow precision fitting of components cutting direction and fiber direction. The sur-
into an assembly. Due to shrinkage during the face may sometimes exhibit many small holes
curing stage of the composite structure it is not due to fiber pull out. When taking heavy
practicable to place holes in the part during milling cuts there is a greater tendency to
the molding stage, therefore milling, cutting, break corners as the tool exits the material so it
drilling etc. are considered a post cure opera- is advisable to first machine a step on the edge
tion. perpendicular to the final pass. A four fluted
Due to the toughness and abrasive nature of end mill will reduce cutting pressure on the
modern composites, there is a need for harder laminate and keep it cooler. Climb milling
and longer lasting cutting tools. A large data- helps prevent the fibers from separating from
base of machining information for various the matrix bond material.
high speed steel and carbide cutting tool Advantages of machining composites are:
materials exists for machining metal, wood
and some thermoplastics. However, much of • improved surface finish unless part surface
this data cannot be applied to machining mod- was directly in contact with the mold sur-
face;
ern composites. Modern composites like
graphite-epoxy, aramid-epoxy and carbon- • machined surfaces provide accurate mating
surfaces for parts to be assembled;
carbon each have their own machining charac-
teristics. Composites are not homogeneous or • eliminates the majority of the problems
associated with part shrinkage and insert
isotropic, therefore the machining characteris-
movement during the fabrication processes.
tics are dependent on the tool path in relation
to the direction of the reinforcing fibers. Tool life factors are:
Metals or metal alloys have nearly homoge-
neous properties throughout the workpiece, • PCD end milling cutters will perform sixty
but each material in a composite retains its to one hundred times longer than carbide;
individual properties. • cutting speed does not have a great effect
on the flank wear of PCD cutting tools.
With increased cutting speeds, the feedrates
Handbook of Composites. Edited by S.T. Peters. Published
can be increased and machining time
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 decreased;
 Mechanical drilling of composite materials 597

• cutting speeds range from 244 surface Depth of cut will vary depending on the
m/min (800 surface ft/min) to 762 surface thickness of the part and the amount of mate-
m/min (2500 surface ft/min) with PCD end rial to be removed.
mills;
• when cutting parallel to the fiber direction,
27.3.1 ADVANTAGES
the wear ratio on the cutting tool increases
compared with cutting perpendicularly to Computer numerical controlled lathes (CNC)
the fiber direction; can be used to machine simple to very complex
• surface finish remains below 20 Ra [arith- rotational parts. CNC machining produces
metical average roughness (see ISO R488)] accurate parts at a high production rate.
when cutting with PCD end mills and the
flank wear is approximately 0.127 mm
27.3.2 DISADVANTAGES
(0.005 in);
• the surface finish deteriorates above 150 Ra Delamination can also occur on a lathe (Fig.
when cutting with a carbide end mill and the 27.1), therefore the part may require a finish
flank wear has reached 0.127 mm (0.005 in); cut moving from the largest diameter to the
• roughing feedrates range from smaller diameter. Graphite chips are a serious
0.23 mm/rev (0.007 in/rev) to 0.38 mm/rev problem. The spinning chuck creates a fan
(0.012 in/rev) and finish feedrates range effect on the graphite particles. The exhaust
from 0.076 mm/rev (0.002 in/rev) to system must be adequate to control the
0.13 mm/rev (0.005 in/rev); graphite chips. Also, the machine ways and
• the depth of cuts should range from one the ball screws on the machine must have
quarter to one half of the diameter of the sealed protection to minimize wear. The com-
end mill cutter. Depth of cut will vary puter control requires protection from the
depending on the rigidity of machine ways, graphite chip particles.
spindle and workholding devices.
The disadvantages associated with milling 27.4 MECHANICAL DRILLING OF
of composites include controlling the graphite COMPOSITE MATERIALS
chips (dust particles), confining them to a small
area and having an adequate collection system. Drilling holes in composites can cause failures
A second problem is controlling the outer lay- that are different from those encountered
when drilling metals. Delamination, fracture,
ers of the composite so that the fibers will shear
instead of lifting up under the force of the cut-break-out and separation are some of the most
ting action and leaving extended fibers beyond common failures. Delamination (surface and
the cut surface. Also when cutting perpendicu- internal) is the major concern during drilling
lar to the lay of composite fibers, edge composite laminates as it reduces the struc-
break-out can occur. This can be controlled by tural integrity, results in poor assembly
designing a backup structure in the tooling. tolerance, adds a potential for long term per-
formance deterioration and may occur at both
the entrance and exit plane. Delamination can
27.3 CONVENTIONAL TURNING
be overcome by finding optimal thrust force
The turning of graphite composite is utilized (minimum force above which delamination is
to produce round surfaces that need to mate initiated). Figure 27.2 shows push out delami-
with either metal of graphite parts. The cutting nation at exit because at a certain point
speeds can be over 305 m/min (1000 ft/min) if loading exceeds the interlaminar bond
the part can be held securely and PCD tool strength and delamination occurs. Figure 27.3
inserts are utilized. shows peel-up delamination at entrance
598 Composite machining

t•
g Tool
Direction
(Correct)

Fig. 27.2 Drill bit showing push-out delamination

JJ~ at exit.
~J!bJ/'~J
graphite reinforcement materials. Each of
these materials requires individual attention
in the selection of cutting tool parameters. The
composite materials with metal backup panels
~
require separate drills with different geome-
~'~\'\'

Ol=!' n
~'\'\
tries. Cutting speeds and feedrates vary in

I2:zn
(Incorrect)
each of the various combinations of materials.
Secondary drilling or reaming operations are
required to hold tight tolerances or smooth
surface finishes on the holes. Table 27.1 shows
Fig. 27.1 Machining direction for turning compos- the drilling results when using four styles oj
ite parts on a lathe. drills.!
PCD tooling offers increased tool life, better
because the drill first abraded the laminate hole quality, consistent hole size and higher
and then pulled the abraded material away machining rates. Drilling and countersinking
along the flute causing the material to spiral
up before being machined completely. This
type of delamination decreases as drilling pro-
ceeds since the thickness resisting the lamina
bending becomes greater.
Among the variables to be considered for
tool selection include the thickness of material,
diameter of hole, tolerance requirements, hole
finish requirements and the composite mater-
Peeling
ial being drilled. Tungsten carbide, micrograin Action
tungsten carbide and drill tool materials are
used for drilling composite materials.
Some commonly used composites are
glass-epoxy, glass-graphite-epoxy, graphite-
epoxy, graphite-epoxy with aluminum
backup and graphite-epoxy with titanium
backup. Other materials include the aramids Fig. 27.3 Drill bit showing peel-up delamination at
(Kevlar®) with combinations of glass or entrance.
 Mechanical drilling of composite materials 599

Table 27.1 Summary of drill performance: mean hole quality measures as a function of point angle.
Maximum recorded values of response parameters are shown in box brackets, [J (Reproduced from Ref 1
by permission of ASM Materials Week)

Criterion/drill Dagger 8-Facet 4-Facet Master NAS 907-/ HSS

Exit breakout 1 2 3 4
(Rank: least = 1)
Panel damage, DRAr 1.96 (3.34) 2.37 (3.18) 2.75 (3.62) 3.63 (5.54)
Microcrack density 1 2 3 4
(Rank: lowest = 1)
Thrust force, 114 [166] 201 [378] 263 [428] 593 [969]
N (Ibf) (25.6 [37.4]) (45.3 [85.2]) (59.3 [96.3]) (133.5 [218])
Torque, 1.29 [2.18] 1.15 [2.0] 0.7 [1.64] 1.53 [2.2]
Nm (ftlbs) (0.95 [1.61]) (0.85 [1.5]) (0.50 [1.21]) (1.13 [1.61])
Surface finish, RA, 0.4 [1.6] 0.95 [2.2] 1.6 [3.0] 2.4 [4.12]
!lm (!lin.) (26 [64]) (38 [88]) (64 [122]) (96 [165])
Hole diameter, 6.354 [6.379] 6.356 [6.369] 6.367 [6.395] 6.375 [6.397]
mm (in) (0.25016 (0.25022 (0.25067 (0.2510
[0.25115]) [0.25075]) [0.25175]) [0.25185])
Hole out-of-roundness,
mm (in) 0.0061 [0.025] 0.003 [0.005] 0.0043 [0.018] 0.013 [0.03]
(0.00024 [0.0010]) (0.00012 [0.0002]) (0.00017 [0.0007]) (0.00051 [0.0012])
Drill point angle, deg. 30 24,118 140 135

with a combination tool provides better hole (CFRP) and glass fiber-epoxy (GFRP) lami-
quality. Tool life is normally determined by the nates using HSS and carbide tipped drills
extent of delamination and fiber break out. For made the following observations. Both chisel
machining graphite composites with or with- edge and flank wear increased on the carbide
out aluminum backing, PCD tooling is drill with a higher ratio of wear between 200
suggested with the same speeds and feeds and 400 holes (test sample 400 holes). The tool
used for machining graphite composites with- wear was greater in the CFRP laminates due to
out any backing. the abrasive nature of carbon fibers. Flank
For machining graphite composites with wear is more pronounced in GFRP when the
titanium backing, it is not recommended that feed was increased and the same effect is
the same drill be used for both the titanium noted when speed is increased. The HSS drills
and graphite sections. Initially a hole should lasted for ten holes in the graphite and twenty
be drilled up to the titanium layer with a holes in the glass.
hydraulic depth sensing device at high speeds
and feed. A second drill with lower speed and
27.4.1 DRILL GEOMETRY
feed for machining titanium should be used.
Finally finish reaming operation and counter- Drill point geometries influence the torque
sinking should be performed for assuring hole requirements. Lip relief and rake angles are
quality. determined by the application. The dagger
A study carried out on carbon fiber-epoxy drill is ideal to machine graphite composites
600 Composite machining

as it eliminates breakout when exiting the be processed.

workpiece. The dagger drill has 35 0 included The grinding of polymer matrix compos-
point angle and a 121 0 chisel edge angle. Twist ites (PMC) has a number problems. For
drills with flute configuration to control metal example in the case of thermoplastic matrix,
chips are also used. Fully fluted drills with the surface of grinder becomes covered with
PCD tips brazed on a solid carbide shaft pro- melted thermoplastic. In the case of aramid
vide the strength of carbide and hardness of fiber it is hard to get a clean cut surface
diamond. Drill geometries are continuing to because the grains cannot abrade the aramid
be experimented with to find ways to elimi- fibers cleanly. Abrasive belts have been used
nate the problems associated with the hole on aramids with some success but dust collec-
making process in composites. tion has been a major problem.
Drill cutting parameters are:
• feedrates range from 0.025 mm/rev 27.6 MACHINING OF KEVLAR
(0.001 in/rev) to 0.063 (0.0025 in/rev);
• cutting speeds range from 30 surface Cutting, Trimming, Turning and Milling of
m/min (100 surface ft/min) to 460 surface Kevlar
m/min (1500 surface ft/min); Because of its inherent toughness, Kevlar is dif-
• high cutting speeds can bum the matrix ficult to cut, so sharp, heavy duty upholstery
material and reduce bond strength between scissors will cut up to 170 g/m2 (5 oz/yd2) fab-
the composite material and the matrix ric of Kevlar. Woven roving and heavier fabrics
material. can be cut using specially designed serrated
scissors. An overview of cutting and trimming
techniques and applications is shown in Table
27.4.2 COOLANTS
27.2. For more information on cutting and
A water soluble coolant forced through a cold machining of Kevlar refer to DuPont's
air blast unit is recommended when machin- Machining Handbook2•
ing most composite materials. However if the
composite is hydrophilic in nature then a cold
air blast unit in combination with dust or vac- 27.7 ABRASIVE WATER JET MACHINING
uum collection system should be used. Abrasive water jet (AWJ) is used for linear pro-
file cutting, turning, milling and drilling
operations in composite materials.
27.5 GRINDING COMPOSITE MATERIALS
Conventional tool machining is affected by
The grinding process has been used exten- fiber or particle reinforcements rather than the
sively for finishing composite golf shafts and matrix material while AWJ machining is not.
fishing rods. Five hundred parts per hour can To make a circular hole 6.35 mm (0.25 in) in
be produced on centerless grinders. Silicon diameter in aramid 3.18 mm (0.125 in) thick, it
carbide wheels are used with an open grain to takes about the same time for both conven-
reduce wheel galling. Surface speeds between tional as well as AWJ machining. The cutting
1219 surface m/min (4000 surface ft/min) and process parameters for AWJ include water jet
1829 surface m/min (6500 surface ft/min) can pressure, velocity, abrasive grain size, abrasive
be achieved. This equipment is specially material, standoff distance and jet impinge-
designed for grinding and finishing compos- ment angle. and some additional parameters.
ites. Grinding accuraces within 0.0127 mm Water jets without abrasive are also used for
(0.0005 in) can be achieved with centerless cutting soft composites. Figure 27.4 shows the
grinding. Both straight and tapered shafts can AWJ processes and machining parameters3 •
Table 27.2 Overview of aramid cutting/trimming techniques and applications2

Process Main applications

Type of laminate/composite Type of cuts
Outside Inside
Thin Medium Thick Sandwich Filament Straight Curved Curved Straight Curved Curved
< 1.5mm 1.5-4.0 mm >4.0mm structure wound (flat) (flat) (3D) (flat) (flat) (3D)
«0.060 in) (0.060-0.160 in) (> 0.160 in)
1.0 Routing
1.1 Split helix router 0 0 0 X X X X X X
1.2 Spiral shear router 0
• X X X X X X
2.0 Nibbling (X) (X)
3.0 Sawing
•
3.1 Band sawing 0 0 0 X (X)
(Conventional • •
blades)
3.2 High speed 0 0 0 X (X) (X) X (X) (X)
reciprocating
(Special blades)
3.3 Circular sawing 0 0 X (X) (X)
conventional blades • • :=t.
~
special blades 0 0 X (X) i::l
• <tl

4.0 Turning (parting tool) X ~.

5.0 Ultrasonic cutter 0

• X (X) (X) X (X) (X) ~
~
....
6.0 Laser beam 0 X X X (X) (X) (X)
• • "".....
7.0 Water jet 0 0 X X X (X) (X) (X) ~
• • ...,;::,
• Main application X Possible ;:r-
S·
o Very limited application - seldom used (X) Possible with limitations ~.
- Not possible
0'\
0
......
 Machining Parameters s
AWJ Parameters ()
o
Linear Cutting Drilling Milling Turning
if:w
u
~R jV ~
ROT I
WJD~
6-L I TDcf5dr-,
:ro- g.
;:i'
~.
~lII>dp sod ~ =-- n I
MT[ -'~T-R-~
db 'dt
£~ ~
--=- +-MTD WV~ I
Pressure
Waterjet Diameter Rotational Speed
Mixing Tube Length Angle Angle Traverse Rate Direction of Rotation
Mixing Tube Diameter Standoff Distance Standoff Distance Lateral Increment Angle
Abrasive Material Traverse Rate Dwell Time Number of Passes Traverse Rate
Abrasive Size Number of Passes Pressure Profile Number of Sweeps Initial Diameter
Abrasive Flow Rate Material Thickness Material Thickness Final Diameter
Abrasive Condition Depth of Cut

Machining Requirements (Dependent Variables)

Traverse Rate Diameter Volume Removal Rate Turned Diameter

Surface Finish Drilling Time Depth Control Surface Finish
Width of Cut Hole Shape Machining Time
--- ---

Fig. 27.4 AJW process and machining parameters. (Reproduced from Proc. Machining of Composites Symp., Chicago, Ill. 1-5 Nov 1992, by
permission of the publisher, ASM Materials Week.)
 Abrasive water jet machining 603

27.7.1 LINEAR CUTTING WIlli AWJ 27.7.2 TURNING

Linear cutting is used to trim composite parts In turning with AWJ, a workpiece is usually
and to cut profile shapes on the inside of a rotated while the jet is fed along all three axes.
part. The cut surface is normally smoother The material encountered by the jet is abraded
near the entrance surface then becomes wavy away in the form of a very fine debris. Higher
in the lower half of the cut toward the exit sur- jet pressure produces a smoother surface with
face. In general, the composite material is a higher material removal rate. Higher tra-
sheared away by a high velocity abrasive verse rates combined with multiple passes are
grain. The width of cut (kerf) decreases as the more efficient than deeper cuts with lower tra-
feedrate increases and the waviness increases verse rates. Surface finish is affected by
as the feedrate increases3 • unsteadiness in traverse rate or abrasive flow
Table 27.3 shows some of the observations rate. The repeatability and accuracy of the
made by Hashish3 • The maximum cutting tra- AWJ turning process depends on control and
verse rate is primarily controlled by the matrix steadiness. A 10% variation in rotational speed
material. does not affect the surface waviness but a tra-
Table 27.4 shows results for some compos- verse rate variation over 4% will significantly
ites with different speeds. affect the surface waviness. Some methods to
improve surface finish are:

Table 27.3 Typical through-cutting traverse rates (in mm/ s) with AWJs for different composites3

Material Thickness (mm)

0.79 1.60 3.18 6.36 12.7 19.1 50.8
Organic matrix composites:
Plastic and composites 53 38 29 21 15 10 2.5
Carbon-carbon composites 42 32 22 13 7.5 4 0.85
Epoxy-glass composites 105 95 76 42 17 12 15
Graphite-epoxy composites 74 63 52 40 15 10 4.2
Kevlar (steel reinforced) 42 25 17 8.5 4.7 2.5 0.63
Cutting conditions: p = 345 MPa, dn = 0.299 nun, dm =0.762 nun, garnet mesh 80
SiC abrasives

Table 27.4 Surface waviness and corresponding cutting traverse rates (in mm/s) for some composite
materials3

Material RMS surface waviness (jlm)

1.9 2.5 3.8 5 6 8
AlP/SiC (20%) 6.35 mm 0.29
Toughened zirconia (6.36 mm) 0.15 0.2 0.4 0.5
Mg/B4C (15%) 6.36 mm 3 6 8
Graphite-epoxy composites (3.18 mm thick) 4 8 12 20 30
Graphite-epoxy composites (18.5 mm thick) 0.6 0.85 1.7 2.5 3.4 4.25
Cutting conditions: p = 370 MPa, dn = 0.299 nun, dm 0.762 mm, garnet mesh 80
604 Composite machining

• multipassing by traversing the jet without velocity decreases as the depth increases which
lateral feed; can be attributed to the effect of return flow
• use of finer abrasive and increasing number which reduces particle velocity and interferes
of passes; with the impact process. Pressures of 30-40
• to improve surface roughness, use softer MPa are common for piercing glass. High pres-
abrasives like silica sand, copper slag etc; sures are necessary to pierce brittle or
• finishing with slurried abrasive yields laminated composites. The higher pressures
improvement in surface roughness. may cause the following problems: fracture
due to shock loading of water; hydrocracking
due to hole hydrodynamic pressurization;
27.7.3 MILLING
delamination due to loading.
The main objective of AWJ in milling is to pro- Holes larger than the piercing diameter of
duce a cavity with controlled depth. In this the AWJ are first pierced, then profile cut to the
method, the jet material interaction is the depth finished diameter being offset by the kerf
determining factor. The production of kerf amount. Hole shape variance depends on mix-
irregularity can be reduced by manipulating ing tube length, target material, standoff
one of the factors, such as traverse rate, increas- distance, depth of hole and dwell time in the
ing the stand off distance or angling the jets. cut. Mixing tube length is important when
To mill square pockets the traverse speed drilling materials with high resistance.
can be varied rather than angling the water jet Increasing the mixing tube length improves
head. In this case the nozzle can be manipu- the distribution of the abrasive with the water
lated over the workpiece with an oscillatory jet. This produces holes that are straighter and
drive using a motor and an eccentric. A uni- rounded.
form traverse rate and exposure time can Advantages offered by AWJ are:
produce a uniform depth cavity. A hard mate-
rial pattern with the shape to be milled can be • suitable for wide range of composites;
used to mask the target surface. This way the • can perform many operations like turning,
mask will allow jetting in the traverse zones drilling and milling;
where the traverse rate is uniform. Surface fin- • no thermal stresses;
ish variations can be achieved by using • high as well as low material removal rates;
different abrasive materials or grit sizes. • no heavy clamping of workpieces required;
Harder abrasives can be used for higher mate- • omnidirectional machining;
rial removal rates and softer abrasives for • process can be automated;
finishing operations. • optimal range of parameters available to pre-
vent delamination, loading and splintering;
• fine holes of 0.5 mm (0.012 in) can be
27.7.4 DRILLING drilled.
Hole drilling can be performed in any of the Disadvantages:
following ways depending on the diameter
and accuracy of the holes: piercing is suitable • dimensional accuracy is low;
for small diameter holes; kerf cutting is suit- • temperature rise in cutting region may be
able for large diameter holes; milling is observed;
suitable for blind holes. • limited data is available with respect to
Techniques of hole piercing vary for each applications in metal and ceramic compos-
composite material. Piercing glass, acrylic and ites;
polycarbonate show that the general geometri- • not suitable for materials that are
cal features of pierced holes are similar. Particle hydrophilic in nature.
 Ultrasonic machining 605

27.S LASER MACHINING OF COMPOSITES Advantages:

Lasers are used in various industrial applica- • superior quality edges due to high temper-
tions such as drilling, cutting, welding and atures;
heat treatment of metals, etc. In composites, • vaporization of the material in cut zone;
polymer matrix materials are most suited for • extremely localized action;
laser cutting. Laser cutting is a non-contact • sealing of the edge in the cut zone;
ablation process in which efficiency is deter- • pulsed CO2 has been demonstrated as the
mined by thermal properties of the workpiece best laser for processing Kevlar composites.
material. Two types of laser have been used in
Disadvantages:
industry: Nd-YAG solid state laser and CO2
gas laser. The Nd-YAG laser operates in the • beam divergence after its focal point;
near infrared (IR) region of the spectrum while • material thickness of about 9.5 mm (% in) is
CO2 gas laser operates in the far infrared the maximum thickness that can be cut with
region. The Nd-YAG IR region wavelength is 1500W;
not absorbed by glass and many plastics while • heat affected zone of varying dimensions.
the CO2 far IR region wavelength is.
Applications of Nd-YAG solid state lasers 27.9 ELECTRIC DISCHARGE MACHINING
extend from drilling fine holes in jet engines to (EDM)
welding implant devices for the medical
industry. It has been determined that the Nd- Advanced composites can be cut by EDM as
YAG laser is very effective in cutting there is no physical contact between the elec-
graphite--epoxy composite materials. The high trodes or workpiece and the tool. In order to
power short pulses achieved with this laser EDM a composite, it should have an electrical
vaporizes both the graphite and epoxy matrix resistivity of less than 1-30hm/m. Polymer
before the epoxy resin can be overheated. matrix composite manufacturers can add a
The CO2 gas laser applications extend from small amount of copper in the matrix of the
drilling holes in baby bottle nipples to welding product to allow shaping by EDM. EDM can
automotive components in assembly lines. be used with conductive silicides, borides, car-
CO2 lasers operate in either continuous wave bides, etc. The EDM process is more accurate
or pulsed mode. Pulsed mode is preferred than AWJ machining. Small holes of 0.25 mm
because of high powers obtained and cool (0.01 in) diameter can be drilled in SiC/TIBz
down time. Aramid fiber reinforced plastic composites. The EDM process is found to be
(AFRP) has been cut very effectively by the slow for many production applications.
CO2 lasers. The general characteristics of a
laser cut zone in composite materials are 27.10 ULTRASONIC MACHINING
shown in Fig. 27.5.
Ultrasonic machining (USM) incorporates a
The charred layer which includes a zone
tool vibrating at 20 kHz and abrasive in a
with fibers protruding from the matrix and as
slurry to perform impact grinding of brittle
outer darkened zone in which the matrix has
undergone some degradation4,5. materials. This technique is particularly useful
for machining of ceramic matrix composites
Figure 27.6 shows the relationship between
that are difficult to process by conventional
kerf width and cutting speed. For three dimen-
methods. USM is a mechanical material
sional (3D) machining two laser beams are
removal process best suited for machining
directed through an optical assembly to inter-
brittle materials like glass, ceramics, graphite
sect in the plane of work piece to cut shoulders
and ceramic matrix composites. The process is
and vee grooves.
limited to workpieces of size below 100 mm
606 Composite machining

r (
"'\ r '\
'\ ( "\ "'\
"'\ (
r r
) ( '\ '\ ( (
TOP VIEW

CHARRED LAYER
PROTRUDING FIBRES

beam entry side

\
CROSS SECTION

U
IL_ _ _ _ _ beam exit side

Fig. 27.5 Schematic of FRP laser cut. (Reproduced

by permission of Marcel Dekker Ltd.) WI: kerf
width at the beam entry side; Wb: kerf width at the
beam exit side (the speed at which Wb = 0 defines ' ,0.-<0_-------......------,-----,
Y-" w. Powtt
the cutting speed limit for through cuts); Wf: length o
ao
•
•
!oow
~w
of protruding fiber; Wd: width of heat affected zone ()
..
•
200W
150W
0.8
(HAZ).
E
E
- 0.8

~
'j
't: 0,'
CD
:lI::

02

Fig. 27.6 Kerf width as a function of cutting speed

for (0/90)20 laminates. (Reproduced by permission
of Marcel Dekker Ltd.) Cutting speed (mmls)
 Ultrasonic machining 607

(3.94 in) because of the limitation on the size of Sonotrode (tool) material
the tool. Some of the variables that influence
Tools with diamond tips have good material
USM for close tolerances are as follows:
removal characteristics and very low wear but
are difficult to machine. Table 27.7 shows accu-
Abrasive type and size racy results of using a non-rotating steel
sonotrode.
Abrasives contained in the slurry do the actual
machining so they must be selected on the
basis of the workpiece material and the surface Ultrasonic vibrations
quality needed. As in the case of AWJ, larger The ideal condition would be the amplitude of
abrasive grains give higher material removal ultrasonic vibration to be equal to the grain
rates and rougher surfaces. The grain diameter mean diameter. If the amplitude is too small
cannot be larger than amplitude of the the abrasive cannot enter the machining gap, if
sonotrode as this would inhibit the injection of too large it causes the grains to be incorrectly
the grains to the machining gap. Common projected. A mixture of both the types of abra-
types of abrasive used are ~03 oxide, SiC, BC sive may be used and a suitable amplitude
and diamond. selected to determine which size grain enters
Table 27.5 shows recommended abrasive the machining gap.
for various materials. The grain diameter
affects surface roughness, overcut and
Surface area
machining rates. When high removal rates are
necessary with no high surface quality This factor influences removal rates and tool
required, 180-280 mesh abrasive do the job. wear. With a small diameter, higher feed rate
For finer surface finish 320-600 mesh abrasive is obtained but also higher tool wear is
is recommended. Table 27.6 shows surface noticed. This can be overcome by using a dia-
roughnesses for different workpiece materials. mond tool or with a closed loop force sensitive

Table 27.5 Recommended abrasive for various materia1s6

Material Recommended abrasive

Graphite Silicon carbide
Zirconia Silicon carbide or boron carbide
Ceramic matric composites Silicon carbide
Metal matrix composites Boron carbide

Table 27.6 Surface roughness for various materia1s6

Workpiece material Surface roughness Ra

(pm)

Graphite 1-2
Zirconia 0.75
Ceramic matrix composites 0.70
Metal matrix composites 0.90
608 Composite machining

servo system maintaining accurate machining on prepreg materials like glass fiber, carbon
pressures. Table 27.8 shows typical ultrasonic fiber and Kevlar with reduced fiber damage.
machining rates for a variety of materials6. Advantages:
USM is used in applications like drilling
aerospace cooling holes in ceramic matrix • conductive and nonconductive materials
composite turbine blades, slotting, irregular can be machined;
configurations in ceramics and composites, • material hardness is not so important;
machining of phased array radar components, • there are no chemical or electrical alter-
cutting tool inserts, superconductors, wire ations in the workpiece;
draw dies and extrusion dies. A CNC USM can • 3D and complex shapes can be machined
cut through 6 nun (0.24 in) thick composite easily and quickly;
layers and produce a controlled depth up to • no heat affected zone.
50 nun (1.97 in). The latter is important, as Disadvantages:
many composites have backing sheets that
should not be damaged. The ultrasonic action • amplitude of ultrasonic vibrations are very
reduces the amount of force required to sever important for proper machining;
the hard materials. This results in a better cut • limited sizes can be machined.

Table 27.7 Accuracy results with a non-rotating steel sonotrode6

Material Inlet diameter Outlet diameter Taper Roundness

(mm) (mm) (%) (mm)
Graphite 10.23-10.25 10.07-10.10 3.00 0.03"
10.2~1O.29 10.02-10.05 2.70 0.03b
Metal matrix composite 10.20-10.24 8.87-9.92 9.00 0.04b
10.09-10.12 8.85-9.90 6.60 O.OSh

Ceramic matrix composite 10.11-10.15 10.00 3.50 0.04b

5.04 4.99 1.25 _c

Zirconia 5.05 4.85 5.50 _c

a Tool 1: Exponential, Diameter = 10 nun

b Tool 2: Exponential, Tube 0 = 10 nun, ID = 7 nun
c Tool 3: Exponential, Diameter = 5 nun

Table 27.8 Typical ultrasonic machining rates for a variety materials7

Drilling diameter = 5 mm Drilling diameter = 10 mm

Material Time Removal rate Time Removal rate

(min) (mm 3/min) (min) (mm 3/min)
Graphite 1 164 1.25 224
Ceramic matrix composite 3.5 39 5.6 50
Metal matrix composite 10 7.6 14 9.3
Zirconia 210 0.65 90 3.1
 References 609

REFERENCES Bhatnagar, N., Naik, N.K and Ramakrishnan, N.

Experimental investigations of drilling on CFRP
1. Mehat, M., Reinhart, T.J. and Soni, A-H., Effect
composites. Materials & Manufacturing Processes,
of fastener hole drilling anomalies on structural 1990, 5(4), 591-608
integrity of PMR-15/GR composite .laminates, Geskin, E.S., Tisminetski, L., Verbitsky, D.,
Proc. Machining of Composite Matertals Symp., Ossikou,V., Scotton, T. and Schmitt, T.
ASM Materials Week, Chicago, ill, 1-5 Nov.
Investigation of waterjet machining of compos-
1992.
ites. Proc. Machining of Composite Materials
2. Kevlar Cutting and Machining Handbook, E.1. Du Symp., ASM Materials Week, Chicago, illinois,
Pont de Nemours and Co.
1-5 November 1992.
3. Hashish, M. State of the art of abrasive waterjet
Hochegn, H., Puw, H.Y. and Yao, KC. Experimental
machining operations for composites. Proc.
aspects of drilling of some fiber-ren:uorced pl.as-
Machining of Composite Materials Symp., ASM
tics. Proc. Machining of Composite Matenals
Materials Week, Chicago, Illinois, 1-5
Symp., ASM Materials Week, Chicago, illinois,
November 1992. 1-5 November 1992.
4. Di Ilio, A., Tagliaferri, V. and Veniali, F. Krishnamurthy, R., Santhanakrishnan, G. and
Machining parameters and cut quality in laser
Malhotra, S.K Machining of polymeric compos-
cutting of aramid fibre reinforced plastics.
ites. Proc. Machining of Composite Materials
Materials and Manufacturing Processes, 1990, 5(4), Symp., ASM Materials Week, Chicago, illinois,
591-608.
1-5 November 1992.
5. Lemma, S. and Sheehan, B. Laser Machining of
Lubin, G., ed., Handbook of Composites, 1982, New
Composite Materials. Proc. Machining of
York: Van Nostrand Reinholt.
Composite Materials Symp., ASM Materials Week, Ramulu, M., Faridnia, M., Gargini, J. L. and
Chicago, Illinois, 1-5 November 1992.
Jorgensen, J. E. Machining of graphite/epoxy
6. Gilmore, R. Ultrasonic machining of composite
composite materials with polycrystalline dia-
materials, Proc. Machining of Composite Materials
mond (peD) tools. Trans. ASME, J. Engng Mater.
Symp., ASM Materials Week, Chicago, lll. 1-5 and Tech., 1991, 113, October.
November, 1992.
Zaring, K., Erichsen, G. and Burnham, C. Procedure
optimization and hardware improvements in
FURTHER READING abrasive waterjet cutting systems. Proc.
Machining of Composite Materials Symp., ASM
Bhattacharyya, D., Allen, M.N. and Mander,. S.J. Materials Week, Chicago, ill., 1-5 November
Cryogenic Machining of Kevlar ComposItes. 1992.
Materials and Manufacturing Processes, 1993, 8(6),
631,651
MECHANICAL FASTENING AND ADHESIVE 28
BONDING
D.W. Oplinger

28.1 INTRODUCTION are particularly severe because the load trans-

fer between elements of the joint have to take
It would be difficult to conceive of a structure
place over a fraction of the available area. For
that did not involve some type of joint. Joints
mechanically fastened joints in metal struc-
often occur at a transition between a major
tures, local yielding, which has the effect of
composite part, where most of the structural
eliminating stress peaks as the load increases,
performance is generated, and a metal feature,
can usually be depended on; such joints can be
which is introduced to allow for very high
designed to some extent by the 'P over X
localized bearing contact for which the com-
approach, i.e. by assuming that the load is
posite has inadequate strength or durability. In
evenly distributed over load bearing sections
aircraft such a situation is represented by artic-
so that the total load (the 'P') divided by the
ulated fittings on control surfaces as well as on
available area (the /X) represents the stress
wing and tail components which require the
that controls the strength of the joint. In
ability to pivot the element during various
stages of operation. Tubular elements such as organic matrix composites, such a stress
reduction effect is realized only to a minor
power shafting often use metal end fittings for
extent, and stress peaks predicted to occur by
connections to power sources or for articulation
at points where changes in direction are elastic stress analysis have to be accounted for,
especially for one-time monotonic loading.
needed. In addition, assembly of the structure
from its constituent parts will involve either
bonded or mechanically fastened joints or both. 28.2 MECHANICALLY FASTENED JOINTS
Joints represent one of the greatest chal- COMPARED WITH ADHESIVE JOINTS
lenges in the design of structures in general
and in composite structures in particular. The In principle, adhesive joints are structurally
reason for this is that joints entail interruptions more efficient than mechanically fastened
of the geometry of the structure and often joints because they provide better opportuni-
material discontinuities, which almost always ties for eliminating stress concentrations; for
produce local highly stressed areas, except for example, advantage can be taken of ductile
certain idealized types of adhesive joint such response of the adhesive to reduce stress peaks.
as scarf joints between similar materials. Stress Mechanically fastened joints tend to use the
concentrations in mechanically fastened joints available material inefficiently and are charac-
terized by sizeable regions where the material
near the fastener is nearly unloaded, which
Handbook of Composites. Edited by S.T. Peters. Published
must be compensated for by regions where
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 high stresses occur to achieve a particular
 Mechanically fastened joints 611

required average load capacity. As mentioned composites first came into use. It was found
above, certain types of adhesive joint, namely early in this period that the behavior of com-
scarf joints between components of similar posites in bolted joints differs considerably
stiffness, can achieve a nearly unifonn stress from that which occurs with metals, primarily
state throughout the region of the joint. because stress concentrations are much more
In many cases, however, mechanically fas- of a factor in joint behavior of composite struc-
tened joints cannot be avoided because of tures, and stress analysis to quantify the level
requirements for disassembly of the joint, for of various stress peaks is more important. It
replacement of damaged structure, or to was fortunate that significant computing
achieve access to underlying structure. In addi- power became available in this period to keep
tion, adhesive joints tend to lack structural up with the need for the intimate details of
redundancy, and are highly sensitive to manu- stress conditions around mechanical fasteners.
facturing deficiencies, including 'poor bonding The current approach to the design of
technique, poor fit of mating parts and sensitiv- mechanically fastened joints in composite
ity of the adhesive to temperature and structures evolved mainly out of a number of
environmental effects such as moisture. 000, NASA and associated university pro-
Assurance of bond quality has been a continu- grams aimed at providing a methodology
ing problem in adhesive joints, primarily which could be applied routinely to aircraft and
because, while ultrasonic and X-ray inspection other applications. Numerous stress analysis
may reveal gaps in the bond, there is no present approaches to the mechanics of fastened joints
technique which can guarantee that a bond have been conducted over the years since the
which appears to be intact by ultrasonic or X- introduction of 'advanced' composites in the
ray inspection does not lack load transfer mid-1960s. These have included: the work of
capability, because of such factors as poor sur- Waszczak and Cruse1 based on the boundary
face preparation. Surface preparation and integral method; the use of two-dimensional
bonding techniques have been well developed, complex variable elasticity solutions which
but the possibility that lack of attention to detail treated the problem of variable contact around
in the bonding operation may lead to such defi- the fastener-5; as well as recent work reported
ciencies needs constant alertness on the part of by Madenci and llieri7; and a number of finite
those responsible for the bonding. Thus element approaches, especially the work of
mechanical fastening tends to be preferred over Crews and Naik8 which featured an inverse
bonded construction in highly critical and method for dealing with the contact problem.
safety rated applications such as primary air- Hart-Smith9 developed an analytic approach
craft structural components, especially in large based on the use of available solutions for
commercial transports, since assurance of the isotropic plates with bolt-loaded holes as well
required level of structural integrity is easier to as unloaded holes in plates under tension or
guarantee in mechanically fastened assemblies. compression which came out of classical efforts
Bonded construction tends to be more preva- such as those reported in PetersonlO• The latter
lent in smaller aircraft however and for provided simple functional descriptions of the
non-aircraft applications as well as in non-flight effect of joint geometry on peak stresses which,
critical aircraft components. with various empirically derived correction fac-
tors introduced by Hart-Smith9, provided
valuable insight into a number of trends in joint
28.3 MECHANICALLY FASTENED JOINTS
behavior.
Mechanically fastened joints for composite In addition to the analytic efforts, several
structures have been under study since the fairly extensive programs aimed at the devel-
mid 1960s when high modulus, high strength opment of design approaches for structural
612 Mechanical fastening and adhesive bonding

systems have been supported by 000 and 28.3.1 KEY FEATURES OF MECHANICALLY
NASN,6. Numerous papers have been pre- FASTENED JOINTS
sented over the years in a series of
Figure 28.1 represents a generic single fastener
DoD/NASA/FAA Conferences on Fibrous
joint, while Fig. 28.2 depicts a multiple fas-
Composites in Structural DesignS, 11-13. Many of
tener configuration. Many of the most
the design principles which have been devel-
important features of typical joints are illus-
oped to take into account the special
trated in the single fastener case shown in Fig.
characteristics of bolted composite structures
28.1. Key dimensions are 0, the fastener diam-
have been described4,6, 13.
eter, t, the thickness of the joint structural
It is not possible within the scope of this
elements, e, the edge distance (distance from
discussion to describe all the details and
the fastener center to the unloaded edge of the
processes that are necessary for achieving the
upper plate) and W, the width of the part of
design of specific joints. The objective is rather
the upper plate to the left of the fastener.
to give the reader some insight into the key
Similar features apply to the lower plate ele-
factors that control the behavior of mechani-
ment. Note that the same load, P, is passed
cally fastened joints in composite structures.
successively through various sections of the
The behavior of mechanically fastened
joint, including the bearing section in front of
joints is governed by two types of information:
the fastener of area Ot which is in compres-
(1) the features that control the behavior of the
sion, the two shearout sections of total area 2et
joint around individual fasteners and; (2) the
which are loaded in primarily in shear, the net
behavior of multiple arrays of fasteners. The
section, (W - O)t which is in tension, and the
behavior of individual fasteners can be consid-
gross section Wt which is also in tension.
ered in terms of a generic rectangular element
The average stresses associated with these
surrounding each fastener (Fig. 28.1), whose
sections are:
length and width are represented as ratios
with respect to the fastener diameter. The average bearing stress, ab = PlOt;
effects of the geometry of this element together average shearout stress, aso = P12et;
with effects of the reinforcement arrangement average net section stress, aN = P / (W - O)t;
used in the laminate for the element determine average gross section stress, aG = P /Wt.
the structural conditions under which the ele-
(28.1)
ment will fail. Once the characteristics of the
rectangular element are selected, its deforma- LOAD PATH
tion characteristics can be combined with
those of other rectangular elements making up
the joint to obtain the performance of the joint
as a whole. This discussion is organized in -
p

terms of these two aspects of joint design.

In addition to a preliminary discussion of
general features of mechanically fastened belll"ilYJ sect ion
!TOSS net Sectl~ 21>t
joints, the discussion which follows also con- section I section Ot I
siders: (1) single fastener joints, including WI ! (W-O)t !
I I
effects of joint geometry together with those of I I

,, E'~i=====::J'_
=====t;\Em'
composite material behavior; (2) multi-fas-
-·C=I
tener joints; (3) fastener effects, and (4) a '
,
discussion of test methods which provide
empirically-based data needed for completing
the joint design. Fig. 28.1 Single fastener joint.
 Mechanically fastened joints 613

!E Lateral Pltctl

HP+P+P)
1 2 r
---

Fig. 28.2 Multi-fastener joint.

From the standpoint of the designer, the gross

section strength is of primary interest since the ~S)I1iIX
objective of good design is to stress the gross
section to its highest level. Structural perfor- r--------~f__-,/
t xy)I1iIX
mance of the joint can be rated in terms of joint
efficiency, which refers to the ratio of average
gross section stress at failure of the joint to the
strength of the laminate in the gross section,
essentially the strength achieved in coupon
tests for tensile or compressive loading of
unnotched specimens. For organic matrix
composites, single fastener joints achieve joint Fig. 28.3 Peak stresses around fastener.
efficiencies of less than 50%, while for multi-
fastener joints the maximum achievable
efficiency is of the order of 60%. In contrast, gross section:
metallic joints can reach efficiencies of close to
(28.2)
80% because of the opportunity for taking
advantage of local yielding around points of or
high stress, though even in metallic joints, Knt=u)
design for avoidance of crack initiation in G nsmax /aG
cases where long life under cyclic loading is where ans)max is the maximum axial stress on
required may force the joint efficiency to be the net section. Predictions of the peak stresses
lower than for single cycle loading. in the joint can be made using continuum elas-
ticity analysisl-S, 7, 12, photoelastic measure-
ments10 or finite element methods8•
28.3.2 EFFECTS OF JOINT GEOMEJRY
The behavior of K~t as a function of D/W
The peak value of axial stress on the net sec- for isotropic joint elements is given in Fig. 28.4
tion uns)max (Fig. 28.3) is one of the key which was obtained from photoelastic mea-
controlling parameters on joint performance. surements of K:Jt against D/W (note that ~I =
It is convenient to express this as a stress con- K:JI/(l - D/W» given by Petersen10• Similar
centration factor, i.e. a ratio with respect to curves for orthotropic plates were obtained by
either the mean stress in the net section or the elasticity analysis2, 3, 12.
614 Mechanical fastening and adhesive bonding

There is obviously a competition for space The behavior of this approximation empha-
between the bearing section and the net sec- sizes the tradeoff between the effect of the
tion. Increasing the fastener diameter, D, to bearing section and the net section on the peak
lower ub has to decrease (W - D)t, the net sec- net section stress. A similar functional fit to the
tion, thereby increasing uN' and vice versa. curve of K{;t (the stress concentration based on
Furthermore, if D/W is much less than I, the UN) against D / W was given for isotropic plates
case is similar to that of a fastener in an infi- by Hart-Smith9, based on photoelasticity
nitely wide plate; in such a case all the peak results reported in the literature; when stated
stresses become a constant multiplier of the in terms of K[lt, the Hart-Smith approximation
bearing stress ub' which becomes large for leads to
D/W«1. If D is large enough to fill most of
Ktlt (bolt loaded hole) ... W/D + 2/(1- D/W)
the available width, the net section becomes
small, UN becomes large, and the peak net sec- -1.5/ (1 + D/W) (28.4)
tion and shear stresses become proportional to
He also gave an expression for K;;t for open,
UN for this case. As seen in Fig. 28.4, Ktlt is large
unloaded holes (Le. with no load transfer
for small D/W as well as for D/W approach-
between the fastener and plate) in isotropic
ing I, leading to a minimum around D/W =
plates; when the stress concentration is
0.59,12. The quantity 0.89u/uG + 1.8 uN/uG
expressed as a ratio with respect to uG rather
which is plotted in Fig. 28.4 gives a reasonable
than UN the Hart-Smith open hole expression
approximation to the curve of Ktlt against
reduces to:
D /W, so that the use of eqn (28.1) to express ub
and UN in terms of D /W, shows: Ktlt (open hole) =2/(1- D/W)
K[lt ... 0.89W/D + 1.8/(1- D/W) (28.3) + (1- D/W)2 (28.5)

FunctiOnal fit .
<It 1.8 aN
0.89

1
O+---------+---------1----------r---------r-------
0.15 0.28 0.4 0.53 0.85
OIW --;;.-
Fig. 28.4 Predicted net section stress concentration against D/W.
 Mechanically fastened joints 615

This is of interest in later discussions of com- delamination and fiber-separation around

bined bearing-bypass loads on bolted joints. regions of peak stress. The method is applied
As a means of predicting joint strength, we can by averaging the stresses predicted by two-
equate the peak stress ans)max to pu, the dimensional elasticity analysis over a selected
strength of the orthotropic laminate used in length aD' along a path lying normal to the load
the joint. If the joint material acts purely elasti- direction, in the vicinity of the peak stress. The
cally, then the elasticity prediction of K[lt gives failure load is obtained by equating the aver-
F~ which denotes the value of aG when net sec- aged stress to the laminate strength PU, and ao
tion failure occurs, as F~ = pu / K[lt. The is determined by an empirical fit of the mea-
minimum in the KGnt against D / W curve of sured strength data. In an alternative version of
Fig. 28.4 will result in a maximum in the rela- the Whitney-Nuismer analysis, Pu is equated
tionship between F~ and D / W at D / W = 0.5. to the stress at a point spaced an empirically fit-
However, joint strength is usually found to ted distance do from the location of the stress
be considerably greater than that based on peak. The latter approach has frequently been
purely elastic behavior of the composite to fail- used for joint strength prediction15, although
ure, and it is appropriate to use the method the averaging and point stress versions of the
developed by Whitney and Nuismer14 to cor- Whitney-Nuismer analysis are equivalent.
rect strength predictions based on linear elastic Figure 28.5 illustrates the prediction of F~
response to failure. The Whitney-Nuismer for net section tension failure as a function of
approach was developed for strength predic- D / W. The curves here were obtained from
tion of composite plates in the presence of stresses given by elasticity analysis of 0/90°
circular holes, crack-like slits and other types of glass epoxy joint elements12, by averaging the
notches, and is meant to take into account the predicted axial stress distribution ax in the net
softening effect of local damage such as section (using the definition of x and y indicated

D = 0.63 cm (0.25 ") Laminate Tensile Strength 110 ksi (0.76 GPa)
45 -=--0.31 GPa Compressive Strength 100 ksi ( 0.69 GPa)

40

.
]I
tn
35

~ 30
;: -rr-----I----+1.29
C/) 25
c:
o
~ 20
Q)
C/)
tn
15 °
tn 10 net section tension
2 failure
~ 5
o ~ ____________ ____________ ____________+-__________
~ ~ ~

0.13 025 0.33 0.5 0.63

D/W
Fig. 28.5 Maximum gross section stress against D/W.
616 Mechanical fastening and adhesive bonding

by the coordinate axes centred at the hole in

Fig. 28.3 from the edge of the hole to a point a
distance ao from the hole, using the ao values
- •.m
..... "
listed in Fig. 28.5. The lowest curve of Fig. 28.5
which corresponds to ao = 0 represents purely
elastic behavior of the joint material. It is dear
that increasing the value of ao causes an

- ..--
upward shift of the predicted strength curve.
Experimental joint strength data12 gave an ao
value of about 3.8 ± 1.3 mm (0.15 ± 0.05 in) for ....
10
~-.--
' •••WD .....'
quasi-isotropic carbon epoxy laminates, for
•
100
net section failure. Nuismer and Whitney14
suggested that ao and the corresponding do are
relatively constant for composite materials in
general, although this is not always true.
However, for a given material and stacking
sequence, ao is independent of hole size; this
has the effect that for large fastener diameters,
the averaging effect is less important, and Fig. 28.6 Experimental results on strength against
D/W2.
strength values tend toward those predicted
by the elasticity analysis.
All the net section failure curves of Fig. 28.5 'bearing failure', corresponding to compressive
have the peak in the vicinity of D/W = 0.5 failure in front of the pin. Except for joints with
which was predicted by the linear elastic short edge distance or multi-fastener joints
response analysis. This suggests that the joint where high bypass loads are present, the peak
strength is a maximum for bolt diameters near compressive stress in front of the fastener 12 is
half of the plate width. For multi-fastener found to be about 1.3 times ab , and ac for bear-
joints such as the one shown in Fig. 28.2, the ing failure can be estimated by setting ab equal
length corresponding to W for the single fas- to pro (compressive strength of the laminate)
tener joint is replaced by Weft' the spacing of from which ac)max =Feu D /1.3W, a linear func-
the dashed lines, which are lines of lateral tion of D /W. Although this result implies that
symmetry; note that Weft is also equal to the the laminate compressive strength is the same
lateral spacing of the fasteners (the lateral as the bearing strength as is approximately true
pitch), so that the peak strength value occurs in some cases, the actual value of pm = aJmax
at a lateral pitch of about twice the fastener depends on the stacking sequence of the lami-
diameter. The experimental data of Fig. 28.612 nate as well as a number of details of the joint
confirms the occurrence of a strength maxi- such as the type of fastener and whether the
mum for D/W = 0.5 in single fastener joints. joint is in single or double shear (Section
For multi-row joints (see Section 28.3.4) in 28.3.5); thus pm has to be empirically deter-
which bypass load is present, the net section mined for a given fastener, joint configuration
strength peak occurs for smaller values of and joint material selection by appropriate
D/W. High bypass loads give results similar bearing strength tests (Section 28.3.6). Hart-
to those for plates with unloaded holes, in Smith9,13 introduced a method of comparing
which case large values of W/D correspond- bearing and net section tension failure similar
ing to large fastener pitch are desirable. to that given in Fig. 28.5 to show the trade off
In addition to the solid curves in Fig. 28.5 for in failure modes for various values of D/W,
net section failure is the dashed line designated and recommended selecting D / W for the
 Mechanically fastened joints 617

design of the joint as the point at the intersec- strength will not be achieved unless e is at least
tion of the bearing and tension failure as great at W. This is shown schematically in
envelopes. However, it was also pointed out in Fig. 28.8, which indicates that the minimum e
Nelson, Bunin and Hart-Smith13 that in multi- for full joint strength over a wide range of fas-
row joints in which the bearing load is tener diameters remains equal to W. It is true
distributed over several fasteners in a given that there is a tendency for bending failures to
column, the bearing stress is smaller than for occur at the unloaded end of the joint (Fig.
the single fastener case and bearing failure may 28.9) for edge distances as small as 1-2 times
not occur in practical joint designs. D, which were shown in Oplinger12, in fact, to
The value of the edge distance, e, is another be a function of e/ D; however, this type of fail-
critical parameter for bolted joints. While it ure is eliminated for e/D greater than about 2.
has become customary to present joint
strength as a function of e/ D, in fact9, 12 it is
more a function of e/ Wife / D is much greater
than I, and plotting data on the effect of edge
distance as a function of e/ D is somewhat mis-
leading even though it persists as accepted ----------l--E!)~;
practice. Both stress analysis9, 12 and experi-
----------- Wi --- :-~~-

-,
mental results (Fig. 28.7) indicate that full joint

~-~=-~=--- ~--~--­
------+--w---
------ ---- IN
rh~1S
----------J-----__ _
Fig. 28.8 Minimum edge distance is equal to W
.t /..~l--
~ 0
- - - -I
II!-A,WD·.
(independent of D) .
.... " "8 . .

X ~41.OIGIIIE"'Y.W/D.I.41'
o .... II "5.3"

x ~./IO.AP"ITI UOICYt./D.I.I~ high bending

,'T
o ••
stress cry
• • Ito
t
etw

Fig. 28.7 Effect of edge distance on joint strength Fig. 28.9 Tendency for bending failure with small
WID =212. eiD.
618 Mechanical fastening and adhesive bonding

28.3.3 EFFECTS OF JOINT MATERIAL

Typical failure modes for mechanically fas-
tened joints are shown in Fig. 28.10. Figure
......... -~.... - -Ell I
28.11 indicates that shear-related failures may
include what are usually 'shearout' failures
such as the upper figure for Oz±45 reinforce-
ments, which tend to lie along straight lines
tangent to the bolt hole and to run to the end
of the joint, as opposed to those for 0/90° lam-
Fig. 28.11 Shear failures in bolted joints.
inates which originate at about 45° points
around the bolt hole and run along curving
lines to the joint end. Hart-Smith9 shows In addition to modelling shear failures, Chang
examples of 'shearout' type failures while the and Hung16 also reported results related to
second type are demonstrated by experimen- bearing failure which show that these failures
tal results for 0/90° glass epoxies1Z• Since the result from high out-of-plane shear stresses;
90° points around the bolt hole are traction the associated failure surfaces are oriented
free, shear stresses cannot exist there, and along planes at 45° to the x-y plane. In addi-
'shearout' failures are probably associated tion to clarifying the bearing failure modes,
with fiber separation failures resulting from to these results verified the importance of clamp-
a deficiency of 90° fibers in Oz±45 laminate, ing pressure as a factor in achieving maximum
while failures originating at the 45° points in bearing strengths. This is discussed further in
0/90° laminates appear to be true shear fail- Section 28.3.5.
ures. Chang and Hung16 provided an It should be mentioned that the term 'bear-
analytical approach which modelled progres- ing' strength, which is usually associated with
sive damage in 0/90° laminates that predicted the maximum achieved value of 0v does not
the types of shear failures that are depicted in necessarily imply that the joint fails by bearing
Fig. 28.11. failure, since, depending on the joint geometry

CUAVAGf -TENSION FAILUIE lOll 'AlIUII

Fig. 28.10 Typical failure modes in composite joints13•

 Mechanically fastened joints 619

and reinforcement, failure may take the form 100 / PEACENYAGE to-DEGREE PLIES
of shearout or net tension rather than bearing
failure; bearing strength only identifies the to ., (J

RAtiNG STRESS CONYOUIS (1(51'

value of O"b that occurs when the joint fails. The
value of O"b for the bearing failure mode can be
looked on as a laminate material parameter,
however, and can be obtained from appropri- I! ...
ate single fastener strength tests, to be I 50 IOLT DIAMETER. 0.25 IN.
0
discussed in Section 28.3.6, in which e/W (usu-

I
1T00QLaD lOLlS)
'"
40
ally specified in terms of e/V) and W/V are
large enough to guarantee a bearing failure 311
r
mode. The effect of the reinforcement configu- 20
ration on the maximum achieved O"b is an
important factor in joint design. Note that for
design purposes, the normal reinforcement o~~-a __ ~~~ __ ¥-~~ __ ~~

arrangement includes only plies in the 0°, 45° o ~ 311 311 40 50 40 ~ • to ~

IIICENfAGE :t45-DEGREI PUES
and 90° directions with respect to the load
axis, implying that other orientations are not
Fig. 28.12 'Carpet' plot of maximum bearing stress
encountered; in addition, the 45° reinforce-
vs. percentage of reinforcement in key directions4•
ment is arranged symmetrically as double
plies of ±45° reinforcement. The notation com-
monly used to describe the stacking sequence the reinforcement arrangement in the vicinity
is condensed to reflect percentages of plies in of the joint.
the three customary directions, 25/50/25 rep-
resenting 25% 0°, 50% 45°, and 25% 90°, or
28.3.4 MULTI-FASTENER JOINTS
quasi-isotropic, reinforcement, for example.
The effects of reinforcement percentages on Single fastener joints cannot generally achieve
maximum achieved bearing stress are shown anything close to the strength of the laminate
in Fig. 28.12, a three-parameter 'carpet plot' being loaded and are not usually encountered
which is used to show the effect of percentage in structural joints. Single fastener coupons
of reinforcement in the 0°, 90° and ±450 direc- do, however, serve as building blocks for
tions on maximum achieved bearing stress for design of multi-fastener joints (Fig. 28.2) since
a fixed joint geometry. The contours of con- stress analysis and strength test results on sin-
stant bearing strength indicate that maximum gle fastener geometries provide data which
joint strength is achieved within a broad translate directly to multi-fastener arrays in
plateau lying between 30-60% 0° and 10-60% structures. There are two principal features
±45° reinforcement. To avoid low strength that affect multi-fastener arrays; that of bear-
against shearout failure, it is generally agreed ing-bypass load is illustrated in Fig. 28.13. The
that a minimum number of 90° plies, at least bypass load, P can be considered as a load
10%, should be included in the laminate. added at the ~aded (right) end of the upper
Laminates reinforced in fewer than three element in the single fastener joint in Fig. 28.1.
directions, i.e. 0/90° and ±45° reinforcement, In the case of a multi-fastener array (Figs. 28.2,
should be avoided if possible, since they 28.14), Pbp for a given fastener is equivalent to
respond ductiley with excessively low yield the sum of the loads developed by the fasten-
strengths. These recommendations may be dif- ers lying to the right of (i.e. those lying in the
ficult to follow if considerations other than direction of the load from) the one under con-
maximum joint structural performance govern sideration. For each fastener, the load PN in
620 Mechanical fastening and adhesive bonding

On the other hand, the net section load on the

last fastener is the sum of all fastener loads and
is therefore the same as in a single fastener joint
carrying the same total load. The benefit of
Pf tPbp multi-rowed fastener configurations is primar-
Pq,
.....:3

(=p )
N
~f I
I
Do

("bypass"
ily because the stress concentration factor
associated with pure bypass load (eqn (28.4» is
smaller than that associated with pure bearing
load) load (eqns (28.7), (28.3)), and the load ratio for
the most highly loaded net section (the left most
Fig. 28.13 Definition of bypass load. in Fig. 28.14) tends toward the case of pure
bypass loading for large numbers of fasteners.
Fig. 28.13 (see also Fig. 28.14) is the load Stress analyses for combined bearing-bypass
passed through the net section associated with loading have been discussed 7, 8, 12, 17, 18.
that fastener and is the sum of P f' the fastener Figure 28.1512 gives the elasticity predic-
load, and Pb ; PN is also equivalent to the total tions of K(Jt for various ratios of Pb to PN'
load at the reft end of the joint in Fig. 28.13. indicating the K(Jt is a nonlinear fttriction of
The situation in Fig. 28.14(b) shows the P/Pb • (The results in Fig. 28.15 are for a per-
buildup of PN along a joint containing five fas- fectfit fastener and for a joint under tension.)
teners for the hypothetical case of equal Pf for This nonlinearity is caused by the variation of
each fastener, while Fig. 28.14(c) shows the angle of contact between the fastener and the
distribution of Pb / P N for that case. It is impor- plate with the load ratio; Fig. 28.16 illustrates
tant to realize that the net section stresses the difference in the contact region for the
build up along with PN' If the joint design is extreme cases of pure bearing and pure bypass
such as to distribute the fastener loads Pf load. In the latter case the plate stretches in
evenly as in Fig. 28.14, the load on each fas- opposite directions along a line parallel to the
tener is then the total load divided by the load direction and contact splits into two
number of fasteners, a fraction of 1/5 in Fig. regions centered about +90 0 and -90 0 with
28.14, so that the bearing stress at each fastener respect to the load direction, while for pure
is much smaller than for a single fastener joint. bearing load, a single contact region occurs
ranging from about -100 0 to +1000 about the
load axis for perfect-fit fasteners. (For clear-
ance fits in the case of pure bearing load the

•
--
Q. -
P/W;~-C· ... ~:p

Fig. 28.15 Net section SCF against bearing/bypass

Fig. 28.14 Bypass against bearing load. load ratio.
 Mechanically fastened joints 621

contect (28.5». The results, given in Oplinger12, indi-

CAl AlE cated that this will result in inaccurate

--
IIrFASS predictions, not only for the general bearing-
bypass situation but also for the case of pure
bearing load with small edge distances (Fig.
28.9). Currently available analytical and finite
element tools are sophisticated enough to treat
(8) PIlE the contact problem routinely, and the 'cos 0'
IfMltIi radial pressure distribution should be avoided,

--
although the superposition method gives some
useful insight into the situation if analytical
tools for dealing with the contact problem are
not available. Crews and Naik8 obtained
results which showed that the hoop stresses
around the bolt hole are predicted with reason-
Fig. 28.16 Contact angles for pure bearing load and
pure bypass load.
able accuracy by the superposition method, so
that with judicious use of Nuismer-Whitney14
correction factors, net section failure stresses
can be reasonably well predicted by superposi-
contact region varies with load and is smaller tionibearingfailurescannot.
than that of exact-fit fasteners. Crews and Naik and Crews18 described test methods
Naik8 treated the case of 1.2% clearance, i.e. a for joint strength under combined bearing-
hole diameter 1.2% greater than the fastener bypass loading, with compressive as well as
diameter, for which contact between the fas- tensile bypass loads. Typical results are given
tener and hole occurs from about -600 to +60 0 in Fig. 28.17 for a 1.2% clearance fit fastener in
for typical loads.) a quasi-isotropic carbon epoxy laminate.
For intermediate bearing/bypass ratios, the Failure modes here are designated 'TRB' ('ten-
contact region is a mix of the two situations. sion reacted bearing', i.e. bearing failure with
Oplinger12 gave the variation of the radial pres- tension bypass loading), 'TRC' (bearing 'fail-
sure distributions for exact fit fasteners, for ure with compression bypass loading), 'NT'
joints in tension, which showed how the con- (net section tension), 'NC' (net section com-
tact region is modified as the bearing/bypass pression) and 'ose' or 'offset compression'
ratio varies, while Naik and Crews8 treated which refers to the failure mode illustrated in
cases of both exact and clearance fits for both Fig. 28.18 for compression bypass load.
tension and compression loaded joints; in addi- Although the load distribution for the five-
tion, MadencF gave comparable results for fastener joint shown in Fig. 28.12 is
cases of shear loaded joints. A number of represented as having the same Pf for each fas-
efforts have treated the contact problem for tener, this condition cannot be achieved for
pure bearing by assuming a radial pressure joints in general. As shown in Fig. 28.19 l3,
distribution which varies around the hole as there is usually considerable variation of fas-
the cosine of the angle with respect to the load tener loads along the joint. In a two-fastener
axis; in addition, for combined bearing and joint such as that shown in Fig. 28.20, the
bypass loading (following industry practice) upper and lower plate elements ('U' and 'L')
Hart-Srnith9 and others have, as a matter of have to stretch equally under load if there are
expediency, superposed the peak stresses pre- no fastener deflection effects (no fastener tilt-
dicted for pure bearing and pure bypass (i.e. ing or bending).
the values of K~t predicted by eqns (28.4) and
622 Mechanical fastening and adhesive bonding

Beartng etre.s, Sb' MPa

800

thr •• hold

Bype.. m.... S..." Wa

Fig. 28.17 Laminate strengths under combined bearing/bypass loading17,

Fastener-
"(S;' Fa j lin Itlle Ibsher
and
Circlllterence fGL =aCL / ELf
the elongations are:
0u = lfcu and 0L = le CL'
The condition that 0u = 0L requires that:
l(P - PfJ)/EutuW = IPfl/EAW.
Since Pf2 = P - PfI, this is equivalent to:

Fig. 28.18 'OSC' failures for combined bearing Pf2 = (Eutu/ ELtL)Pfl,
bypass loading17,
If EL»Eu or tL»tu' the resulting Pf2 will be
small compared with PfJ so that the second fas-
With the average gross section stresses and tener will be nearly unloaded. In such a case
strains for the two plate elements in Fig. 28.20 the joint will be equivalent to a single fastener
give by: joint containing an unloaded hole at the loca-
tion of the second fastener. On the other hand,
the last equation shows_that if the stiffnesses of
the upper and lower plates are equal, the two
fasteners will be loaded equally. In general,
fastener loads will be highly variable in a way
that depends on the relative thicknesses and
 Mechanically fastened joints 623

moduli of adjacent plate elements in the

regions around each fastener; fastener deflec-
tions due to beam-bending of the fasteners as
well as clearance effects will add other compli-
cations to the situation. In Fig. 28.19,
configuration A illustrates the behavior of
joints with both elements tapered (i.e. scarf
joints); due to load transfer between the ele- p
ments by fastener load, the net section load PN
decreases from the loaded (thick) end of each
plate to its unloaded (thin) end, while the cor-
responding thicknesses decrease keeps the Fig. 28.20 Two fastener joint.
gross section stresses and strains and therefore
the stretching deflections uniform along each listed under each configuration), it is noted
element, providing for nearly equal load trans- that the scarfed configuration gives the lowest
fer at each fastener. For configuration B of Fig. strength of the four, and is about 9% weaker
28.19, the case of uniform element thickness, than the uniform thickness configuration, B. It
the interior fastener loads are smaller than is usually expected that scarfing will lead to a
those at the joint ends, which is typical of this way of keeping the bearing load minimized at
situation. Since the configurations shown at the joint ends where the highest P ~ are
the right in Fig. 28.19 are arranged in order of encountered, but other factors having to do

!i
increasing load capability (see the strains with the balance between the effects of local

BOLTLOAODISTRIIUTIONS BOLT DIAMETERS

14 ,..--...;.;..;.;;.;.;...;;,,;;;~~.;;.;;.;;.;..-.-... !II. 1/2.

4-ROW BOLTED JOINT

r
,n:611 \ =
1
I :I :I
O.liOIN.
2 3
i;
4 O.37&_IN.

•I ;_

,.
" CONFIGURATION 0
, ,0 e.... • O.-INJ!N.

Ii :=
"---..' .' O.IID Itt. '1 21 3 41 O.zaIN.
I " 1/2. 1/2. 1/2:m-' t , i I1
IOLTLOAD
(1,oooLB) COf4FIGURATION C
e.... • O.DIMIIN/ltt.

r: Iii i 1;:=
0.&0 Ir. ' 2 3 4 0.21 IN.
6 :6 CONFIGURATION A
D-·-o CONFIGURATION 8 1/2.,/2.1/2:;'
0-.-0 CONFIGURATIONC CONFIGURATION 8
2· • _ _ • CONFIGURATION 0
e... ·O.IIIMIINJIN.
____ ________ 0.150 IN.

I I I'
o~~ ~ ~~
1 2 3 4 0.25 IN.

~ I:::
1 2 3 4
IOLTNUMIER 1/2. 1/2, 1/2,;/2 \ ;
CONFIGURATION A
e... • O_,INJlN.
Fig. 28.19 Effect of joint configuration on fastener load distribution13.
624 Mechanical fastening and adhesive bonding

element thickness and local bearing load pre- 28.3.5 FASTENER EFFECTS
dominates here. Note that configuration D, the
Joint strengths for local areas (Fig. 28.13)
strongest, uses a combination of variable fas-
around individual fasteners are affected by a
tener diameter and element tapering to
number of parameters associated with the fas-
achieve a maximum thickness of the outer ele-
teners. Some of these include: whether the
ments which is greater than that in
joint is in single or double shear (Fig. 28.21);
configuration C, to obtain maximum joint
the use of countersunk (flush head) compared
strength. Thus, judicious use of element taper-
with protruding head fasteners (Fig. 28.22);
ing and variation of fastener diameters as well
effects of fastener diameter; effects of fastener
as other joint parameters can improve joint
clamping; fastener clearance effects, and effect
performance. With untapered joints, the maxi-
of fastener deflections on fastener load distrib-
mum benefit of additional rows of fasteners is
ution. Bearing strengths are significantly
not much more than 20% greater than that for
affected by the use of single compared with
single row joints. Joint tapering will provide

::. J]
some improvement over that figure, although
the benefit is limited; it should be kept in mind
that the net section load at the loaded end of a
given element will be the same for single fas-
'1- 2,

tener and multi-fastener joints, so that the

benefit of adding more than a few rows may (A) Double Shear Configuration
not be great. The interaction of the effects
encountered in multi-fastener joints is fairly
complicated and requires the use of analyses
which can take into account the stresses and
strains in single fastener configurations with
bypass loading present (Fig. 28.13), represent-
ing 'unit cells' of the joint configura~ion, (8) Single Shear
together with finite element calculatIOns
which evaluate the interactions among the Fig. 28.21 Single shear and double shear configura-
various unit cells to provide the overall fas- tions.
tener load distribution. Computer codes have
been developed under DoD and NASA spon-
sorship to provide for this type of integrated
joint design. For example, Nelson, Bunin and
Hart-Smith13 discuss the application of the
well-known 'A4EJ' code19 in conjunction with
the code 'BJSFM' (Bolted Joint Stress Field
ModeP5) which were developed by McDonnell (A) Protruding Head Fastener
Douglas under NASA and Air Force sponsor-
ship for this type of joint design analysis.
Northrop similarly developed codes 'SASCJ'
(Stress Analysis of Single Fastener Composite
Joints) and 'SAMCJ' (Stress Analysis of
Multifastener Composite Joints) under Air (B) Countersunk Fastener
Force Contracf',20. For information on the the-
ory and application of these codes, the reader Fig. 28.22 Countersunk and protruding head fas-
is directed to the references. teners.
 Mechanically fastened joints 625

double shear configurations, bearing strengths The fastener diameter should be on the
for single shear joints tending to run consider- order of the thickness of the thicker of the
ably below those for double shear because of plate elements making up the joint, or greater,
greater through-the-thickness variation of fas- (D/t ~ 1) to avoid excessive fastener bending.
tener-plate contact pressure. Bearing strength As in Fig. 28.10 (lower right-hand sketch)
tests referenced in Section 28.3.6 include sepa- excessive bending can lead to failure of the fas-
rate test configurations for the two situations. teners, which is intolerable. In addition,
In addition, as indicated in Fig. 28.22(a), bend- fastener bending causes uneven distribution
ing moments tend to occur in single shear of bearing pressure through the plate element
joints which are not present with double shear thicknesses, so that the full bearing strength is
arrangements. Fastener head pull through not available in such cases. The effect of large
(Fig. 28.10) can be a problem in the presence of fastener deflections on the clamping pressure
such bending effects, and special test methods provided by the fastener is another adverse
for fastener pull-through strength are effect of fastener bending deflections. Figure
described in Section 28.3.6. Countersunk, or 28.2413 illustrates the fact that bending defor-
flush head, fasteners (Fig. 28.22(b» are fre- mations reduce the clamping pressure
quently encountered in exterior surfaces of provided by fastener head, causing a reduc-
aircraft components where avoidance of air tion of bearing strength which is in addition to
flow disturbance is required. Countersunk fas- that caused by uneven bearing pressure
teners for composites include (Fig. 28.23) through the thickness.
'tension head' fasteners having the larger head The beneficial effect of clamping pressure
depths and therefore wider heads, and 'shear on bearing strength, discussed earlier, has
head' fasteners having smaller head depths, been well established. Required clamping lev-
with head angles ranging from 100° to 130°. els are usually described in terms of bolt
Countersunk fasteners tend to bear against the torques, 'finger tightened' being the lightest
surrounding element more unevenly through level, and installation requirements specify
the thickness than protruding head fasteners torque levels which supposedly represent par-
do. Tension head fasteners are generally pre- ticular bolt tensions (and therefore clamping
ferred over shear head fasteners because of
greater strength against head pull-through;
however, if the joint element is so thin that the
countersunk depth is greater than 70% of the
element thickness, the tendency toward
uneven bearing pressure in tension head fas-
teners is too great and shear head fasteners are
recommended in this case.
DELAMINATIONS DUE TO BEARING LOAD AND REDUCED
ClAMP-IJP AS A RESULT OF BOLT BENDING
Tension Shear stear
HIGH BEARING LOAD SIDE
tad Itmd ted
(100 0 ) (100°) (130 0 )

U U lJ Fig. 28.24 Effects of fastener bending on joint fail-

ure13.
Fig. 28.23 Tension head and shear head fasteners.
626 Mechanical fastening and adhesive bonding

pressures) that can be calculated in terms of Complications, such as the way in which the
the pitch of the fastener threads from machine fastener head and nut/washer combination
design formulas. Bolt tensions for a given bears on the surfaces of the plate element will
torque level are notoriously variable because influence the outcome of such calculations and
of friction effects in the bolt threads, but spec- must be taken into account. In addition, the
ified torque levels which have been through-thickness distribution of bearing
determined empirically probably represent pressure between the fastener and the sur-
minimum clamping levels necessary to insure rounding the plate element should be
maximum bearing strengths that can be included in the calculations. The method of
achieved when variations in service condi- Harris, Ojalvo and Hooson23 which treats the
tions are taken into account. Steps should be bore of the fastener hole as an elastic founda-
taken to avoid loss of clamping pressure due tion for the beam used to model the fastener
to through-the-thickness viscoelastic deforma- has been applied 6 for such calculations.
tion of the laminate21,22 at elevated temperature
and humidity.
28.3.6 TEST METHODS
Fastener clearances are typically on the
order of 0.075 mm (0.003 in) or less for typical Joint strength tests are needed to establish cer-
0.635 mm (0.25 in) aircraft fasteners. Analytical tain key parameters of the joint as inputs to
studies have shown that bearing stresses design analyses. Such data as failure stresses
increase significantly for such relatively small for pure bearing load as obtained from single
clearances since the angle of contract decreases fastener coupons, and open hole coupon
rapidly as the clearance increases. Clearances strengths for both tension and compression
also have a significant effect on the fastener loading, are needed to establish joint perfor-
load distribution since the last in a series of mance for pure bearing and bypass loads.
fasteners cannot take up load until all the Intermediate combinations of bearing and
clearances have been taken up. bypass load must also be considered to pro-
In addition to the effect of clearances on fas- vide empirical curves for dealing with the
tener load distribution, effects of fastener general situation. Because of differences in the
bending deflections must be taken into way the fastener contacts the surrounding
account in load distribution calculations such plate materials, bearing tests have to be con-
as those provided by the A4EJ, SASCJ and ducted to treat both single and double lap
SAMCJ codes described above. In the case of (single and double shear) configurations. In
the two-fastener joint shown in Fig. 28.20, single lap joints in particular, tests are needed
bending and rotational deflections of the fas- to establish the effect of fastener rotation and
teners will modify the load distribution bending deflection. Fastener deflections must
described in the discussion of that figure for be determined in tests of the type just
zero fastener deflection. For ELtL»Eutu' in Fig. described for providing fastener response data
28.20 for example, fastener deflections will in connection with predictions of load distrib-
allow some load to be transferred to the sec- ution in multifastener joints. In addition,
ond fastener, as opposed to the case of no fastener head pull-through strength tests have
fastener deflections discussed earlier which to be performed to allow for joint configura-
led to Pf2 = O. Fastener deflection effects can be tions in which overall bending takes place, in
inferred from bolt bearing tests which provide which case out-of-plane forces between the
for deflection measurements. Alternatively, fastener and joint plates tend to be significant.
analytical approaches based on beam models The details of test methods for mechanically
for the fastener which include both bending fastened joints are described by Shyprykevich24
and shear deformations have been used 13 • and in MIL-HDBK-l725.
 Adhesive joints 627

28.4 ADHESIVE JOINTS the adhesive is never the weak link; (3) recog-
nition that the ductility of aerospace adhesives
28.4.1 INTRODUCTION is beneficial in reducing stress peaks in the
adhesive; (4) careful use of such factors as
Adhesive joints are capable of high structural adherend tapering to reduce or eliminate peel
efficiency and constitute a resource for struc- stresses from the joint; (5) recognition of slow
tural weight saving because of the potential cyclic loading, corresponding to such phenom-
for elimination of stress concentrations which ena as cabin pressurization in aircraft, as a
cannot be achieved with mechanically fas- major factor controlling durability of adhesive
tened joints. Unfortunately, because of a lack joints, and the need to avoid the worst effects
of reliable inspection methods and a require- of this type of loading by providing sufficient
ment for close dimensional tolerances in overlap length to ensure that some of the adhe-
fabrication, aircraft designers have generally sive is so light1y loaded that creep cannot occur
avoided bonded construction in primary there, under the most severe extremes of
structure. Some notable exceptions include: humidity and temperature for which the com-
bonded step lap joints used in attachments for ponent is to be used.
the F-14 and F-15 horizontal stabilizers as well Much of the discussion to follow will retain
as the F-18 wing root fitting, and a majority of the analysis philosophy of Hart-Smith, since it
the airframe components of the Lear Fan and is considered to represent a major contribution
the Beech Starship. to practical bonded joint design in both com-
While a number of issues related to adhe- posite and metallic structures. On the other
sive joint design were considered in the earlier hand, some modifications are introduced here.
literature26-33, much of the methodology cur- For example, the reViSlOns of the
rent1y used in the design and analysis of Goland-Reissner single lap joint analysis36
adhesive joints in composite structures is have been re-revised according to the
based on the approaches evolved by L.J. Hart- approach presented in Refs. 53, 54.
Smith in a series of NASA/Langley-sponsored Certain issues which are specific to compos-
contracts of the early 1970s34-39 as well as from ite adherends but were not dealt with in the
the Air Force's Primary Adhesively Bonded Hart-Smith efforts will be addressed. The most
Structures Technology (PABST) program40-43 of important of these is the effect of transverse
the mid-1970s. The most recent such work shear deformations in organic composite
developed three computer codes for bonded adherends.
and bolted joints, designated 'A4EG', 'A4EI'
and 'A4EK'44-46 under Air Force Contract. The
28.4.2 SUMMARY OF JOINT DESIGN
results of these efforts have also appeared in a
CONSIDERATIONS
number of open literature publications47-50. In
addition, such approaches found application
28.4.2.1 Effects of adherend thickness:
in some of the efforts taking place under the
adherend failures versus bond failures
NASA Advanced Composite Energy Efficient
Aircraft (ACEE) program of the early to mid- Figure 28.25 shows a series of typical bonded
1980S5O,51. joint configurations. Adhesive joints in general
Some of the key principles on which these are characterized by high stress concentrations
efforts were based include: (1) the use of simple in the adhesive layer. These originate, in the
one-dimensional stress analyses of generic com- case of shear stresses, because of unequal axial
posite joints wherever possible; (2) the need to straining of the adherends, and in the case of
select the joint design so as to ensure failure in peel stresses, because of eccentricity in the
the adherend rather than the adhesive, so that load path. Considerable ductility is associated
628 Mechanical fastening and adhesive bonding

I
illustrate this point, shows a progression of
(A,
I i joint types which represent increasing strength
BONDED DOUBLER capability from the lowest to the highest in the
(B) , )
figure. In each type of joint, the adherend
UNSUPPORTED SlNGLE·LAP JOINT thickness may be increased as an approach to
II
(e) ,
achieving higher load capacity. When the
SlNGLE·STRAP JOINT adherends are relatively thin, results of stress
0::::::::
(D) • ;::::::=0
analyses show that for all of the joint types in
TAPERED SlNGLE·LAP JOINT

: :
Fig. 28.26, the stresses in the bond will be small
IE) j enough to guarantee that the adherends will
DOUBLE.LAP JOINT reach their load capacity before failure can
l ) occur in the bond. As the adherend thicknesses
IF) I
increase, the bond stresses become relatively
DOUBLE·STRAP JOINT
larger until a point is reached at which bond
(eI) * -== 0 ~ failure occurs at a lower load than that for
TAPERED STRAP JOINT which the adherends fail. This leads to the gen-
eral principle that for a given joint type, the
Fig. 28.25 Adhesive joint types39,55. adherend thicknesses should be restricted to an
appropriate range relative to the bond layer
with shear response of typical adhesives, thickness. Because of processing considera-
which is beneficial in minimizing the effect of tions and defect sensitivity of the bond
shear stress joint strength. The response of typ- material, bond layer thicknesses are generally
ical adhesives to peel stresses tends to be limited to a range of 0.125-0.39 mm
much more brittle than that to shear stresses, (0.005-0.015 in). As a result, each of the joint
and reduction of peel stresses is desirable for
achieving good joint performance.
From the standpoint of joint reliability, it is
vital to avoid the condition where the adhesive
layer is the weak link in the joint, i.e. that the
joint be designed to ensure that the adherends
fail before the bond layer whenever possible.
This is because failure in the adherends may be
controlled, while failure in the adhesive is resin
dominated, and thus subject to effects of voids
and other defects, thickness variations, envi-
ronmental effects, processing variations,
deficiencies in surface preparation and other
factors that are not always adequately con-
trolled. This is a significant challenge, since
adhesives are inherently much weaker than the
composite or metallic elements being joined.
However, the objective can be accomplished by
recognizing the limitations of the joint geome-
try being considered and placing appropriate
AOHEREND THICKNESS
restrictions on the thicknesses the adherends
for any given geometry. Figure 28.26, which
has frequently been used by Hart·Smith39,55 to Fig. 28.26 Joint geometry effects39 •
 Adhesive joints 629

types in Figs. 28.25 and 28.26 corresponds to a Scarf joints with lU1balanced stiffnesses between
specific range of adherend thicknesses and the adherends do not achieve the uniform shear
therefore of load capacity, and as the need for stress condition of those with balanced
greater load capacity arises, it is preferable to adherends, and are somewhat less structurally
change the joint configuration to one of higher efficient because of rapid buildup of load near
efficiency rather than to increasing the the thin end of the thicker adherend.
adherend thickness indefinitely. Step lap joints (Fig. 28.25(h» represent a
practical solution to the challenge of bonding
thick members. This type of joint provides
28.4.2.2 Joint geometry effects
manufacturing convenience by accommodat-
Single and double lap joints with uniformly ing the layered structure of composite
thick adherends (Fig. 28.25(b), (e) and (f) are laminates. In addition, high loads can be trans-
the least efficient joint type and are suitable ferred if sufficiently many short steps of
primarily for thin structures with low running sufficiently small 'rise' (i.e. thickness incre-
loads (load per unit width, i.e. stress times ele- ment) in each step are used, while maintaining
ment thickness). Of these, single lap joints are sufficient overall length of the joint.
the least capable because the eccentricity of
this type of geometry generates significant
28.4.2.3 Effects of adherend stiffness
bending of the adherends that magnifies the
unbalance
peel stresses. Peel stresses are also present in
the case of symmetric double lap and double All types of joint geometry are adversely
strap joints, and become a limiting factor on affected by unequal adherend stiffnesses,
joint performance when the adherends are rel- where stiffness is defined as axial or in-plane
atively thick. shear modulus times adherend thickness.
Tapering of the adherends (Figs. 28.25(d) Where possible, the stiffnesses should be kept
and (g» can be used to eliminate peel stresses approximately equal. For example, for step lap
in areas of the joint where the peel stresses are and scarf joints between quasi-isotropic car-
tensile, which is the case of primary concern. bon epoxy (YOlU1g'S modulus = 55 GPa = 8 x
For joints between adherends of identical stiff- 106 Ib/in2) and titanium (YOlU1g'S modulus =
ness, scarf joints (Fig. 28.25(i» are theoretically 110 GPa = 16 x 106 Ib/in2) ideally, the ratio of
the most efficient, having the potential for the maximum thickness (the thickness just
complete elimination of stress concentrations. beyond the end of the joint) of the composite
(In practice, some minimum thickness corre- adherend to that of the titanium should be
sponding to one or two ply thicknesses must 110/55 = 2.0.
be incorporated at the thin end of the scarfed
adherend leading to the occurrence of stress
28.4.2.4 Effects of ductile adhesive
concentrations in these areas.) In theory, any
response
desirable load capability can be achieved in
the scarf joint by making the joint long enough Adhesive ductility is an important factor in
and thick enough. However, practical scarf minimizing the adverse effects of shear and
joints may be less durable because of a ten- peel stress peaks in the bond layer. If peel
dency toward creep failure associated with a stresses can be eliminated from consideration
uniform distribution of shear stress along the by such approaches as adherend tapering,
length of the joint unless care is taken to avoid strain energy to failure of the adhesive in shear
letting the adhesive be stressed into the non- has been shown by Hart-Smith35 to be the key
linear range. As a result, scarf joints tend to be parameter controlling joint strength; thus the
used only for repairs of very thin structures. square root of the adhesive strain energy
630 Mechanical fastening and adhesive bonding

density to failure determines the maximum sta- transverse tension, as a result of which the
tic load that can be applied to the joint. The limiting element in the joint may be the inter-
work of Hart-Smith has also shown that for pre- laminar shear and transverse tensile strengths
dicting mechanical response of the joint, the of the adherend rather than the bond strength.
detailed stress-strain curve of the adhesive can Ductile behavior of the adherend matrix can
be replaced by an equivalent curve consisting be expected to have an effect similar to that of
of a linear rise followed by a constant stress ductility in the adhesive in terms of response
plateau (Le. elastic-perfectly plastic response) if of the adherends to transverse shear stresses,
the latter is adjusted to provide the same strain although the presence of the fibers probably
energy density to failure as the actual limits this effect to some extent, particularly in
stress-strain curve gives. Test methods for regard to peel stresses.
adhesives should be aimed at providing data The effect of the stacking sequence of the
on this parameter. Once the equivalent elastic- laminates making up the adherends in compos-
perfectly plastic stress-strain curve has been ite joints is significant. For example, 90° layers
identified for the selected adhesive for the most placed adjacent to the bond layer theoretically
severe environmental conditions (temperature act largely as additional thicknesses of bond
and humidity) of interest, the joint design can material, leading to lower peak stresses, while
proceed through the use of relatively simple 0° layers next to the bond layer give stiffer
one-dimensional stress analysis, thus avoiding adherend response with higher stress peaks. In
the need for elaborate finite element calcula- practice it has been observed that 90° layers
tions. Even the most complicated of joints, the next to the bond layer tend to seriously weaken
step lap joints designed for root-end wing and the joint because of transverse cracking which
tail connections for the F-18 and other aircraft, develops in those layers, and advantage cannot
have been successfully designed4S,46,56 and be taken of the reduced stresses.
experimentally demonstrated using such Large disparity of thermal expansion char-
approaches. Design procedures for such analy- acteristics between metal and composite
ses which were developed on Government adherends can pose severe problems.
contract have been incorporated into public Adhesives with high curing temperatures may
domain in the form of the A4EG', A4EI' and
J J be unsuitable for some uses below room tem-
IA4EK' computer codes44-46 mentioned previ- perature because of large thermal stresses
ously in Section 28.4.1. Note that the A4EK code which develop as the joint cools below the fab-
permits analysis of bonded joints in which local rication temperature.
disbonds are repaired by mechanical fasteners. Composite adherends are relatively pervi-
ous to moisture, which is not true of metal
adherends. As a result, moisture is more likely
28.4.2.5 Behavior of composite adherends
to be found over wide regions of the adhesive
Organic matrix composite adherends are con- layer, as opposed to confinement near the
siderably more affected by interlaminar shear exposed edges of the joint in the case of metal
and tensile stresses than metals, so that there is adherends, and response of the adhesive to
a significant need to account for such effects in moisture may be an even more significant
stress analyses of joints. Transverse shear and issue for composite joints than for joints
thickness-normal deformations of the between metallic adherends.
adherends have an effect analogous to thick-
ening of the bond layer, corresponding to a
28.4.2.6 Effects of bond defects
lowering of both shear and peel stress peaks.
On the other hand, the adherend matrix is Defects in adhesive joints which are of concern
often weaker than the adhesive in shear and include surface preparation deficiencies, voids
 Adhesive joints 631

and porosity, and thickness variations in the adherends, porosity may grow catastrophi-
bond layer. cally and lead to non-damage tolerant joint
Of the various defects which are of interest, performance.
surface preparation deficiencies are probably Bond thickness variations61 usually take the
the greatest concern. These are particularly form of thinning due to excess resin bleed at
troublesome because there are no current non- the joint edges, leading to overstressing of the
destructive evaluation techniques which can adhesive in the vicinity of the edges. Inside
detect low interfacial strength between the tapering of the adherends at the joint edges
bond and the adherends. Most joint design will compensate for this condition; other com-
principles are academic if the adhesion pensating techniques are also discussed 61 •
between the adherends and bond layer is poor. Bond thicknesses, per se, should be limited to
The principles for achieving good adhesion of ranges of 0.12-0.24 mm (0.005-0.01 in) to pre-
the bond to the adherends (see Chapter 29) are vent significant porosity from developing
well established for adherend and adhesive although greater thicknesses may be accept-
combinations of interest. Hart-Smith, Brown able if full periphery damming or high
and Wong57 give an account of the most crucial minimum viscosity paste adhesives are used.
features of the surface preparation process. Common practice involves the use of film
Results shown in that reference suggest that adhesives containing scrim cloth, some forms
surface preparation which is limited to of which help to maintain bond thicknesses. It
removal of the peel ply from the adherends is also common practice to use mat carriers of
may be suspect, since some peel plies leave a chopped fibers to prevent a direct path for
residue on the bonding surfaces that makes access by moisture to the interior of the bond.
adhesion poor. (However, some manufactur-
ers have reported satisfactory results from
28.4.2.7 Durability of adhesive joints
surface preparation consisting only of peel ply
removal.) Low pressure grit blastingSS,S9 is Hart-Smith45 discusses differences in durabil-
preferable over hand sanding as a means of ity assessment of adhesive joints between
eliminating such residues and mechanically concepts related to creep failure under cyclic
conditioning the bonding surfaces. loading and those related to crack initiation
For joints which are designed to ensure that and propagation which require fracture
the adherends rather than the bond layer are mechanics approaches for their interpretation.
the critical elements, tolerance to the presence In summary, Hart-Smith suggests that if peel
of porosity and other types of defect is consid- stresses are eliminated by adherend tapering
erable45 • Porosity60 is usually associated with or other means, and if the principle discussed
overthickened areas of the bond, which tend in Section 28.4.2.1 of limiting the adherend
to occur away from the edges of the joint thickness to ensure failure of the adherends
where most of the load transfer takes place, rather than the adhesive is followed, crack-
and thus is a relatively benign effect, espe- type failures will not be observed under
cially if peel stresses are minimized by time-varying loading, failures being related
adherend tapering. In such cases60, porosity primarily to creep fatigue at hot wet condi-
can be represented by a modification of the tions, in joints with short overlaps which are
assumed stress-strain properties of the adhe- subject to relatively uniform distributions of
sive as determined from thick-adherend tests, shear stress along the joint length. Additional
allowing a straightforward analysis of the discussion of viscoelastic response of bonded
effect of such porosity on joint strength, as in joints is given62--M.
the A4EI computer code. If peel stresses are There is an extensive body of literature65-71
significant, as in the case of over-thick on fracture mechanics approaches to joint
632 Mechanical fastening and adhesive bonding

durability, based on measurement of energy erations force bonded joints to incorporate

release rates for various adhesives together adherends which are thin relative to their
with analytical efforts aimed at applying them dimensions in the load direction, stress varia-
to joint configurations of interest. In particular, tions through the thickness of the adherend and
Johnson and Malln report fatigue crack initia- the adhesive layer tend to be moderate. Such
tion in bonded specimen configurations with variations do tend to be more significant for
adherend tapering aimed at reduction of peel organic composite adherends because of their
stresses in varying degrees, in some cases prac- relative softness with respect to transverse
tically eliminating them; data in Ref. 92 shear and thickness normal stresses. However,
indicate that crack initiation will occur even a considerable body of design procedure has
with the adhesive in pure shear, for cycling to been developed based on ignoring thickness-
106 cycles above loading levels which are prob- wise adherend stress variations. Such
ably considerably below static failure loads. approaches involve using one-dimensional
The results givenn suggest that for combina- models in which only variations in the axial
tions of peel and shear stressing, total (mode 1 direction are accounted for. Accordingly, the
+ mode 2) cyclic energy release rate can be bulk of the material to be covered here is based
used to determine whether or not cracking will on simplified one-dimensional approaches
occur. However, Hart-Smith reported45 that in characterized by the work of Hart-Smith. The
'thick adherend' test specimens that provide a Hart-Smith approach makes extensive use of
relatively uniform shear stress distribution in closed form and classical series solutions since
the adhesive (see MIL-HDBK-17, Vol. I, these are ideally suited for making parametric
Chapter 7, Section 7.3) which were subjected to studies of joint designs. The most prominent of
fatigue tests in the PABST program40, cycling to these have involved modification of Volkersen26
more than 107 cycles applied at high cycling and Goland-Reissner27 solutions to deal with
rates (30 Hz) were achieved without failure of ductile response of adhesives in joints with uni-
the adhesive, although in certain cases, namely form adherend thicknesses along their lengths,
those involving 6.27 mm (0.25 in) adherend together with classical series expressions to
thicknesses, fatigue failures of the metal deal with variable adherend thicknesses
adherends did result. More study is needed to encountered with tapered adherends, and scarf
resolve some of the apparently contradictory joints. Simple lap joint solutions described
results which have come out of various studies. below calculate shear stresses in the adhesive
for various stiffnesses and applied loadings.
For the more practical step lap joints, the
28.4.3 STRESSES IN ADHESIVE JOINTS
described expressions can be adapted to treat
the joint as a series of separate joints, each hav-
28.4.3.1 General
ing uniform adherend thickness.
Stress analyses of adhesive joints have ranged
from very simplistic 'P over p.: formulations in
28.4.3.2 Adhesive shear stresses
which only average shear stresses in the bond
layer are considered, to extremely elegant elas- Figure 28.27 shows a joint with ideally rigid
ticity approaches that consider fine details, e.g. adherends in which neighboring points on the
the calculation of stress singularities for appli- upper and lower adherends slide horizontally
cation of fracture mechanics concepts. A with respect to each other when the joint is
compromise between these two extremes is loaded to cause a displacement difference 0 =
desirable, since the design of structural joints U u - uL related to the bond layer shear strain by
does not usually depend on the fine details of Yb = o/tb · The corresponding shear stress, T b, is
the stress distributions. Since practical consid- given by Tb = Gbyb• The rigid adherend
 Adhesive joints 633

assumption implies that 0, Yb and tb are uni- one for which ELtL » Eutu)' stretching elonga-
form along the joint. Furthermore, the tions in the upper adherend lead to a shear
equilibrium relationship indicated in Fig. strain increase at the right end of the bond
28.27(c), which requires that the shear stress belayer. The case in which both adherends are
related to the resultant distribution in the equally deformable, shown in Fig. 28.29(b),
upper adherend by indicates a bond shear strain increase at both
ends due to the increased axial strain in
dTj dx = tb (28.6) whichever adherend is stressed at the end
leads to a linear distribution of Tu and TL under consideration. For both cases, the varia-
(upper and lower adherend resultants) as well tion of shear strain along the bond results in an
as the adherend axial stresses axu and axL indi- accompanying increase in shear stress which,
cated in Fig. 28.28. These distributions are when inserted into the equilibrium eqn (28.6)
described by the following expressions: leads to a nonlinear variation of stresses. The
Volkersen shear lag analysis26 provides the
Tu 7'+; TL 7'(1-+)
= = simplest calculation of adhesive shear stresses
for the case of deformable adherends. This
involves the solution of the following differen-
:. 0xu = ax +; 0XL = ax (l-+)
(28.7) tial equation:
d2Tu
where ax = Tit. In actual joints, adherend =
dx2
deformations will cause shear strain variations
in the bond layer which are illustrated in Fig.
28.29. For the case of a deformable upper
adherend in combination with a rigid lower
adherend shown in Fig. 28.29(a) (in practice, which applies to the geometry of Fig. 28.30

CA) RIGID ,gfIENl1Oll

f - Ciii.iii~=~~--r

6 = u - u
U l

Y = 6 I tb
{t] EIlUllIlIIllJj

.
•• d Tu I dx = ~

Fig. 28.27 Elementary joint analysis (rigid adherend model).

634 Mechanical fastening and adhesive bonding

(A) AXIAL RESULTANT DISTRIBUTION (B) AXIAL STRESS DISTRIBUTION

1.7 Q.1
U

-
U
Ie-<
...... 05
"'" u
•..........
.r ..
i»"
iot III ~u
I .,
OF-~~

•
______
I.t 112 . . U
+-~~~

.. . . ro
__~
QI Oil 1
~~-~~----~---,--~~~I_~~~~
o ..., U o.a U .. U Il1 U U 1
xl' xl'
Fig. 28.28 Axial stresses in joint with rigid adherends.

111111II1II IMXIIUI
(A) RIGID LOJER _rAl STMIII AXiAl $!MIN
AllIfI9II

(8) IIJTH AIJfREtm

IIRJIMBlE

IIXIIUI 1111111l1li
AXiAl SIIlI.III AXIAL SIIlI.III

Fig. 28.29 Adherend deformations in idealized joints.

below. The solution for this equation which (28.9)

provides zero traction conditions at the left
where
end of the upper adherend and the right end
of the lower adherend, together with the
applied load T at the loaded ends gives the
resultants as:
T
u
= r{_l_[l
1 + PB
+ sinhP(x - l)/t ]
sinhPI/t
t =
tu + tL
2 ; PB = BJ Bu

Using eqn (28.6) to obtain an expression for

+ PB sinhpx/t }. the shear stress distribution leads to:
1 + PB sinhPI/t '
 Adhesive joints 635

I
0-
I x
~l---i _
-
T- / ~} 7(~ /7
{} ;P S' <0\
~ \-T
7 7 E t
PH 1
+ 1 + PH tanhfll/t
) (28.11)
tb u' u
where
ax = fit
Fig. 28.30 Geometry for Volkersen solution.
Also of interest in the discussion which fol-
T = fla [_1_
1 + PH
coshfl(x -l)/i'
sinh flll t
lows is the minimum shear stress in the joint.
This occurs approximately at x = 1/2, leading
b x to:

+ ~ coshflx/t ] (28.10) (28.12)

1 +P H sinhfll/t
For the case of dissimilar adherends, assuming In the case of ductile adhesive response, the
that the lower adherend is as stiff as or stiffer ratio of minimum shear stress in the elastically
than the upper, the maximum shear stress responding part of the bond line to the yield
obtained from eqn (28.10) is given by: stress in the ductiley responding part is of
interest from the standpoint of joint durability,
BL ~ Bu; to be discussed subsequently. Figure 28.31

... "*L . . . . . .,..,.... --1\. 1DEIJ 4 .JtIiIIR_........a ·-EL --~

I.QWER.tOJIEJIIIID ,~

I~
IFPERAOHEM.,
'15 I
O.
I ~:::::::::::~----..-.l o~~~~~~~~----~_
o D2DA •• U 11.21.41.11.1 2
D 0.2 CIA U ... , '.2 '.4 ,. 1.1 2
1_

t•
II M~~~--~·~
,,.,.
2 .... _ . . . . . . . . . . . . __ ~_~

1
I
1.4

J: 12
1
I ..
•
2
J: OJ!
~i-____~~~==~~ ______
.~~------------~--~--~
o U 1.4 ... ..1 1 U U 1.1 1.1 2 • Cl2 1.4 . . U I Lt 1.4 , . U 2
If_ 1-

Fig.28.31 Comparison of adherend and bond stresses for E] = Eu and E] = lOEu- equal adherend thicknesses.
636 Mechanical fastening and adhesive bonding

0.5
0.4
i
1/1

i ~:
0.3

O~~~~~~+4~~~~~~~~~~

0.5 1.5 2.5 3.5 4.5 5.5 6.5 1.5 8.5 9.5 10. 11. 12.
555
f /t -----
Fig. 28.32 Comparison of average and maximum shear stress vs.l/t.

shows the distribution of axial adherend An additional point of interest is a typical

stresses and bond layer shear stress for two feature of bonded joints illustrated in
cases corresponding to Eu ; ; ; EL and EL ; ; ; lOEu Fig. 28.31(d) which gives the shear stress dis-
with tu = tv 13 ; ; ; 0.387 and 1/ t ;;;;; 20 for both tribution for equal adherend stiffness, namely,
cases (giving 131/t ; ; ; 7.74) and a nominal the fact that high adhesive shear stresses are
adherend stress ax = 10. As in the approximate concentrated near the ends of the joint. Much
analysis given earlier, the shear stresses given of the joint length is subjected to relatively low
by eqn (28.10) are maximum at both ends for levels of shear stress, which implies in a sense
equally deformable adherends (Bu ; ; ; BL ); for that that region of the joint is structurally inef-
dissimilar adherends with the lower adherend ficient since it does not provide much load
more rigid (BL > Bu)' the maximum shear stress transfer. However, the region of low stress
obtained from eqn (28.10) occurs at the right helps to improve damage tolerance of the joint
end of the joint where x = 1, again as it did for since defects such as voids and weak bond
the approximate analysis. strength may be tolerated in regions where the
Figure 28.32 compares the behavior of the shear stresses are low, and in joints with long
maximum shear stress with the average shear overlaps this may include most of the joint. In
stress as a function of the dimensionless joint addition, Hart-Smith has suggested51 that
length, lit, for equal adherend stiffnesses. The when ductility and creep are taken into
point illustrated here is the fact that although account, it is a good idea to have a minimum
the average shear stress continuously shear stress level no more than 10% of the
decreases as the joint length increases, for the yield strength of the adhesive, which requires
maximum shear stress which controls the load some minimum value of overlap length.
that can be applied without failure of the Equation (28.12) can be used to satisfy this
adhesive, there is a diminishing effect of requirement for the case of equal stiffness
increased joint length when '7 = 131/t is much adherends. The two special cases of interest
greater than about 2. again are for equal adherend stiffness and a
 Adhesive joints 637

rigid lower adherend, since these bound the relatively obvious due to the offset of the two
range of behavior of the shear stresses. As a adherends which leads to bending deflection
practical consideration, we will be interested as in Fig. 28.33(b). In the case of double lap
primarily in long joints for which f3l!t » l. joints, as exemplified by the configuration
For these cases eqn (28.11) reduces to: shown in Fig. 28.34, the load path eccentricity
is not as obvious, and there may be a tendency
f3l/t » 1; to assume that peel stresses are not present for
this type of joint because, as a result of the lat-
eral symmetry, there is no overall bending
deflection. However, a little reflection brings
to mind the fact that while the load in the sym-
BL = Bu; l'b)max'" ~ 13 ax (28.13) metric lap joint flows axially through the
central adherend prior to reaching the overlap
Thus, for long overlaps, the maximum shear region, there it splits in two directions, flowing
stress for the rigid adherend case tends to be laterally through the action of bond shear
twice as great as that for the case of equally stresses to the two outer adherends. Thus
deformable adherends, again illustrating the eccentricity of the load path is also present in
adverse effect of adherend unbalance on shear this type of joint. As seen in Fig. 28.34(c), the
stress peaks. shear force, designated as FSH' which repre-
sents the accumulated effect of l'b for one end
of the joint, produces a component of the total
28.4.3.3 Peel stresses
moment about the neutral axis of the upper
Peel stresses, i.e. through-the-thickness exten- adherend equ~l to FSHl' /2. (Note that FSH is
sional stresses in the bond, are present because equivalent to T /2, since the shear stresses react
the load path in most adhesive joint geome- this amount of load at each end.) The peel
tries is eccentric. It is useful to compare the stresses, which are equivalent to the forces in
effect of peel stresses in single and double lap the restraining springs shown in Fig. 28.34(b)
joints with uniform adherend thickness, since
peel stresses are most severe for joints with
uniform adherend thickness. The load path (A) OOAlLE LAP JlINT
eccentricity in the single lap joint (Fig. 28.33) is

CAl IHlRIIIIIJ ImI3IIY (lEIIlUWJ)

(8) IIRlIIlll GEIJETRY (lIJIJIl JlJNJ)

(8) IIlEIIT EFFECT (J' SIQA Rm5

AlBtnlnll9
flfoct
Cll'JIIInI

'om~'E_~
(C) w.melT5 Cf IOEHT

Fig. 28.33 Peel stress development in single lap Fig. 28.34 Peel stress development in double lap
joints. joints.
638 Mechanical fastening and adhesive bonding

and (c) have to be present to react the moment It is important to understand that peel
produced by the offset of FSH about the neutral stresses are unavoidable in most bonded joint
axis of the outer adhered. Peel stresses are configurations However, they can often be
highly objectionable. Later discussion will reduced to acceptable levels by selecting the
indicate that effects of ductility significantly adherend geometry appropriately.
reduce the tendency for failure associated with
shear stresses in the adhesive. On the other
28.4.3.4 Effects of joint geometry
hand, the adherends tend to prevent lateral
contraction in the in-plane direction when the In this section the behavior of joints is consid-
bond is strained in the thickness direction, ered with linear response of the adhesive in
which minimizes the availability of ductility shear assumed. Effects of ductility will be con-
effects that could provide the same reduction sidered later.
of adverse effects for the peel stresses. This is
illustrated by the butt tensile test shown in Fig.
28.4.3.4.1 Single and double lap joints with
28.35 in which the two adherend surfaces adja-
uniform adherend thickness
cent to the bond are pulled away from each
other uniformly. Here the shear stresses asso- Double lap joints will be considered first since
ciated with yielding are restricted to a small they are somewhat simpler to discuss than sin-
region whose width is about equal to the gle lap joints because of deflection effects in the
thickness of the bond layer, near the outer latter. Shear and peel stresses in double lap
edges of the system; in most of the bond, rela- joints with uniform adherend thickness were
tively little yielding can take place. For organic treated by Hart-Smith34• For the shear stresses,
matrix composite adherends, the adherends the type of analysis discussed in Section 28.4.3.2
may fail at a lower peel stress level than that at can be applied with suitable changes in nota-
which the bond fails, which makes the peel tion, i.e. the expressions for the shear stresses
stresses even more undesirable. given in eqns (28.11) and (28.12) can be applied
with subscripts 'i' and '0' ('inner' and 'outer')

t t t substituted here for 'L' and 'U' ('lower' and

'upper') used in eqns (28.6-28.11); in addition,
the outer adherend thickness in the earlier
equations is now equivalent to half the thick-
ness of inner adherend because of vertical
symmetry of the double lap joint. However, we
will also introduce the effects of thermal mis-
match effects in the following expressions for
Bond later reference. The notation used here is:

Edge Region
(01sta-J1IlI
Strailli)

BB.
B :
o
B. (a
1
o - a)AT;

Fig. 28.35 Shear stresses near outer edges of butt (28.14)

tensile test.
 Adhesive joints 639
A

where ao' a i are thermal expansion coefficients In the absence of thermal effects ( Tth = 0) and
and fiT is the temperature change. assuming that Bi ~ Bo' the maximum value of
Note that T is related to the resultants (axial the shear stresses occurs at the right end of the
adherend stress times thickness) at the ends of joint as noted earlier (Fig. 28.31). With thermal
the joint as shown in Fig. 28.36. The shear effects present, the situation is complicated by
stresses are then given by: the sign of ITth which is positive if (a o - a) and
AT have the same sign and negative otherwise.
T = a [_1_ cosh,8(x - I)/t The peel stresses in the double lap joint are
b ,8 x 1 + Ps sinh,81/t described by a beam-on-elastic foundation
type differential equation of the form:
+ ~ cosh,8xJt ] +,8fJ [COSh,8x/f
1 + PB sinh ,81/ t th sinh ,81/ t d 4a b td 1 dTb
dx4 + 4yab = 2"" to dx (28.18a)
_ cosh,8(l - x)/f ] (28.15)

I
sinh,81/t (28.18b)

2Hii:6
The solution to eqn (28.18) depends on
2t to I (A) OOUllLE STRIIP JOINT
whether a strap joint or a lap joint is consid-
ered. The exact form of the solution contains
Iii HI products of hyperbolic and trigonometric
functions but for the practical situation of
E
j
\
o
E ",-, I
joints longer than one-or-two adherend thick-
- nesses and,8 << Yd' are given by:
21-1 \ ~' I .:-_1'1'
Double lap joint,
r--1--11 (BJ OClJBLE LAP JO INT
X X- I)
1

a =:;a)
b bmax
(cosydt-e-'Ydxt-cosyd
__
t
erix-l)/t
Fig. 28.36 Symmetric double strap / double lap
joints. Double strap joint,
For the usual situation in which the overlap is 1 x-I )
( cosy -e-'Y~t--cosy __ eYix-l)!t
a "'a) X
long enough so that ,8IJt is greater than about b bmax dt 2 d t
3, the peak shear stresses at the ends of the
joint are given by: (28.19)
For the case of identical adherends, the maxi-
x = 0; Tho = ,8( 1 : P s ax - lTth); mum peel stresses, which occur at x = 0, are
given by:
x = I; Tbl = ,8( 1 :Sps ax + lTth) (28.16) (28.20)

and for the special case of equal adherend stiff- l'b)max = ,8 a/2 - ,8lTth
nesses (Bi = B) we have: (identical adherends)
Here l'b)max is taken to be the peak stress at the
left end of the joint, corresponding to the
expression for x = 0 in eqn (28.16), since the
out-of-plane normal stresses are compressive
(28.17)
at the other end of the joint for a tensile load.
640 Mechanical fastening and adhesive bonding

For compressive loading, the situation would

reverse for the double lap joint (Fig. 28.36(b»,
Quarter Plane Symmetry
with the positive out-of-plane stresses occur-
ring at the right end of the joint (x = I), in the
case of the double snap joint (Fig. 28.27(a)), the
peak out-of-plane stresses would be compres-
sive at the left end of the joint and would not
occur at x = I, since the inner adherends butt
against each other there and act as a continu-
ous element.
Effects of thermally induced stresses will be

IX
discussed in a later section. Figure 28.37 com-
pares the peel and shear stress distributions 2.00.
for b"th = 0, in a typical joint having balanced
adherend stiffnesses (the sum of the outer
1.80
1.60
lll1.40
t
adherend stiffnesses equal to the inner
adherend stiffness) whose parameters are
~~ 1.20
1.00
/

listed in Fig. 28.37(b). The diagram at the top m0.80

indicates the origin of x at the left end of the iii 0.60

overlap. The distribution of peel stresses is § 0.40

Cl 0.20
"",-_____ ~
K-.x_x_x
somewhat more concentrated near the ends 0.00 +--+-----I--+-_+_--+-+--+----ir-_+__
than that of the shear stresses and the peel 0.000.200.400.600.801.001.201.401.601.802.00
(a)
x-
stresses at the right end of the joint are nega-
tive. In addition, the compressive peak at the 1.5 II
right end is half as great for the strap joint as Ex - - 20 000 Gb - - 150 Eb - - 500
for the lap joint, which is the result of the . t - - 0.1 tb - - 0.01 I - - 2
Sig - x bar - - 10
restraint of bending rotations in the strap joint
for a gap which is essentially zero. If the load- L--'-=..L:O.2~0.4_
..---r-r-.a ~-a-a~.a: ~~
•

ing were compressive rather than tensile, the .a"'" 0.6 0.8 1 1.2 1.4 1.6 1.8 2

:A
inner adherends would bear directly on each x-
other and no shear or peel stress peak would D,"~e ,bap ;0;01
occur at the gap, whereas in the lap joint the
right end of the overlap would experience the Double lap joint
same peak stresses for compressive loading as (b)
the left end does for tensile loading. Fig. 28.37 Bond stresses in double lap/strap joints;
The situation for the single lap joint (Fig. (a) bond shear stress distribution; (b) bond peel
28.38) is complicated by the effects of lateral stress distribution.
deflection which are indicated in Fig. 28.39.
Literature for the following discussion on the The effects of lateral deflections on the bond
single lap joint is given53,54,73-78. stresses were first evaluated by
The deflection effect is dependent on the Goland-Reissner27 for the case of equal
joint load, given in terms of the quantity adherend thicknesses, so that tu and tL can be
W/2(8)1/2tu' where denoted by t in the following. The lateral

U tu~(JJ=~(12 ;J;D
deflections can then be stated in terms of a
= u = 1~EutJ(28.21)dimensionless ratio, k, with respect to the
adherend thickness, and are of the following
form:
 Adhesive joints 641

OISPLAce.t:NTS

Fig. 28.38 Single lap joint geometry.

effects. The most accurate expression (given in

CAl INERAI!I tmmr (ZBII LOOJ) Ref. 74) is fairly elaborate and will not be
repeated here; an expression of intermediate
accuracygiven in Ref. 74 retains the essential
form of the GR result but gives considerable
improvement over the GR expression for thin
adherends:
tanhW
(28.23a)
Fig. 28.39 Effects of bending deflections in single k = tanhWo + "'J8Cpt~(W/2Cp)
lap joints. where

C = (1 + 1.2 p + 1.4 --2)112-

p t Pt '
t sinh Ux/t 1 t + tb
xs;.Io;w = k2 sinhUljt - 2 - L - x tb
Pt = t;A = lltu ;Ao = lJtu (28.23b)
u
< < . _ sinh[Uj\l8(x - L)/t]
The original GR expression for k is recovered if
10 - x_I + 10 , w - W2 sinh[Ulfj8t]
Cp is set equal to 1 corresponding to tu » tb (i.e.
relatively thick adherends) and tanh Wo is like-
tu 1 t + tb t
+2 - 2-L- x ; W2 = 2(1 - k) (28.22) wise set to 1 corresponding to very long outer
adherend lengths. A plot of k against the
The Goland and Reissner (GR) expression for adherend loading stress ais given in Fig. 28.40
the parameter k has been re-examined by for two different values of adherend thickness
Hart-Smith36 and more recently by Oplinger22, corresponding to bond thickness-to-adherend
based on the discussion in Ref. 74, the thickness ratios (Pt in eqn (28.23b» of 0.5 and
Goland-Reissner expression appears to pro- 0.1. This plot suggests that k is fairly constant at
vide adequate accuracy unless the adherends a value of about 0.25 for a wide range of
are excessively thin, not more that one or two applied stress values once the initial drop has
times the bond layer thickness, in which case occurred. The effect of bond-to-adherend thick-
the expressions given in Refs. 73,74 and pro- ness ratio is not particularly great and can
vide corrections for bond layer thickness perhaps be ignored for the most part.
642 Mechanical fastening and adhesive bonding

0.8

0.7 Ell = 10,000; Gb = 150 ; 1;, = 500

It, = 0.01 ; fo = 610 t ; f = 61 t

,
:.I. 0.5 \
OA l t.o.1

/
t~MIZ

l
---
0.3 ~
t--: / I
I
0.2

0.1

o
o 5 10 15 36 40 50

Fig. 28.40 k parameter vs. adherend loading stress.

The lateral deflections of the joint have a

1 +l-X] x-l-I
significant influence on the stresses in the + Ucos Y. 0 t exp Ys to
bond layer, which show this through the pres-
ence of the k parameter in expressions for
them. The shear stress is given by: x-I X-I)
+ [Y.(cos Ys - t _0 - sin Ys - t _O
f "" 0 {.l(1 + 3k)cosh [2,8(x - L)lt]
b "4 sinh (fJAI)

+ 3 U(1 _ k) cosh [U(x - L)lY8t]} (28.24)

8<J8 sinh (w../2<J8) where Ys = (6 Eb
tiE t )1/4 (28.26)
" b
where p and U are given in eqns (28.14) and
(28.21). Equation 28.24, which represents a The maximum stresses in the bond layer are
slight modification of the GR expression, given by:
reduces to the latter for small values of Wit. In Maximum bond shear stress,
addition, the peel stresses, for joints in which
the overlap length is more than one or two fb)max = Ox[ ~ (1 + 3k)/tanh,Bl
adherend thicknesses (essentially the only case
of practical interest) are given by

a =0 kyo (28.25)
+ 8~8 U(1 - k) I tanh ( ~~ )] (28.27)
b "2
Maximum bond peel stress,
1 + 1- x 1 + 1- x )
p x {[Ys(cos Ys 0 t + sin Ys 0 t (28.28)
 Adhesive joints 643

Figure 28.41 gives a comparison of the maxi- stresses vary essentially in proportion to the
mum bond stresses as functions of the loading load even in the single lap joint, as just dis-
stress ax for two different adherend thick- cussed. The stresses are plotted in this figure
nesses in a joint with a bond layer thickness of as a function of adherend thickness with the
0.01. It is interesting to note that the peel and adherend axial modulus as a parameter. The
shear stresses take on quite similar values. trend toward higher bond stresses and there-
Since the maximum peel stress varies approx- fore a greater tendency toward bond failure
imately as Y~ according to eqn (28.27) (the with increasing adherend thickness which was
contribution of U being relatively minor), the discussed in Section 28.4.2 is clearly borne out
relationship for y5 given in eqn (28.26) sug- in these curves. Note also that reduction of the
gests that the peel stresses should be expected adherend modulus tends to aggravate the
to vary as (t/tb)l/2, while the same variation is bond stresses. In addition it is apparent that
seen from eqn (28.27) for the maximum shear there is considerable separation between the
stresses since f3 also contains (t/tb)1I2 as a fac- peel and shear stresses in the case of the dou-
tor. Thus both stresses should vary with the ble lap joint, the peel stresses for the latter case
thickness ratio by the same factor. The fact that being smaller. This reflects the fact that the
they are numerically close together for all peel stresses vary linearly as yd defined in eqn
stresses is partly due to the effect of other (28.18b) and therefore vary as (t/tb)l/4 rather
parameters that enter into eqns (28.27) and than as (t/tb)1I2 as in the single lap case. Thus,
(28.28) and partly due to the fact that k does peel stresses for double lap joints are not as
not vary much with load for ax greater than 5. much of a factor in joint failure as they are in
A slight nonlinearity can be observed in the single lap joints, although they are sti11large
curves of Fig. 28.41 for the lower loading enough relative to the shear stresses that they
stresses. can not be ignored.
Figure 28.42 gives a comparison of maxi- Failure characteristics of single and double
mum bond stresses in single and double lap lap joints will be discussed below. If the
joints for a fixed value of the loading stress ax. adherends are thin enough, failure in double
For loading stresses above this value the bond lap joints should be in the form of adherend

14 t=O.02;
t-O.I;
"'2. ",=10
£.=10000. Gb=lS0. s.=~

".0. ",=50
12 _&lr_ _

_ s - ..... .

o ~~----------+-------------~---------
o 10 20 - 30 40 50
ax ---

Fig. 28.41 Maximum bond stresses in single lap joint, bond thickness = 0.01.
644 Mechanical fastening and adhesive bonding

Double Lap Joint

_Sl....
Single Lap Joint
s_ _
4

t
¥~
3.53 Ex.&.ODD

.
::
~.--.
a " CI~

:~Z-=:=-
0.5
o ____ ______________+ -____
0.5 ~
Ex- 20,000
O+-----_------__- - - - - - + - - - - - - - - - t
~ ~ ~

0.02 0.04 0.06 0.08 0.1 0.02 0.04 0.06 0.08 0.1
Adherencl ThIckMse _ AdherendThlckn... _

O'x= 10 Gt,= 150 ~= 500

to= 50 1=10

Fig. 28.42 Maximum bond stresses in single and double lap join~, fixed Vx = 10.

axial (tensile or compressive) failure. For sin- on single lap joint performance is quite long
gle lap joints, adherend bending stresses are range. Figure 28.43 shows that for a joint with
significant at the ends of the overlap as indi- an adherend thickness of 2.54 mm (0.1 in) the
cated in Fig. 28.39; using standard beam bond stresses do not reduce to their minimum
formulas, the maximum axial stress for com- values until the overlap length reaches a value
bined bending and stretching (the latter stress in the range of 10-12.7 cm (4-5 in), for a load-
corresponding to the single lap joint in tension ing stress of 69 MPa (10 ksi). Double lap joints
loading) for the bending deflection given in also require some minimum length before
eqn (28.22) can be expressed as stresses settle out as a function of overlap
length, but in this case the stresses reach mini-
a)D1JJX = ax3(1 + t/t)k (28.29)
mum values with respect to overlap length for
The maximum adherend axial stress is lengths on the order of 5 to 10 adherend thick-
largest for adherends which are particularly nesses, in the present case amounting to
thin with respect to the bond thickness; these 1.3-2.5 em (0.5-1 in).
will be prone to brittle bending failures for
composite adherends or to yielding associated
28.4.3.4.2 Effects of adherend tapering
with bending for metal adherends. Hart-Smith
discusses difficulties with the use of standard In this section we will consider joints with
single lap shear test specimens50• The problem adherend thicknesses which vary along the
is that adherent bending failures are likely to joint length. These include the configurations
occur with such specimens rather than bond shown in Figs. 28.44 and 28.45, namely, double
failures and test results obtained in such cases strap joints with tapered outer adherends and
tend to be irrelevant and misleading. scarf joints as well as step lap joints. As dis-
One additional characteristic difference cussed in Section 28.4.2.2, tapering the outer
between single and double lap joints should adherends of strap joints as in Fig. 28.44(a) is
be discussed. The effect of lateral deflections beneficial mainly for reducing or eliminating
 Adhesive joints 645

4.5
4

t3.5

i
III 3
2.5
~ 2 IAdianIldthidcnass. 0.11
CD 1.5
SiIJnrt'x bar - - 10
1
Gb--150 Eb--500 &--5,000
0.5
tb - - 0.01 (M. tdla'8ftd length - - 20
O~--~--~--+---+---+---+---+---~--4

1 2 3 4 5 6 7 8 9 10

Fig. 28.43 Effects of overlap length in single lap joints.

I 1
r---'"1
I
(A) PARTIALLY TAPERED STRAP JOINT
I
I
1--
I

Triangular Element
Fig. 28.45 Generic step lap joint.

(8) SCARF JDI NT

(axial modulus times maximum thickness) in
each adherend. For tapered portions of strap
joints it is fairly accurate for the peel stresses,
Bond and holds approximately for the shear stresses
if the tapered portion is not too long. Note that
Fig. 28.44 Tapered double strap and scarf joints. eqn (28.30) implies that the bond stresses are
constant along the length of the joint and can be
reduced to any arbitrary level by making til
peel stresses, while scarf and step lap joints small enough, i.e. making the joint long enough
(Figs. 28.44(b), 28.45) can eliminate shear stress with respect to the adherend thickness.
peaks as well as peel stresses. Moreover, the effect of til on the peel
With both tapered outer adherends and stresses is quite strong, being governed by the
scarf joints, it can be shown that the bond square of the thickness-to-Iength ratio. This is
stresses can be related to the ratio of taper especially important in the case of outer
length to thickness by adherend tapering in strap and lap joints as a
means of reducing peel stresses to a manage-
(28.30)
able level.
This relationship is quite accurate for scarf Step lap joints (Fig. 28.45) represent a com-
joints having the same maximum stiffness promise version of the scarf joint which can
646 Mechanical fastening and adhesive bonding

take advantage of the layered structure of the tapered double strap joints, scarf joints and
composite adherend. The average slope of the step lap joints. The overall approach is to aim
region represented by the line through the for a highly efficient joint which reduces the
steps in Fig. 28.45 tends to control the average effects of shear and peel stress concentrations
shear stresses developed in the bond. Within at the ends of the joint. Ideally we would like
each horizontal section, equivalent to the tread to achieve the joint strength provided by the 'P
of a staircase, the behavior is analogous to a over A' concept obtained with the case dis-
joint with constant adherend thickness, and cussed in Section 28.4.3.2 for perfectly rigid
the differential equation given in eqn (28.10) adherends (Fig. 28.27), in which increasing the
(Section 28.4.3.2) applies locally when tu and tL joint length indefinitely brings the shear stress
are adjusted to match the local situation. An in the bond down to any required level regard-
expression similar to eqn (28.12), i.e. for the jth less of the magnitude of load being supported
step, by the joint. While tapering does reduce the
r)
bi max =Rf(
P i - = -1- - 1
(28.31)
peel stresses markedly in the tapered strap
joint as will be seen below, shear stress peaks
t 1 + PSj sinhP//t
can not be avoided, and the law of diminish-
PB
+ 1 +P Sj
1)
tanhf3// t
ing returns continues to prevail with regard to
increasing the joint length to obtain greater
load capacity; however, adhesive ductility will
enhance the strength beyond what elastic
where Pi [ t2 (~
= Gb -
1 1 ))112
+ B. ; analysis suggests.
tb UJ LJ Double strap joints with tapered outer
- tu + tL , adherends are considered in Figs. 28.46 to
t = J 2 J; PBj = BL/ BUj 28.48. Figure 28.46 indicates the tapered con-
figurations that are considered. Figure 28.47
gives the maximum shear stresses within each gives some shear stress predictions for joints
step, and the overall solution is a chain of such with uniform adherend thickness for compari-
expressions with allowance for continuity of son with the tapered cases which are
the shear strain and resultants, TUj and TLj at considered in Fig. 28.48.
the points where neighboring steps join. In Figure 28.46 defines the notation used in
each step of the joint the shear stresses will Fig. 28.45 in terms of 'fully tapered' outer
have a distribution similar to that of Fig. 28.36, adherends (Fig. 28.46(a», partially tapered
the size of the peaks being governed primarily adherends in which the taper extends only
by the length of the step through the parame- part of the length of the joint (Fig. 28.46(b» and
ter P.l./t. The aspect ratio for the step, lit, can fully tapered adherends with an 'initial rise',
in p~inciple be kept small enough to] almost i.e. in which the thin end of the adherend does
completely avoid any peaking by using a large not come to zero thickness. (The term '% initial
number of steps and keeping the length of rise' implies that the rise is expressed as a per-
each one small. In practice, the number of centage of the maximum adherend thickness.)
steps is governed by the number of plies in the The three cases considered in Fig. 28.48 can
laminate. In addition, if the joint is used to be compared with the case for uniform
connect a composite adherend to a metal com- adherends with equal upper and lower
ponent, machining tolerance requirements adherends modulus (Eu = EL ) in Fig. 28.47. For
and cost considerations for the metal part the situation of no initial rise, two cases are
enter into the selection of the number of steps. considered in Fig 28.48, the case of 50% taper
The following discussion will address the and that of full taper. There is an appreciable
specific benefits of adherend tapering in difference in the shear stress distribution at the
 Adhesive joints 647

(A) Fully tapered - - 00 initial rise left end of the joint for these two cases, but the
peel stress distribution is essentially unaf-
fected. For both the full taper and 50% taper
- - - 2 --~l/ 0.2 cases, a minor tensile secondary peel stress
peak is present at the right end of the region
T under consideration (near the midpoint of the
strap joint). The peel stress expression in eqn

-
(28.30) gives a good estimate of the peel stress
level at the left end of the joint, and the result
of the estimate is so small for both cases that
the difference is not distinguishable in Fig.
28.48. However, in the case of an initial rise of
(8) 50% tapered jo i nt only 1/4 of the maximum outer adherend
thickness (25% initial rise), significant peel

-
--t 1.0 I - -
stresses arise at the left end of the joint, in fact,
about 80% of the level occurring for the case of
no tapering. The initial rise also causes a
greater increase in shear stress at the left end
of the joint than in the case of 50% taper. Thus,
tapering is advantageous mainly as a way of
(C) Full taper - - 25'" initial rfse
eliminating the effects of peel stresses in dou-
ble strap joints. Once this is accomplished, the
0.05 initial rise
effects of peel stress peaks can be controlled to
a significant extent by taking advantage of
adhesive ductility. Tapered strap joints can not
achieve the ideal behavior which is possible
with scarf or step lap joints, but they provide a
simpler solution to good joint performance if

- the adherends are thin enough.

Shear stress distributions in scarf joints (Fig.
28.49) are given in Fig. 28.50. Practical scarf
Fig. 28.46 Tapered strap joints under consideration.

Shear Stren - - Uniform Adherend Thickness

EU • 1,000 to. u.. • 0.2

Gb = 1M Eb. IGO til. 0.01 :'/
Owdop IengtIt • z :1
.,
:1

4 S9M . x bar. 10 ELlEU=4 ~ ..../

ELlEU=2 ~ \:)'

ELlEU=l "-
~ ~I
/,
"r /

.' ,
r'" ,

, .. ';,/'
.....; .......

o o.z G.4 41.1 0.8 1~ 1A 1.1 1.1 2

x Fig. 28.47 Shear stresses in unta-
pered strap joints.
648 Mechanical fastening and adhesive bonding

(A) Shear Stress •• Double Strap Tapered Adherends

5

t ~: FuR taper - - 21% Inlllal rise Initial rise

3.5
I 3
2.5
J
II)

2
II)
1.5
1
0.5
O~~~~-+-+-+-r-r-r~~+-+-r-~~~~

o ro ~ ~ ~ ~ H ~ ~ " 1~

(B) Peel stress •• Double Strap Tapered Adherends

8
5. Notaper

t V~VFUI
4
3 '.
taper. 25% Initial rift

j 2 ". Full taper or 50% taper

.. 1
:. o
·1
0:2 .....0..4·· 0.8 0.8 1.2 1.4 1.8

·2
Gb " 180 Eb. IGO tb. 0.G1
Sigma -x bar-10

Fig. 28.48 Stresses in tapered double strap joints.

joints are arranged in a symmetric double lap the scarf joint is the effect of adherend stiffness
configuration which avoids bending effects. unbalance (Eo *" E;; '0' and 'i' refer to the outer
Figure 28.49 represents a balanced stiffness and inner adherends as in Fig. 28.36). The
design for dissimilar materials, by achieving a results given in Fig. 28.50 represent the effect
continuous thickness change over the length of varying degrees of stiffness unbalance. The
of the joint. The most important parameter for ratio of peak-to-average shear stresses com-
pare well with the values given by
Hart-Smith37, although the latter did not give
1-----1·- 2 -----j J the distribution of stresses along the length of
----~I~~~=2J/1---- the joint. For fairly sizeable unbalances, up to

\ 0.2,
/f
0.4
4:1, the maximum shear stress peak is not as
great as that observed in Fig. 28.47 for the uni-
form adherend case. However, it is clear that a
Ii, E=16 nsi QI graphite epoxy stiffness unbalance will increase the shear
E=8 msi
stress peak and weaken the joint. For the equal
Fig. 28.49 Stiffness-balanced scarf joint configura- stiffness case the shear stress is constant and
tion. equal to the average stress at all points.
 Adhesive joints 649

3.5
EL/IEU=4
EU=I,OOO tU=L=.2
3
Gb =110 Eb" 500 tb II 0.01
2.5
Sigma-x =10
I
l! 2
t;;
J
U)
1.5

1
--- ...... ... ---
-------
.'"
1
---------- ............ .
....... .- .. -
.~

..... "' ....... . . - ....

0.5

0
0 0.2 0.4 0.6 0.8 1.2 1.4 1.6 1.8 2
x
Fig. 28.50 Shear stresses in scarf joints.

If there is no reason why the joint cannot be might be lost if performance in hot wet envi-
configured as in Fig. 28.49 for dissimilar mate- ronments is required.
rials so as to take advantage of the benefits of Step lap joints79,80 are treated in Figs. 28.51
the balanced stiffness case, then in principle, to 28.53. Figure 28.51 shows a generic joint
the scarf joint provides a near ideal solution to configuration that is introduced to illustrate
achieving as much load capacity as is required some of the effects of design parameters on
in any situation without overstressing the stresses in the joint. The results presented in
bond layer. However, the dimensions of the Figs. 28.52 and 28.53 were generated for this
joint may grow too large to be practical for discussion using a linear elastic response
high joint load. In addition, an extremely good model for the adhesive; in practice, consider-
fit, for example, to tolerances on the order of able strength capability of the adhesive is
the bond thickness over large lengths, has to unused if elastic response of the adhesive is
be maintained to ensure that the joint can assumed; Fig. 28.54 taken from the discussion
maintain uniform load capacity over its by Hart-Smith80 is an example of joint design
length. Thermal stresses will also be a factor in using elastic-plastic response for the adhesive.
various combinations of dissimilar materials The elastic adhesive model used in Figs. 28.52
which will prevent the ideal form of behavior and 28.53 is adequate for illustrate some of the
from being achieved. In terms of the Hart- controlling parameters on joint design. These
Smith approach to avoiding creep failure results are based on the classical Volkersen-
under slow cyclic loading, the balanced scarf type analysis with provision for resultant and
joint is at a disadvantage in not providing a shear strain continuity at the interfaces
shear stress minimum. For this situation the between neighboring steps, as discussed pre-
allowed load would have to be limited to pre- viously. The 5-step design in Fig. 28.52 and the
vent environmental conditions corresponding 10-step design in Fig. 28.53 were chosen with
to hot wet exposures; thus the advantage of the following characteristics:
the scarf joint in eliminating stress peaks
650 Mechanical fastening and adhesive bonding

~02
loEI 1,25
e-I
I II

0,4

Ti, E ;: 16 mi
Quasl-lsotr~l1C carbon Epoxy,
Fig. 28.51 Step lap joint
E :: 8 lI'Si
• configuration.

• Except for the first and last steps, the As discussed above, the joint design shown in
adherend thickness was equal for each step. Fig. 28.5437 represents a practical joint design
The first and last thickness increments were which accounts for several considerations that
half those of the generic steps. the simplified elastic analysis approach used
• The lengths of each step were chosen with a for Figs. 28.52 and 28.53 neglects. The neglect
fixed value of the parameter '1sl = Pl/tj' of ductility effects has already been men-
where lj is the length of step j. J tioned. In addition, the use of as large a
The half-incremented end steps gave a more
uniform shear stress distribution than main-
taining the same increment for all steps. The
symmetric joint configuration shown in Figs.
28.51 and 28.54, the thickness increment for
the outer adherend (composite) was greater
than that for the inner adherend by the inverse
of the modulus ratio, to achieve the desired
stiffness balance for the dissimilar adherends.
The parameter ETA listed in Figs. 28.52 and
28.53 is equivalent to TIs! defined above. This • &177 . . . . . . . . 1. . . U7I Uta . . .
parameter essentially controls the length of the X

joint; both Figs. 28.52 and 28.53 show an

increase in joint length with '101 ('ETA' in the
two figures). Note further that the load capac-
ity of the joint in terms of the allowed
resultant, 5, listed as 'NBAR' in Figs. 28.52 and
28.53, which provides for the required bond
shear stress limitation of 5 (ksi for the units
mentioned earlier) shows a general increase
with joint length, but with diminishing
increase when Tlol gets much beyond 3. Table
28.1 gives a summary of the results shown in
the two figures. Fig. 28.52 Shear stresses in 5-step joint.
 Adhesive joints 651

presented by Corvelli and Saleme79; this was

later enhanced by Hart-Smith37, under NASA
u • funding, to provide for elastic-plastic
.- response, culminating in the A4EG and A4EI

I~
programs44--46, to allow for variations in thick-
ness, porosity, flaw content and moisture
t! content in the bond layer. Hart-SmithBO,81 notes

...
1 that in mathematical treatments of step joints,
all properties have to be constant within each
• ... . . . . . . . . .lin .... 1.11. , . . , • • step; however, in an actual joint such as that
• shown in Fig. 28.54, artificial breaks may be

U
• inserted to permit changes in porosity of bond
thickness.
•
I~ 28.4.4 MECHANICAL PERFORMANCE OF
J,! ADHESIVES

... t
28.4.4.1 Ductile response of adhesives
• un..., UJ 1... , ... U17 q ..
Figure 28.55 taken from the 1983 edition of the
• DoD jNASA Advanced Composites Design
U
• Guide81 show shear stress-strain response
• characteristics of typical aerostructural adhe-
sives. Figure 28.55(a) represents a relatively
I~ ductile film adhesive, FM73, under various

...
I.! environmental conditions, while Fig. 28.55(b)
t represents a more brittle adhesive (FM400)
under the same conditions. Similar curves can
~~~~~~~~~~~~~
• UtI . . ,_a..Itm ........... be found in other sources 62 • Temperature
• dependence and strain rate dependence of the
Fig. 28.53 Shear stresses in 10-step joint.
stress-strain characteristics are important
characteristics; these are also addressed in the
results shown in Ref. 62. Even for the less duc-
number of steps as 10 in Fig. 28.53 may not be tile material such as that represented in Fig.
practical. The joint design shown in Fig. 28.54 28.55(b), ductility has a pronounced influence
represents the evolution of steplap joint design on mechanical response of bonded joints, and
over many years. Early analytical work was design only for elastic response deprives the

Table 28.1 Summary of step lap joint results (Figs. 28.52 and 28.53)

No. of steps 10 10 10 5 5
11.1 1 2 3 3 6
Joint length, em 4.44 8.89 13.33 6.05 12.5
(in) (1.75) (3.5) (5.25) (2.47) (4.93)
Allowed resultant, kN / em 12.03 18.78 22.05 12.35 13.43
(103 1b/in) (6.87) (10.72) (12.59) (7.05) (7.67)
652 Mechanical fastening and adhesive bonding

0.40 HTS GRAPHITE EPOXY TITANIUM

<D
\-0.&0-1 J.-O.15--l
•
(2P1.CS, (3PLCS'
STATIONNUMBERS t z 3 &
SPECIMEN GEOETRY tALL DIMENSIONS IN INCHES,
TENSI.E LOMJINQ
(AT- -4IO'f'
7 SI1IENIJIH = I!,7G t.8IIN.

6 0.15 -1.+7..

Ii

SHEAR 4 0.10
STRESS
(KSI) 3
SHEAR
-'Yo
STRAIN
2 0.05

1
o~~--~~--~ __ ~~
1 2 3 4 Ii 6 2 3 4 Ii 6
STATION NUMBER STATION NUMIISt

Fig. 28.54 Practical step lap joint designBo.

60 LT

60 [(A) FM73 NON POROUS I

Of
HT/WT

o 0.2 0.4 0.8 0.8 1.2 1.4

Shear Strain

50 LT = -55degC
=
Lf
45 RT Room Temp
~40
HTIWT
=
HTIWT 60 degC/100%RH
,.36
";;30
=26
~211
~ 15
.<:
'" 10

--
6
0
0 0.02 0.G4 0.08 0.08 0.1 0.12 0.14

Fig. 28.55 Typical stress-strain characteristics of aerospace adhesivesB1 ,

 Adhesive joints 653

application of a significant amount of avail- T ,which can be obtained by solving for T

p p
able structural capability. from eqn (28.32), leading to
The work of Hart-Smith34-39 emphasized the
importance of ductile adhesive response and Tp = GbOYmax - -V[(Gboymax)2 - 2GbOSE]
introduced the relationship between the strain
energy to failure of the adhesive and the load (28.33)
capacity of the joint. As a means of simplifying Hart-Smith has also used an equivalent
the stress analysis of the joint in the presence bilinear representation in which the horizontal
of ductile adhesive response, Hart-Smith part of the curve is set equal to Tmax, the maxi-
showed that any bilinear stress-strain curve mum shear stress of the actual stress-strain
which has the same ultimate shear strain and curve, and the initial modulus GbO adjusted to
maximum strain energy as that of the actual give the strain energy match, using the expres-
stress-strain curve will produce the same total sion:
load in the joint. Figure 28.5681 gives an exam-
ple of the method for fitting a bilinear curve to GbO = Tma/(TmaxYmax - SE)/2 (28.34)
the actual stress-strain curve of the adhesive which is also obtained from eqn (28.32) when
in shear. With the strain energy of the adhesive r max is substituted for r p. In either case the use
given by of a bilinear representation of the stress-strain
(28.32) curve for the response of the adhesive in shear
makes it straightforward to obtain one dimen-
where GbO' Ymax and SE are the initial modulus sional stress distributions in various types of
of the stress-strain curve, the maximum strain joint geometry with adhesive ductility
and the strain energy of the adhesive at Ymax accounted for; solutions are given for single
respectively, then the equivalent bilinear curve and double lap joints with uniform and
consists of an initial straight line of slope GbO tapered adherends, as well as more sophisti-
together with a horizontal part at an abscissa cated joint designs such as scarf and step lap

-
6 I I I
E1utic-Pl&ftic Mod.l 11 •... 0 a

4P •
L/:}!l

~ ~1)50
y. 70S PSJ
Of/IN
. ~
t-N... .

FaU1l .. e 8t ...... -
Ct
'"
A.... n.. -
Lower Bou.· 41,., PC
VII,." . . . . - 5590 P&l
!SSG PSI
(
B.d Uae
1'-1
Nap1tiA Alai Sh.ar '1' ••'e ~ TIlle........
FM-1174~ 0 0.00651
1'.111&" ay- e 0.004H
A_ ~0.0G4~_
'""Low... It"""--0.492
A.... n..
IN/Df -
• 0.00471

"i
O. 70S Dl/rN
Upper BOIIAd - 0,99TIN/JIII I- SM
o
o 0.1 0.203 O. SO; 0.421 0.'46 0.614 1.1.0 1.0)0
"':r Stl'ata DC/Dt
Fig. 28.56 Elastic-perfectly plastic adhesive response model (FM73)82.
654 Mechanical fastening and adhesive bonding

geometries (Hart-Smith~4--39). These have sub- double lap joint with equal adherend stiff-
sequently been incorporated in the 'A4Ex' nesses (i.e. Eo :;: Ei; to :;: [/2). Figure 28.57(a)
series of computer programs44-46 mentioned gives the distribution of upper adherend axial
previously. stress resultant while Fig. 28.57(b) gives the
Figure 28.57 (see notation of Fig. 28.36) shear stress distribution in the bond layer. The
shows an example of the use of the bilinear linear portions at the ends of the resultant dis-
stress-strain curve approximation, in this case tribution in Fig. 28.57(a) correspond to the ends
for predicting the stresses in a symmetrical of the shear stress distribution in Fig. 28.57(b)

4.5 (A) Upper adherend resultant distribution

II:
4 ( 1 k-Ib/ln'" 1.752 kNlcm)
:i

,i!
J; 3.5
3
2.6
'V
c:: 2
!
.: 1.6
1.. 1
IAxial load 310 MPa (46 kSi)1
!.
g 0.5
0
0 0.6 1 1.& 2
)(

fp~ ooE f.
5
4.5 (8) Shear stress distribution
4 (1 ksi= 6.896 MPa)
'Uj
~ 3.&
u; 3
en
-=
en 2.5 IAxial load 310 MPa (45 kSi)1
~ 2
.!
til
1.$
1
0.5
0
0 0.& 1 1.5 2

Fig. 28.57 Stress distributions in double lap joint - ductile adhesive response.
 Adhesive joints 655

where the shear stress is a constant because of then eqns (28.36a, b) give the same value. The
the plateau in the bilinear representation of the factor (2GboYmax/ip _1)112 in eqn (28.36b) acts as
stress-strain curve, in agreement with the equi- a load enhancement factor and represents the
librium relation given in eqn (28.6). Following increase of joint load capacity due to ductile
the analysis developed by Hart-Smith, the adhesive response over the maximum load
lengths of the plastic zones designated in Fig. allowed by elastic response of the adhesive.
28.57(b) as Ip are given by Note that eqn (28.36b) can be rearranged to
express aJmax in terms of the maximum strain
Ip =(aj2ip -l/{3bd)to ;{3bd = [2GbOt/Eo~]l/2 energy of the adhesive:
(28.35) SE = i//2GbO + ip(Ymax -Ye)
Here {3bd (subscript 'bd' denoting balanced
where Ye = 7:/GbO (28.37)
double lap) is equivalent to {3 given in eqn
(28.14) when the latter is specified for the case then eqn (28.36b) can be written:
of equal-stiffness adherends, while ax is the 2
nominal loading stress at either end of the aJmax = p-V(2GbOSE) (28.38)
overlap. The expression for Ip given in eqn
(28.33) is valid only if greater ilian 0, of course, The Hart-Smith analysis based on the equiva-
negative values of plastic zone length not hav- lent bilinear stress-strain law was shown in
ing any meaning. Thus if {3 aj 2 < i ,no plastic Ref. 35 to give the same joint load capacity as
zone is present and the behavior ~f the joint the solution for the problem using the actual
can be considered to be purely elastic. The stress-strain curve of the adhesive. The conve-
maximum value of ax for this case can be nience of the bilinear stress-strain description
expressed by inverting the shear stress expres- is in the simplicity of the solutions it allows;
sion in eqn (28.17) with 0th = 0, for the case of once the length of the plastic zone at each end
equal adherend stiffnesses and setting 7:b)max to is determined, the same types of solution
i p. For the case of {3aj2 ~ ip which corre- apply for the elastic zone as were given in eqns
sponds to ductile response of the adhesive (the (28.9) and (28.10) for the resultant and shear
plateau of the bilinear stress-strain curve), the stress distributions, together with linear resul-
Hart-Smith analysis35 provides the required tant and constant shear stress distributions in
expression for Ux The two cases are summa- the plastic zones.
rized as follows: The most obvious effect of ductility in the
adhesive behavior is the reduction of peak
{3bdaj2 < ip (elastic response):
shear stresses. In addition, there is a beneficial
effect on reduction of peel stresses. For the
(28.36a)
double lap joint considered in Fig. 28.57, the
maximum peel stresses denoted by 0b)max
{3bdaj2 ~ ip (ductile response):

Ii:
which occur at the ends of the joint, are given39
by:
uJ~ = t,~(2 G,,~- -1) (28.39)

~ a.~(2 G~_ -1) (28.36b)

_ ( ~to )1/4 .
y- 3 Et '
o b
If Ymax = i / Gw which is the maximum strain
in the elasHc part of the bilinear representation, Eb = peel modulus of adhesive
656 Mechanical fastening and adhesive bonding

where 1:b)max is the maximum shear stress, leads to high shear and peel stresses at the
either pa /2 for the elastic case or 1: for the ends of the joint, and may inhibit desirable
case of d~ctile response. The maxitlum peel flow characteristics of the adhesive. On the
stresses are thus reduced by the same ratio as other hand, thick bond lines tend to generate
the maximum shear stresses in the case of duc- porosity which weakens the bond. Data pre-
tile response of the adhesive. sented in Ref. 83 show a fairly persistent
Even though ductile response of the adhe- tendency for lap shear strength to drop off
sive provides additional load capacity of the somewhat as the bond thickness is increased
joint over what is provided by purely elastic above 0.12 mm (0.005 in).
response, it is advisable to keep the load In addition to effects of bond thickness per
capacity of the joint low enough to ensure se, Hart-Smith60 discusses the effect of bond
purely elastic response for most practical situ- thinning at the ends of the joint which is
ations where time-varying loading is caused by resin flow during curing. Figure
encountered. Some damage to the adhesive 28.58 illustrates the tendency toward bond
probably occurs in the ductile regime which thinning at joint edges together with some
would degrade the long-term response. The manufacturing techniques for avoiding the sit-
main benefit of ductile behavior is to provide uation. Loss of bond thickness may cause
increased capacity for peak loads and damage considerable elevation of shear and peel
tolerance with regard to flaws - voids, poros- stresses in the bond. In addition to the
ity and the like - in the adhesive layer. In approaches shown in Fig. 28.58, tapering of
addition, calculations of the plastic zone the adherends near the ends will help to alle-
length play a part in the avoidance of creep viate the situation; tapering from the inside
failures which can constitute a major consider- surface of the adherend will also provide a
ation for slow cyclic loading in hot wet local thickening of the bond line to compen-
environments. sate for thinning due to resin loss.
Effects of porosity in the bond layer are
illustrated in Fig. 28.59 61 which compares the
response of FM73 for porous and non porous
28.4.4.2 Effects of bond layer defects
bond layers for various environmental condi-
Defects in adhesive joints include surface tions. The data 61 indicate that porosity is
preparation deficiencies leading to low mainly a characteristic of thickness of the bond
strength interfaces between the adhesive and layer. There is some loss of structural capabil-
adherends, voids and porosity, and lack of ity in the presence of porosity in the bond, but
bond thickness control. Surface preparation there may still be adequate strength for the
effects were discussed in Section 28.4.2.6 and bond to function as required if the joint is
will be treated in considerable detail in designed adequately. Since porosity is associ-
Chapter 29. However, it should be kept in ated with thickened regions of the bond which
mind that adhesive joints will not succeed in tend to occur away from the edges, porosity
providing dependable performance if good tends to be confined to the interior of the joint
surface preparation procedures are not main- where the stresses are relatively small, and
tained. may not be objectionable in many cases. The
Bond layer thicknesses of 0.12-0.25 mm main focus in Ref. 61 is the effect of adherend
(0.005-0.010 in) are typical of structural thickness with regard to damage tolerance in
bonded joints. There appears to be a tradeoff the presence of bond layer defects. The issue
between negative effects which occur when has to do with the design principle discussed
the bond is too thin and those occurring for in Section 28.4.2.1 of keeping adherend thick-
too thick bond layers. Thinness in bond layers nesses within limits which ensure that the
 Adhesive joints 657

VAeIl.IM;
..
f
rf. SMWUMmD
.
..e

~
AOtIIIIft
t
6
SC.II A.",VI!
1TIIIS_IIIlL lI0II1 HLY WlTII.-LIITIIAIIIrT IHIII
A. PlNCH-OFf UftIOER VACUUM BAG
8. USE OF WlRi TO LI FT VACUUM BAG OFF TIP OF BOND

'>
TlllCltEII".lI TI\AII
trUCE tI1I "UlLEII ?
+Jl l'
VACUlIIIIAII

::;Sr
TAftIl ""'ClllAI. " ......IY "_I-lACICID
TAft ,., LOCATE
fllGE flA"", WIlleN CAl •
tnTOUIlO . . . .YCII " " ' "
C. REITRICTEO AREA Of PREllURI! ""'~ATION ..MI•• AlIltCtllAIY
IN OUT.oF·AUTOCLAVIBONOING I).IJI! OF THICKER EDGE PUTE TO RILl IN!
VACUUM flAG PRESSURE AT !!DOE
E. ADOtnON 0' 1 TO. PERCENT 0" O.ClO7·tNCH GLASS B£ADa K\I ADH&SJV£ FILM
IS A VERV EFFECTIVl T~IOUE FOR CONTROLLING BONOliNE THICI(NUS

Fig. 28.58 Manufacturing techniques to relieve bond pinch-off 51.

60

SO
T~lT Es VS. Strain .. FM7~
~ »>XX ><x LT
~ 40 ;?~RT H/W/Non.porous

""""~ 30
::;r- .. " RT ~~ Porous
U; ~ »x·xxx·x-x ~ -- / H/W LT - - -55degC
)'~ RT - -room temp.
iii
Q)
~
(f)
20
*-r·
~' )<xx-xx-x-x~x.-x-xx--x--
x __ x-><-x--x--x~
x-xx
B/W - - 60 degC/IOO% rh

10 x~:
~>.

0 f
o O.S 1.5 2
Shear Strain

Fig. 28.59 Effect of porosity on adhesive stress-strain characteristics81 x: porous; : non porous bond layers.

adherends fail rather than the bond. Ductile to)cr,which ensures adherend failure can be
response of the adhesive has an important expressed by restating eqn (8) of Ref. 52 as fol-
influence on the situation. By making use of lows:
eqn (28.36b) above, together with the defini-
8Eoip(Ymax - y/2)
tion of flbd given in eqn (28.35), the adherend (28.40)
thickness limit for a double lap joint, a)2ult
658 Mechanical fastening and adhesive bonding

which is equivalent to: damage, while tests conducted at one cycle per
hour produced failures within a few hundred
t),,/tb = 4 {Max adhesive strain energy}/
cycles. On the other hand, specimens repre-
{Max adherend (elastic) strain energy} (28.41)
sentative of structural joints which have the
Hart-Smith states that for to not greater than characteristic shear stress trough seen in Figs.
t)cr given in eqn (28.40), joint performance will 28.37 and 28.57 are able to sustain hot wet con-
be relatively insensitive to bond flaws pro- ditions even at low cycling rates if the length
vided there is adequate ductility in the of the elastic region (Ie in Fig. 28.57) is long
adhesive response. Problems may arise with enough. Based on experience of the PABST
high temperature adhesives such as the PM program, the Hart-Smith criterion for avoid-
400 considered in Fig. 28.55(b), since these ance of creep failure requires that tJmin be no
tend to have limited ductility. As indicated in greater than'l/10. But the stress analysis for
eqn (28.36b), limited ultimate strain capability the elastic-plastic case using the bilinear adhe-
(ymax) will reduce margin of ultimate strength sive response model leads to an expression for
over elastic response of the joint. the minimum shear stress equivalent to eqn
(28.12) with 1 replaced by Ie:
28.4.4.3 Durability of bonded joints T'
T'b)min = sinh PPI /2t (28.42)
Two major considerations in the joint design bde 0

philosophy of Hart-Smith are: (1) either limit- (f1bd - see eqn (28.35». Since sinh (3) "" 10, this
ing the adherend thickness or making use of amounts to a requirement that Pbd1/2to be at
more sophisticated joint configurations, such least 3, i.e. that the elastic zone length be
as scarf and step lap joints, to ensure that determined by Ie ~ 6t/Pbd• Since Ie is equiva-
adherend failure takes precedence over bond lent to the total overlap length, I, minus the
failure; (2) designing to minimize peel stresses, sum of the plastic zone lengths, i.e. 21 , then
either by keeping the adherends excessively making use of the expression for I fn eqn
thin or, for intermediate adherend thicknesses, (28.33), the criterion for elastic zorie length
by tapering the adherend. In addition, it is reduces to a criterion for total overlap length
essential that good surface treatment practices corresponding to a lower bound on I which
be maintained to ensure that the bond can be stated as
between the adhesive and adherends does not
fail. When these conditions are met, reliable
performance of the joint can be expected for
I ~ (~; + P~d )to (28.43)
the most part, except for environmental Equation 28.43 for the joint overlap length is
extremes, i.e. hot wet conditions. The Hart- the heart of the Hart-Smith approach to dura-
Smith approach focuses primarily on creep bility of bonded joints for cases where
failure associated with slow cyclic loading (i.e. adherend failure is ensured over bond failure
one cycle in several minutes to an hour) under for static loading and in which peel stresses are
hot wet conditions, this corresponds, for eliminated from the joint design. This type of
example, to cyclic pressurization of aircraft requirement has been used in several contexts.
fuselages. In the PABST program41-43, 18 test For example60, it becomes part of the require-
specimens used for characterizing adhesives ment for acceptable void volume in the bond
(so-called 'thick adherend' specimens) which layer, since in this case the voids, acting essen-
are designed to produce essentially uniform tially as gaps in the bond layer, reduce the
shear stress along the bond were tested at high effective length of the overlap. The criterion
cycling rates (30 Hz) and were able to sustain has to be modified numerically for joints other
more than 10 million loading cycles without than symmetric double lap joints with equal
 Adhesive joints 659

stiffness adherends and uniform thickness. For suggest the need for consideration of crack
more sophisticated joint configurations such as growth phenomena in bonded composite
step lap joints, the A4EI computer code pro- joints. Indeed, a major part of the technical
vides for a step length requirement equivalent effort that has been conducted on the subject
to that of eqn (28.43) for simple double lap of durability of adhesive jointS65-72 has been
joints. based on the application of fracture mechanics
In addition to creep failures under hot-wet based concepts. The issue of whether or not a
conditions, the joint may fail due to cracking fracture mechanics approach is valid needs
in the bond layer. Johnson and Mall72 pre- further examination. Apparently, no crack-like
sented the data in Fig. 28.60 which shows the failures occurred in the PABST program,
effect of adherend taper angle on development which was a metal bonding program, even
of cracks at ends of test specimens consisting when brittle adhesives were examined at low
of composite plates with bonded composite temperatures. The amount of effort which has
doublers, at 1()6 cycles of fatigue loading; here been expended by a number of respected
the open symbols represent the highest load workers on development of energy release
levels at which cracks fail to appear while the rate calculations for bonded joints certainly
solid symbols are for slightly higher loads at suggests that there is some justification for
which cracks just begin to appear. It is noted that approach, and the results obtained by
that even for outer adherend taper angles as Johnson and Mall appear to substantiate their
low as 10° (left-most experimental points in need for composite joints in particular.
Fig. 28.60) for which peel stresses are essen-
tially nonexistent for static loading, crack
28.4.5 MECHANICAL BEHAVIOR OF
initiation was observed when the alternating
COMPOSITE ADHERENDS
load was raised to a sufficient level. A number
of factors need to be clarified before the impli-
cations of these results are clear. In particular it
28.4.5.1 Joint failure characteristics
is of interest to establish the occurrence of
bond cracking at shorter cycling times, say less Typical failure modes in structural joints are
than 3 x lOS cycles corresponding to expected illustrated in Fig. 28.61 which are indicative of
lifetimes of aircraft. Effects of cycling rate and adherend rather than bond failures. In the case
environmental exposure are also of interest. of single lap joints (Fig. 28.61(a» bending fail-
Nevertheless, the data presented in Ref. 65 ures of the adherends will occur because of
high moments at the ends of the overlap. For
metal adherends, bending failures take the
200 I form of plastic bending and hinge formation,
a
while for composite adherends the bending
150
\1
[J.....
•
I L FM-300 2
Gth = 81 J/ar
• DEBONO
[JNO DEBONO
failures are brittle in nature. In the case of dou-
~~~~:~D I.~ -~~ -;;;:;.:z~
STRESS,
100
ble lap joints, peel stresses build up for thicker
adherends causing the types of interlaminar
1>5. Mfa £C-3445 - -
failures in the adherends illustrated in Fig.
SO G1h = 18 J/1\12 28.61(b).
a
L!. -I "1 CLS SPECIMEN 1-

lO
TAPER ANGLE,
60
a. !leg
90 28.4.5.2 Thermal stress effects
Thermal stresses are a concern in joints with
Fig. 28.60 Crack development in bonds of tapered adherends having dissimilar thermal expan-
composite doublers at 106 loading cycles72• sion coefficients. Figure 28.62 illustrates the
660 Mechanical fastening and adhesive bonding

- '.....- - - - - '

-t=( ==:f!:~~~
IIGH LON) LEVEL (METAL MJHER!NDB)
~
PERMHN'lLY OEFOIIIIED (METAL) AtIHlMNDS
AFTER FALURI t1F ADHESIVE
(A) Single Lap Joints

-~
"~·-TITI:
I
-5 L __ ~
!ITRESSES AC'IING ON OUTER AIlHERENO
••
-J L ;J~
_c.._._~_._.l
c.

A.B. AND C INDICATE FAILURE _QUENCE

(8) Double Lap Joints

Fig. 28.61 Failure modes in composite adherends49,50.

8)T
I

ED
StruI:IUaIi load
plus
~ "mal stress E; = Eo 68.97 GPa (10000 ksi); Gb
:2 4J = 1.04 GPa (150 ksi); t/2 = to =
III
1/1
8tIuI:tUnI1laad ~ =
2.54 mm (0.1 in); tb 0.254 mm
oily
i 2) (0.01 in); a j = 23.4 x 1~ °C-1(13 x
1~ °F-1); a o = 1.8 x 10-6°C-1 (1 X

I
.t:. 0 22RqqqUUauua~ooa666tK&~
1~OF-1).

Cure temperature 121.1 °C

tfJ 30 40 50
(250°F);
.2) x,(~~ Application temperature 23.9°C
(75°F)
Loading stress 68.97 MPa
(10 ksi).

Fig. 28.62 Thermal shear stresses in double lap joints: outer adherend 0/90 carbon epoxy; inner adherend
aluminum.
 Adhesive joints 661

effect of thermal stresses in a double lap joint loading stress as 69 MPa (10 ksi). Similar
consisting of an aluminum inner adherend effects occur with the peel stresses, although
and a 0/90 0 carbon epoxy outer adherend. The the peel stresses due to thermal mismatch
stresses due to thermal mismatch between the alone have the same sign at both ends of the
aluminum and composite arise if the cure tem- joint; with a composite outer adherend the
perature of the bond is substantially different thermally induced peel stresses are negative,
from the temperature at which the joint is which is beneficial to joint performance.
used. The case considered here represents a Peak peel and shear stresses obtained from
121 DC (250°F) cure temperature for the adhe- these relationships for various combinations
sive and a room temperature application, a of metal and composite adherends whose
temperature difference of -79°C (-175°F), properties are given in Tables 28.2 and 28.3 are
which (see Tables 28.2 and 28.3) would result in shown in Table 28.4. For joints with an alu-
a strain difference of 0.002 between the alu- minum inner adherend, the difference in
minum and composite if no bond were present. thermal expansion between the adherends is
(The material combination considered here, relatively large, giving considerably higher
aluminum and carbon epoxy, represents the thermal stresses for the most part. In addition,
greatest extreme in terms of thermal mismatch carbon epoxy has a particularly low thermal
between materials normally encountered in expansion, which tends to produce higher
joints in composite structures.) Thermal thermal stresses with carbon epoxy adherends
stresses in bond layers of double lap joints can in combination with metals than do other
be determined from the expressions given in composites. Note that boron epoxy in combi-
eqns (28.14-28.20). (These calculations are all nation with titanium gives particularly small
based on an assumed elastic response of the thermal stresses because of similarity of the
adhesive.) thermal expansion coefficients shown in
Hart-Smith34-39 provides corrections for Tables 28.2 and 28.3 for these materials. As
ductile response in the presence of thermal
Table 28.3 Generic metal properties (MIL-HDBK-5
effects. Figure 28.62 illustrates how the ther- 1983)
mal stresses combine with the stresses due to
structural load to determine the actual stress Ti6-AI4-4V 1025 2014
distribution in the adhesive. The thermal Steel Aluminum
stresses in themselves develop an appreciable Young's 110.3 69.0
206.9
fraction of the ultimate stress in the adhesive, modulus,
and although they oppose the stresses due to GPa
structural loading at the left end, they add at
Poisson 0.3 0.3 0.3
the right end and give a total shear stress that
ratio
is somewhat beyond the yield stress of typical
adhesives, even with as small a structural a, 1~ °C-l 8.82 10.26 23.4

Table 28.2 Generic mechanical properties of composites (C.c. Chamis NASA Lewis Research Center,
NASA TM-86909, 1985)

Unidirectional lamina 0/90 Laminate

Composite Eu GPa ET, GPa V LT au 1~ 0C-1 aT' 10-6 °C-1 ExGPa a x 1~ 0C-1
Boron epoxy 201 20.1 0.17 11.7 30.4 113.8 8.6
5-glass epoxy 60.7 24.8 0.23 3.78 16.7 43.72 7.92
Carbon epoxy 137.9 6.90 0.25 0.72 29.5 72.6 2.34
662 Mechanical fastening and adhesive bonding

Table 28.4 Bond layer thermal stress in double lap joints (0/90 composite outer adhered, metal inner
adherend)

Boron epoxy Glass epoxy Carbon epoxy

Titanium
Shear stress, MPa 0.419 2.33 15.64
(ksi) 0.061 0.338 2.27
Peel stress, MPa -0.465 -3.73 -19.43
(ksi) -0.067 -0.541 -2.817
Steel
Shear stress, MPa 5.44 7.99 26.22
(ksi) 0.789 1.16 3.80
Peel stress, MPa -6.30 -15.0 -38.1
(ksi) -0.914 -2.17 -5.52
Aluminium
Shear stress, MPa 27.7 28.2 40.47
(ksi) 4.02 4.08 5.86
Peel stress, MPa -24.4 -40.1 -44.6
(ksi) -3.54 -5.82 -6.47
ti = S.08 mm (0.2 in); to adjusted for equal adherend stiffnesses; tb = 0.253 mm (0.1 in). Adhesive properties: shear mod-
ulus, 1.03 GPa (ISO ksi); peel modulus, 3.49 GPa (SOO ksi). Give temperature, 121°C (250°F). Application temperature,
24°C (7S°F).

discussed earlier, the 'peel' stresses shown in tion for metal adherends which are relatively
Table 28.4 are all negative (i.e. compressive) stiff with respect to transverse shear deforma-
because of the location of the composite on the tion, but for polymer matrix composite
outside of the joint, although the shear stresses adherends which have low transverse shear
are unaffected by this aspect of the joint. moduli, transverse shear deformations are
Composite repair patches on aluminum air- more significant and can have an important
craft structures benefit from this type of influence on bond layer shear stresses. A use-
behavior, in that peel stresses are not a prob- ful correction to the classical Volkersen
lem for temperatures below the cure solution which allows for transverse shear
temperature. Placing the metal rather than the deformations in the adherends can be
composite on the outside of a double lap joint obtained by modifying the shear modulus of
would reverse the signs of the peel stresses the adhesive from its actual value, Gb, to an
making them tensile and aggravating the effective value, Gb)eff' given by
effects of differential thermal expansion of the
adherends. (28.44)

28.4.5.3 Transverse shear and stacking 1 ( Gb to

where Ksh = 1 + - G t + ~2t
Gb ti)
sequence effects in composite adherends 3 xzob XZl b

Classical analyses such as the Volkersen shear Here Gxzo and Gxzi are the transverse shear
lag model for shear stresses in the bond layer moduli of the adherends. For the double lap
(Sections 28.4.3.3 and 28.4.3.4.1) are based on joint, the parameter Pappearing in eqn (28.14)
the assumption that the only significant defor- (see the third equation of the top row of eqn
mations in the adherends are axial, and that (28.14» is then modified by replacing Gb by
they are uniformly distributed through the G/ Ksh using the value for Ksh given in eqn
adherend thicknesses. This is a good assump- (28.44), and all the expressions in eqns
 References 663

(28.15-20) for stresses in the bond layer are 3. Oplinger, DW. and Gandhi, KR, Analytical
modified by the resulting alteration of {J. The studies of structural performance in mechani-
correction given here amounts to treating 1/3 cally fastened fiber-reinforced plates. In Froc.
Army Solid Mechanics Con! 1974, Army
the thickness of each adherend as an extension Materials and Mechanics Research Center
of the bond layer, and assigning the shear stiff- Manuscript Report AMMRC MS 74-8 (1974).
ness of the adherend for that part of the 4. Garbo, S.P., Ogonowski, J.M. and Reiling, H.E.,
effective bond layer. The factor 1/3 corre- Jr, Effect of variances and manufacturing tolerances
sponds to a linear distribution of shear stress on the design strength and life of mechanically fas-
through the adherend thicknesses, which is tened composite joints. v2 Air Force Wright
consistent with the assumption that the axial Aeronautical Laboratories Report AFWAL-TR-
81-3041 (1981).
deformations are approximately uniform
5. Hyer, M.W. and Klang, E.C., Contact stresses in
through the adherend thickness. pin-loaded orthotropic plates, Virginia Tech Center
As an example, consider joint with a 0/90 for Composite Materials and Structures Report
carbon epoxy outer adherend joined to an alu- CCMS-84-02 (1984).
minum inner adherend, with adherend 6. Ramkumar, RL., Saether, E.S. and Appa, K,
thicknesses of 2.53 mm (0.1 in) and 5.06 mm Strength analysis of laminated and metallic plates
(0.2 in), respectively, and a 0.253 mm (0.01 in) bolted together by many fasteners, Air Force Flight
Dynamics Laboratory Report AFWAL-TR-86-
bond thickness. Assume a shear modulus of 3034 (1986).
the bond layer of 1.06 CPa (150 ksi) and trans- 7. Madenci, E. and meri, L., Analytical determina-
verse shear moduli of 4.82 CPa (700 ksi) for the tion of contact stresses in mechanically fastened
composite adherend and 26.5 CPa (3800 ksi) composite laminates with finite boundaries.
for the aluminum. A value of 1.839 is then Intern. ]. Solids Structures 30, pp. 2469-2484
obtained for Ksh' and the value of {J and the (1993).
maximum shear and peel stresses which 8. Crews, rH. and Naik, RA., Combined bearing
and bypass loading on a graphite/epoxy lami-
depend on it are reduced by a factor of (Ksh)1/2
nate. Composite Structures, 6, pp. 21-48 (1986).
or 1.36 for this case. The shear and peel 9. Hart-Smith, L.J., Mechanically-fastened joints
stresses are therefore approximately 30% for advanced composites - phenomenological
lower than the values predicted with the considerations and simple analyses. In Fibrous
unmodified bond shear modulus. This type of Composites in Structural Design. New York:
correction can be shown to give relatively Plenum Press (1980) pp. 543-574.
good predictions of the adhesive stresses in 10. Petersen, RL., Stress Concentration Factors. New
York: John WIley and Sons (1974) p. 135.
comparison with finite element analyses. In 11. Lenoe, E., Oplinger, D.W. and Burke, J.J.,
addition, the departure of Ksh given in eqn Fibrous Composites in Structural Design. In
(28.44) from 1 gives a good indication of the Froc. 4th Conf. Fibrous Composites in Structural
range of joint parameters for which adherend Design, New York: Plenum Press (1980).
shear deformations are important. 12. Oplinger, D.W., On the Structural Behavior of
Mechanically Fastened Joints in Composite
Structures. In Fibrous Composites in Structural
REFERENCES Design, New York: Plenum Press (1980) pp.
575-602.
1. Waszczak, J.P. and Cruse, T.A., A synthesis proce-
13. Nelson, W.D., Bunin, B.L. and Hart-Smith, L.J.,
dure of mechanically fastened joints in advanced
Critical Joints in Large Composite Aircraft
composite materials, Air Force Materials
Structure. In Proc. 6th Con! Fibrous Composites in
Laboratory Report AFML-TR-73-145 (1973).
Structural Design, Army Materials and
2. Oplinger, D.W. and Gandhi, KR, Stresses in
Mechanics Research Center Manuscript Report
mechanically fastened orthotropic laminates. In
AMMRC MS 83-2 (1983), pp. U-2 to 11-38.
Proc. 1st Con! Fibrous Composites in Flight Vehicle
14. Whitney, J.M. and Nuismer, R, Stress fracture
Design, Air Force Flight Dynamics Laboratory
criteria for laminated composites containing
Report AFFDL-TR-74-103 (1974).
664 Mechanical fastening and adhesive bonding

stress concentrations. J. Composite Materials, 8, (1944).

pp. 253-265d (1974). 28. Kutscha, D. and Hofer, K., Feasibility of joining
15. Garbo, S.P. and Ogonowski, J.M., Effect of vari- advanced composite flight vehicle structures, Air
ances and manufacturing tolerances on the design Force Materials Laboratory Report AFML-TR-
strength and life of mechanically fastened composite 68-391 (1968).
joints, Air Force Flight Dynamics Laboratory 29. Dickson, J.N., Hsu, T.N. and McSkinney, J.N.,
Report AFFDL-TR-78-179 (1978). Development of an understanding of the fatigue phe-
16. Chang, EK. and Hung, Chang-Li, Response of nomena of bonded and bolted joints in advanced
double shear-lap bolted composite joints under mul- filamentary composite materials. Vol. 1, Analysis
tiple bypass loads. Standford University Dept. of Methods, Lockheed Georgia Aircraft Company,
Aeronautics and Astronautics Report, Doctoral USAF Contract Report AFFDL-TR 72-64, vol I
dissertation of Chang-Li Hung. (1993). Uune 1972).
17. Garbo, S.P., Effects of bearing/bypass load interac- 30. Grimes, G.c., Wah, T. et al., The development of
tion on laminate strength, Air Force Flight non-linear analysis methods for bonded joints in
Dynamics Laboratory Report AFWAL-TR-81- advanced filamentary composite structures, South-
3114 (1981). West Research Institute, USAF Contract Report
18. Naik, R.A. and Crews, J.H., Jr., Ply-level failure AFML-TR-72-97 (September 1972).
analysis of a graphite/epoxy laminate under bear- 31. Renton, W.J., The analysis and design of composite
ing/bypass loading, NASA Technical materials bonded joints under static and fatigue
Memorandum 100578 (1988). loadings, PhD Thesis, University of Delaware
19. Hart-Smith, 1.J., Design methodology for bonded- (1973).
bolted composite joints. vI Air Force Wright 32. Renton, W.J. and Vmson, J.R., The analysis and
Aeronautical Laboratories Report AFWAL-TR- design of composite materials bonded joints under
81-3154 (1982). static and fatigue loadings, Air Force Office of
20. Ramkumar, R.1., Saether, E. and Cheng, D., Scientific Research Report TR-73-1627 (1973).
Design guide for bolted joint in composite struc- 33. Oplinger, D.W,. Stress analysis of composite
tures, Air Force Wright Aeronautical joints. In Proc. 4th Army Materials Technology
Laboratories Report AFWAL-TR-88-3035 (1986). Con!, Newton, MA: Brook Hill Publishing Co.
21. Shivakumar, K.N. and Crews, J.H., Jr., Bolt (1975), pp. 405-451.
clamp-up relaxation in a graphite/epoxy laminate, 34. Hart-Smith, 1.J., AFFDL-TR-72-130, pp.
NASA Technical Memorandum 83268 (1982). 813--856.
22. Slepetz, J.M, Oplinger, D.W. and Andrews, B. 35. Hart-Smith, L.J, Adhesive bonded double lap joints,
0., Bolt tension relaxation in composite friction NASA Langley Contractor Report, NASA CR-
joints. In Proc. 7th Con! Fibrous Composites in 112235 (1973).
Structural Design, Wright Patterson 36. Hart-Smith, 1.J., Adhesive bonded single lap joints,
Aeronautical Laboratories Report AFWAL-TR- NASA Langley Contractor Report, NASA CR-
85-3094 (1985). 112236 (1973).
23. Harris, H.G., Ojalvo, I. and Hooson, R.E., Stress 37. Hart-Smith, 1.J., Adhesive bonded scarf and
ai.d deflection analysis of mechanically fastened stepped-lap joints, NASA Langley Contractor
joints, Air Force Flight Dynamics Laboratory Report, NASA CR-112237 (1973).
Report AFFDL-TR-70-49 (1970). 38. Hart-Smith, 1.J., Analysis and design of advanced
24. Shyprykevich, P., Characterization of bolted joints composite bonded joints, NASA Langley
behavior, MIL-HDBK-17 Accomplishments at Contractor Report, NASA CR-2218 (1973).
Standardization ASTM J. Composite Technical 39. Hart-Smith, 1.J., Advances in the analysis and
Research, 17(3), pp. 260-270 (1995). design of adhesive-bonded joints in composite
25. Military Handbook MIL-HDBK-17: Polymer aerospace structures. In SAMPE Process
Matrix Composites (1992). Engineering Series, 19, pp. 722-737 Asuza:
26. Volkersen, 0., Die Nietkraftverteilung in SAMPE, (1974).
Zugbeanspruchten Nietverbindungen mit 40. Primary adhesively bonded structure (PABST) tech-
Konstanten Laschenquerschnitten. nology, Air Force Contract F33615-75-C-3016
Luftfahnforschung, 15, pp. 4-47 (1938). (1975).
27. Goland, M. and Reissner, E., Stresses in 41. Thrall, E.w., Primary adhesively bonded structure
cemented joints. J. Appl. Mech., 11, A17-27 technology (PABST) Phase Ib: Preliminary Design,
 References 665

Air Force Flight Dynamics Laboratory Report Critical joints in large composite aircraft struc-
AFFDL-1R-76-141 (1976). ture. In Proc. 6th Conf Fibrous Composites in
42. Shannon, R. W et al., Primary adhesively bonded Structural Design, Army Materials and
structure technology (PABST): General Material Mechanics Research Center Manuscript Report
Property Data, Air Force Flight Dynamics AMMRC MS 83-8 (1983).
Laboratory Report AFFDL-TR-77-101 (1977). 53. Oplinger, D.W., A layered beam theory for single
43. Land, K.L., Lennert, F.B. et al., Primary adhesively lap joints, US Army Materials Technology
bonded structure technology (PABST): Tooling, Laboratory Report MTL TR 91-23 (1991).
Fabrication and Quality Assurance Report, USAF 54. Oplinger, D.W, Effects of adherend deflections
Technical Report AFFDL-TR-79.3154 (October, on single lap joints. Int. J. Solids Structures, 31
1979). (18), pp. 2565-2587 (1994).
44. Hart-Smith, L.J., Adhesive bond stresses and 55. Hart-Smith, L.J., Adhesively bonded joints in
strains at discontinuities and cracks in bonded fibrous composite structures, Douglas Aircraft
structures. Trans. J. Engng Mater. Tech., 100, pp. Paper 7740. Presented to the Intern. Symp.
128-144 (1978). Joining and Repair of Fibre-Reinforced Plastics,
45. Hart-Smith, L.J., Differences between adhesive Imperial College, London (1986).
behavior in test coupons and structural joints, 56. Hart-Smith, L.J., Induced peel stresses in adhesive-
Douglas Aircraft Company Paper 7066. bonded joints, Douglas Aircraft Company,
Presented to ASTM Adhesives Committee D- 14 Technical Report MDC-J9422A, August 1982.
Meeting, Phoenix, Arizona, 1981. (see also USAF Report AFWAL TR-82-4172,
46. Hart-Smith, L.J., Design methodology for bonded- (1982).)
bolted composite joints, Douglas Aircraft 57. Hart-Smith, L.J., Brown, D. and Wong, S.,
Company, USAF Contract Report AfWAL-TR- Surface preparations for ensuring that the glue will
81-3154, Vol I and II (February 1982). stick in bonded composite structures, 10th
47. Thrall, E.W, Jr., Failures in adhesively bonded DoD /NASA/FAA Conf. Fibrous Composites in
structures, AGARD-NATO Lecture Series No. Structural Design, Hilton Head Is, SC (1993).
102, 'Bonded Joints and Preparation for 58. Hart-Smith, L.J., Ochsner, W and Radeckv, R.
Bonding', Oslo, Norway and The Hague, L., Surface preparation of fibrous composites
Netherlands, April 1979 and Dayton, Ohio, for adhesive bonding or painting. Douglas
October 1979. Service Magazine, 1, pp. 12-22 (first quarter
48. Hart-Smith, L.T., Further developments in the 1984).
design and analysis of adhesive-bonded structural 59. Hart-Smith, L.J., Ochsner, Wand Radecky, R.
joints, Douglas Aircraft Company Paper 6922. L., Surface preparation of fibrous composites
Presented at the ASTM Symp. Joining of for adhesive bonding or painting. Canadair
Composite Materials, Minneapolis, MN, April Service News, 2, pp. 2-8 (1985).
1980. 60. Hart-Smith, L.T., Effects of adhesive layer edge
49. Hart-Smith, L.J., Adhesive bonding of aircraft pri- thickness on strength of adhesive-bonded joints,
mary structures, Douglas Aircraft Company Quarterly Progress Report No.3, Air Force
Paper 6979. Presented to SAE Aerospace Contract F33615-80-C-5092 (1981).
Congress and Exposition, Los Angeles, 61. Hart-Smith, L.J., Effects of flaws and porosity on
California, October 1980. strength of adhesive-bonded joints, Quarterly
50. Hart-Smith L.T., Stress analysis: a continuum Progress Report No.5, Air Force Contract
analysis approach. In Developments in Adhesives F33615-80-C-S092 (1981).
- 2 (ed. A. T. Kinloch), London: Applied Science 62. Frazier, T.B. and Lajoie, A.D., Durability of adhe-
Publishers, pp. 1-44 (1981). sive joints, Air Force Materials Laboratory
51. Hart-Smith, L.T. and Bunin, B.L., Selection of Report AFML TR-74-26, Bell Helicopter
taper angles for doublers, splices and thickness Company (1974).
transition in fibrous composite structures. In 63. Becker, E.B. et al., Viscoelastic stress analysis
Proc. 6th Conf Fibrous Composites in Structural including moisture diffusion for adhesively bonded
Design, Army Materials and Mechanics joints, Air Force Materials Laboratory Report
Research Center Manuscript Report AMMRC AFWAL-TR-84-4057 (1984).
MS 83-8 (1983). 64. Jurf, R. and Vinson, J., Effects of moisture on the
52. Nelson, W.D., Bunin, B.L. and Hart-Smith, L.T., static and viscoelastic shear properties of adhesive
666 Mechanical fastening and adhesive bonding

joints, Dept. of Mechanical and Aerospace 73. Tsai, M.Y. and Morton, J., On numerical and
Engineering Report MAE TR 257, University of analytical solutions to the single lap joint.
Delaware (1984). Intern. J. Solids and Structures (1994).
65. Mostovoy, 5., Ripling, E.J. and Bersch, C.E, 74. Benson, N.K., Influence of stress distribution on
Fracture toughness of adhesive joints. J. strength of bonded joints. In Adhesion,
Adhesion, 3, pp. 125-144. (1971). Fundamentals and Practice, New York: Gordon
66. DeVries, K.L., Williams, M.L. and Chang, M.D., and Breach, (1969), pp. 191-205.
Adhesive fracture of a lap shear joint. 75. Adams, R.D. and Wake, W.c., Structural
Experimental Mechanics, 14, pp. 89-97 (1966). Adhesive Joints in Engineering, Amsterdam:
67. Trantina, G.G., Fracture mechanics approach to Elsevier Applied Science Publishers (1984).
adhesive joints, University of Dlinois Dept. of 76. Kuenzi, E. and Stevens, G., Determination of
Theoretical and Applied Mechanics Report mechanical properties of adhesives for use in the
T&AM 350, Contract NOOO19-71-0323 (1971). design of bonded joints, Forest Products
68. Trantina, G.G., Combined mode crack extension in Laboratory Note FPL-Oll (1963).
adhesive joints, University of Dlinois Dept. of 77. Snedon, I., The distribution of stress in adhesive
Theoretical and Applied Mechanics Report joints. In Adhesion, (ed. D.O. Eeley), Ch. 9,
T&AM 350, Contract N00019-71-C-0323 (1971). Oxford: Oxford University Press (1962).
69. Keer, L.M., Stress analysis of bond layers. Trans. 78. Carpenter, W., Goland and Reissner were cor-
ASME J. Appl. Mech. E., 41, pp. 79-83 (1974). rect. J. Strain Analysis, 24(3), pp. 185-187 (1989).
70. Knauss, J.E, Fatigue life prediction of bonded pri- 79. Hart-Smith, L.J., Further developments in the
mary joints, NASA Contractor Report design and analysis of adhesive-bonded struc-
NASACR-159049 (1979). tural joints. In Joining of Composite Materials,
71. Wang, 5.5. and Yau, J.E, Analysis of interface American Society for Testing of Materials
cracks in adhesively bonded lap shear joints, NASA Special Technical Publication ASTM STP 749
Contractor Report NASA-CR- 165438 (1981). (1981).
72. Johnson, W.S. and Mall,S., A fracture mechanics BO. Hart-Smith, L.J., In Fiber Composite Analysis and
approach for designing adhesively bonded Design, Federal Aviation Administration
joints. In Delamination and Debonding of Technical Center Report DOT /FAA/CT-88/18,
Materials, ASTM Special Technical Publication Vol. 2, Ch. 3 (1988).
STP 876, American Society for Testing and 81. DoD/NASA Advanced Composites Design
Materials, pp. 189-199 (1985). Guide, 1983.
SURFACE PREPARATIONS FOR ENSURING 29
THAT THE GLUE WILL STICK IN BONDED
COMPOSITE STRUCTURES
L.J. Hart-Smith, D. Brown and S. Wong

29.1 INTRODUCTION 29.2 HISTOIUCAL BACKGROUND

Adhesively bonded joints can be no stronger During the late 1970s and early 1980s, the
than the interface between the adhesive and Douglas Aircraft Company at Long Beach,
the members being bonded together. In-ser- California, was the site of one of the techni-
vice bond failures have always been cally most successful research contracts ever
associated with weak interfaces, for both metal funded by the Wright-Patterson Air Force
and composite adherends. While most people Laboratories - the Primary Adhesively
acknowledge that adhesive bonding of metal- Bonded Structure Technology (PABST) pro-
lic structures requires strict adherence to gram. This research was directed at
proper processes, many people unthinkingly well-known problems concerning the adhe-
accept the notion that it is easy to make epoxy sive bonding of metallic aircraft structures.
stick to epoxy, for example, and pay no atten- These problems were the need to change con-
tion to the need for proper processing for the temporary processing from etching to
adhesive bonding of composite structures. anodizing, and the need to select adhesives
TItis chapter begins with a historical review and primers on the basis of long-term dura-
of the need for appropriate surface treatment to bility rather than short-term strength. These
ensure that the glue will stick to composite sur- failings had been made very clear by wide-
faces. It then focuses on photomicrographs of spread in-service problems experienced for
different surfaces, to which the adhesive will or many years by both commercial and military
will not stick, as a basis for inspections prior to operators. A successful outcome to the
bonding. Such prebond inspections are vital research was assured because of the success
because of the inability to detect weak bonds of the Redux bonding developed in England
after manufacture, until they have fallen apart. during WWII and since employed exten-
The characterization of the surfaces prepared sively by Fokker. The PABST program
by different techniques can assist in formulat- succeeded in all of its objectives, bar one. It
ing process specifications that will ensure omitted a large-scale flight demonstration
reliable adhesive bonding and in identifying program, because it would have solved no
past practices that should be discontinued. problem that had not already been solved by
the successful ground testing of panels and a
Handbook of Composites. Edited by S.T. Peters. Published complete wide-body fuselage barrel with
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 simulated wing center section. However, in
668 Surface preparations for ensuring that the glue will stick in bonded composite structures

retrospect, such a demonstration was neces- ing resins or adhesives. They create a weak
sary as a public relations exercise to convince interface, with failure occurring at only a frac-
the decision makers to apply the technology, tion of the strength that would have been
where appropriate, on the primary structure developed by grit-blasting. However, the
of large aircraft. Nevertheless, the technology resulting weak bonds cannot be detected by
did not die. The combination of phosphoric conventional ultrasonic nondestructive
acid anodizing for aluminum alloys, a pheno- inspection. Also, most bonded composite
lic-based corrosion-inhibiting primer and structures are so thin that they will not fall
adhesives formulated to resist the absorption apart even if the bond has as little as one-tenth
of water has since been applied very exten- of the strength of a properly processed bond.
sively on the primary structures (wings, Consequently, little has been done about the
fuselages, and tails) of Cessna and SAAB air- problem, in the mistaken belief that doing so
craft with excellent results. would incur an unnecessary expense.
The composites bonding industry could be On the contrary, even if local bond failures
described today as being in the same state that can be tolerated structurally, once they become
metal bonding was in 25 years ago, but with detectable in service they cause expensive
one big difference. The need to properly pre- inspection programs and possibly rework, not
pare the surfaces of composite laminates prior necessarily to add to the strength of the parts,
to bonding is acknowledged only by those but to prevent the flaws from being detected
who have suffered a major problem from hav- and repaired again on subsequent inspections.
ing failed to do so on some previous occasion. Failing to ensure that the glue sticks properly
Ironically, there are many researchers and pro- in the first place is definitely a false economy.
duction engineers worldwide who are aware The delay in solving the corresponding prob-
of the problems and their causes. However, lems on bonded metal structure gave the
there is no highly visible activity like the technology such a bad name in the USA that it
PABST program to draw attention to the was the direct cause of the extensive use of riv-
issues. This chapter cannot be expected to eted structure when bonding would have been
solve these problems, but it is hoped that it lighter, less expensive, and more durable, with
will raise the level of awareness of the subject better than a twentyfold reduction in the inci-
and that its content will help achieve better dence of fatigue cracks and a dramatic
bonded composite structures by providing improvement in damage tolerance. Lest the
standards for inspecting the surfaces before same preference for mechanical fastening
bonding, rather than creating the impossible rather than bonding continue to be followed
situation of trying to detect weak bonds after for thin composite structures, it should be
bonding. What is needed is a method that will noted that the typical fasteners cost about a
ensure a reliable bond every time, using pro- hundred times as much as the simple rivets
cedures to which all manufacturers of used with aluminum structures. Worse, the
composite structures will be willing to adhere. minimum diameter of the specialty fasteners
The first part of this challenge has already for composites is 50% larger than the diameter
been accomplished: light grit-blasting or thor- of comparable rivets. Conventional aluminum
ough mechanical abrasion has been shown to rivets are not used for carbon composite struc-
work. Unfortunately, more often than not, tures because of concern about galvanic
films of adhesive are applied to the surfaces corrosion, between the rivets and the carbon
left by the removal of a peel ply without any fibers, so composite structures that could have
surface treatment. Most, and possibly all, peel been bonded must be locally reinforced to be
plies are known not to create a suitable surface mechanically fastened instead, adding to both
for bonding, at least when used with laminat- the cost and weight.
 Historical background 669

There are, then, very great incentives for aerospace industry. It seems significant that,
reliable processing of adhesive bonds in com- today, some factories prohibit the use of bond-
posite structures. There is also no history of ing directly to a composite surface created by
properly processed composite bonds failing in the mere removal of a peel ply and that the
service. (This is also true of metal bonding. All automotive racing industry has experienced
in-service failures are the result of environ- the same kind of premature failures with peel-
mental attack, at load levels far too low to ply treatments that the aerospace industry has
have initiated mechanical failure of the bonds. suffered. If what may be called the peel-ply
Testing during the PABST program of coupons bonded joints were twice as strong as they
cut from retired aircraft stored at Davis- actually are, there would be little concern.
Monthan, Arizona, showed that there was no Conversely, if they were only half as strong,
structural deterioration after 20 years.) the problem would have been more widely
There are others that call for improvements recognized long ago and far more done about
in surface preparation for bonding of compos- it. The real problem with peel-ply bonded
ites. Robert Schliekelmann, the famous pioneer joints is that, while quite unreliable in the eyes
of Redux bonding at Pokker, was sufficiently of those closest to the subject, they usually
concerned about the failure to recognize the have sufficient strength to pass ultrasonic
need for proper surface treatment of compos- inspections (because there are no gaps) and are
ites as well as metals to make a pleal for more sufficiently strong some of the time that the
attention to this issue. Almost a decade ago, joints do not fail prior to delivery of the com-
Douglas engineers prepared an article on the ponent. This makes it difficult to present a
subjecf to help the airlines until the repair man- convincing case to non-technically minded
uals could be updated. The article was judged participants in the industry that there is a real
to be of sufficient importance to be reprinted in problem. One non-failure tends to be inter-
the Canadair house joumaP. Every experiment preted as complete success.
had to be repeated to create new photographs This chapter cannot possibly provide all the
and every phenomenon was duplicated, answers on the subject of surface preparations
thereby verifying that the problems discussed for composite bonding. It took a $20 million
in the original article really existed. A similar 5 year program to solve the corresponding
concern was expressed in England, where problems for metal bonding. It should also be
Parker and Waghom4 reported on a far more acknowledged that the concerns expressed
comprehensive test program on the effects of here are based exclusively on consistently bad
surface preparation on adhesive bond strength experiences with 180°C (350 0 P) cured epoxy
for carbon-epoxy laminates. They also con- composites. (The problem may be associated
cluded that 'composite surfaces must be with the curing of the laminate and the peel
abraded to achieve strong adhesive-bonded ply as much as with the adhesive.) It may also
joints.' Pocius and Wentz advocated the use of be significant that most, if not all, of these
Scotchbrite pads with embedded abrasive par- problems were associated with nylon peel
ticles as an effective and reliable technique for plies, whose use with a 180°C (350 0 P) cured-
achieving good composite bondss. A recent laminate is believed by some polymer
research program6 addressed both thermoset chemists to be undesirable, because the nylon
and thermoplastic composites. Reference 6 also may break down and transfer a thin film of
cites other English research. The problems still low-molecular weight material to the compos-
exist, on both sides of the Atlantic Ocean. ite surface. Photographic evidence of this
A more recent article on this subjece, voiced phenomenon is included here. Such a surface
concerns because of the reintroduction of infe- would be very difficult for an epoxy adhesive
rior techniques throughout much of the to wet, because of too low a surface energy.
670 Surface preparations for ensuring that the glue will stick in bonded composite structures

The polymer chemists express less concern peel plies. Not one in-service delamination has
about the use of nylon at 120°C (250°F), but occurred in those composite structures pre-
some adhesion failures of glued joints have pared for bonding by the grit-blast technique.
occurred with those materials also. The domi- It is hoped that this chapter will inspire
nant factor may be the low surface energy those who believe that their structural adhe-
level of the composite cured against a slick sive bonds are both strong and reliable to
peel ply, or one contaminated by a release assess composite surfaces prior to bonding. (A
agent to ensure that it can be removed easily request for information from a major US aero-
without fracturing (damaging) the composite space manufacturer about a peel ply different
laminate. from those used at Douglas resulted in their
It is certain that the criteria by which manu- switching to another peel ply. Microscopic
facturers evaluate their peel plies are precisely examination of the surface to be bonded made
the opposite of those that someone else trying it clear to them that one of their problems
to promote adhesion in a bonded joint would could be tied directly to the choice of a new
follow. Some peel plies are even coated with peel ply. Once alerted, they checked all related
release agents to ensure that they can be procedures and even changed one of their old
removed easily without damaging the basic and trusted peel plies once its true capabilities
laminate. Silicone transfer has been observed had been exposed.)
with some peel plies supposedly totally free of This chapter concentrates exclusively on the
contamination. What is needed to create a issue of surface preparation. Obviously, one
bondable surface is a tear ply that will remove must also follow correct thermal and pressure
a small amount of surface resin over the entire profiles during cure to achieve a properly
area to be bonded. However, peel plies known cross-linked adhesive that has a chance to flow
to be capable of achieving this are almost and wet the surface to which it is to be
invariably shunned because they are so hard bonded. Prebond moisture, in both laminates
to peel off, because they break and inevitably and adhesives, is also known to be a major
lead to local contamination as the process is cause of weak bonds. These and other impor-
restarted, or because there is concern that the tant issues are discussed in References 7 and 8.
interior of the laminate will be damaged. This does not imply that they are any less
It cannot be asserted on purely technical important.
grounds that there are no circumstances under We now present electron-micrographs of
which some peel plies will produce an ade- glue that failed to stick to the composite, com-
quate surface for adhesive bonding. It stands posite surfaces to which the adhesive did not
to reason that some users of this approach stick, similar surfaces to which the glue will
have been spared the problems Douglas and not stick, surfaces to which adhesive is known
others have encountered, or the products to have stuck in the past, and surfaces to
would have all been removed from the mar- which it is hoped it will stick. The effectiveness
ketplace. However, it is likely that still others of light grit-blasting as a reliable surface treat-
have also suffered, but are unaware of the ment has already been established. It was used
cause of their problems. Nevertheless, the on all bonded joints of the Lear Fan all-com-
argument against peel plies is irrefutable on posite aircraft 9, creating bonds stronger than
economic grounds. The cost of investigating the surrounding structure even when half the
weak bonds has been far, far greater than overlapping area was disbanded because of
would have been incurred if Douglas had inappropriate solid-rubber tooling used for
insisted that all suppliers and subcontractors the fuselage skin splices. The bond to the grit-
lightly grit-blast or thoroughly hand-sand blasted surface was so strong, where the
composite surfaces following removal of the mating surfaces were brought properly into
 The problem - weak bonds 671

contact, that 100% bonding was not necessary. area with a different texture is part of the sur-
Grit-blasting is used today on some of the face ply of unidirectional carbon-epoxy left
most trouble-free composite components on when the remainder separated. The reason
the MD-ll aircraft; spedfically, the compo- why this very small segment adhered is
nents made in Japan. However, some unknown. It is almost as if there were one
manufacturers of bonded composite structures small drop of chemical acting like a wetting
will prefer to continue to use peel-ply-only agent on the composite when it was bonded.
surface preparations, no matter how weak or If this should prove to be the case, and
unreliable the resulting bonds are. One must research can identify an agent capable of pro-
find different peel plies that do not release moting the wettability of cured epoxy in a
cleanly, but which take some of the surface laminate by uncured epoxy in an adhesive
resin with them, without damaging the layer, it would be a giant contribution to com-
remainder of the matrix, or find a coupling posites technology. The idea of a coupling
agent to improve the behavior of what are agent, equivalent to the silanes used for
known today to be unsatisfactory peel plies no epoxy-bonding of aluminum alloys during
matter how widely they may be used, or trans- repairs, is apparently feasible for composites
fer the production to factories in which reliable also - at least when the peel ply has not been
surface-treatment practices are followed. coated with a release agent. Coupling agents
would be more likely to work if the basic lam-
inate were not fully cured prior to bonding.
29.3 THE PROBLEM - WEAK BONDS
One would prefer an incomplete initial cure in
No more convincing proof of the existence of order to leave some active chains in the mole-
weak bonds created on peel-ply composite cular structure of the composite to which the
surfaces can be found than in Fig. 29.1. This is glue could bond. The bond cycle would then
a scanning electron microscope (SEM) pho- be relied upon to complete the cure of the
tomicrograph not of a composite surface after laminate and this process might affect the
the peel ply was lifted off, but of the cured choice of adhesive. However, such an out-
adhesive, showing the imprint left by the com- come must be preceded by an
posite laminate after it peeled off! The small acknowledgment that the remainder of
Fig. 29.1 and the large similar surrounding
area imply the existence of a problem.
Figure 29.2 shows the same lack of adhesion
evident in Fig. 29.1 at a small magnification
over a very much larger area, to show how
widespread this phenomenon can be. The tex-
ture of the peel ply is dearly imprinted in the
glue over almost the entire area. The different
texture visible on the left side of the figure is
the peel-ply imprint on the underlying com-
posite laminate. In this area, the adhesive
failed to bond to the lower surface.
Throughout everywhere else shown in Fig.
29.2, the adhesive failed to bond to the other
composite part, the peel-ply imprint of which
is embossed on the adhesive. Figure 29.3
Fig. 29.1 Imprint of peel ply, in background, on shows an adjacent area on the same part. The
cured adhesive film that failed to adhere (mag x30). thick glue layer (shiny area) at the bottom left
672 Surface preparations for ensuring that the glue will stick in bonded composite structures

Fig. 29.2 Large area of adhesive that failed to bond,

showing imprint of peel-ply surface (mag x5).
Fig. 29.3 Adhesive that failed to bond, also show-
ing sanding confined to high spots of the composite
of Fig. 29.3 shows the unmistakable imprint of weave (mag x5).
the fine-weave peel ply left by the other com-
posite part where it did not stick. Immediately virtually impossible to sand the surface with-
above it, there is a small area of cohesive frac- out breaking the fibers. One needs to use
ture of the bond. (The reason for bonding in preimpregnated peel plies or to include a layer
just that small area is not known.) The dark of resin (or adhesive) film between the basic
area in the lower right of Fig. 29.3 shows the lay-up and any dry peel ply to create what
imprint of the peel ply in the composite under- amounts to a gel coat sufficiently thick to sand
neath the adhesive layer. The small vertical as the surface for bonding or painting. It is also
white lines define the fine weave of the peel impossible to completely sand a surface con-
ply. The more widely spaced larger white taining ply drop-offs or local buildups. Only
blotches, with some sharply defined edges, light grit-blasting would work in such a case.
reveal the very small extent of the sanding car- It should be noted that, at the time when the
ried out on this sample. The white blotches are bond shown in Figs. 29.1 to 29.3 was manufac-
the high points in the coarser weave of the tured, conventional ultrasonic inspection
plain-weave cloth in the basic laminate. These failed to find any evidence of a defect. The gap
are evident on the surface opposite from that had not yet opened up even though an after-
cured against the lay-up tool. the-fact inspection revealed almost no
These large white blotches in Fig. 29.3 illus- adhesion throughout most of the area meant to
trate one unavoidable problem with sanding have been bonded. This is very significant.
as a surface preparation. How can one sand Weak bonds cannot be detected by inspection
down to the bottom of every depression in the before they have failed. They can be prevented
fine-weave texture left by the peel ply without only by adhering to process specifications that
sanding into the fibers at the top of every ensure they will not occur.
raised bump in the texture of the composite Figure 29.4 shows the typical composite
material itself? This problem is aggravated by surface left by removal of the same type of
the use of woven fabrics rather than unidirec- peel ply (in this case, a corona-treated nylon
tional tape and is unavoidable on the bag side peel ply). A comparison with Fig. 29.1 will
of the laminate, unless a caul plate is used on leave no doubt about the origin of the surface
both sides. The customary use of dry (nonim- embossed into the cured adhesive layer.
pregnated) peel plies, which absorb resin from (Scientific testing of this particular piece of
the surface plies of the laminate, results in adhesive revealed nothing untoward about
such an ultrathin layer of surface resin that it is the way it was cured, although there are
 The problem - weak bonds 673

Fig. 29.4 Imprint of nylon peel ply in epoxy surface

of composite laminate (mag x65).

unconfirmable suspicions that prebond mois-

ture may have exacerbated the problem.)
An examination of the surface of the nylon
peel ply itself after removal from the laminate
Fig. 29.5 Enlargement of nylon peel ply after strip-
revealed no resin adhering to the peel ply (Fig.
ping off composite laminate cured at I80 a C (350 D F)
29.5). [Regrettably, this is not from the same (mag X4I).
sample, even though it is nominally the same
peel ply cured in the same laminating resin, at
180 DC (3S0 DF).] Significantly, the other side on Higher resin contents in aerospace composite
the peel ply was covered by a considerable structures do not necessitate no-bleed cure
amount of resin bled from the surface plies of cycles. They retain a higher resin content, but
the laminate. This is most undesirable, they also prevent the removal of any volatiles
because most composite specifications call for absorbed and trapped before cure or gener-
unreasonable resin starvation as the norm. ated during the cure. The lack of volatile
(The purpose of minimizing the resin content removal introduces a whole new family of
is to boost the easily measured in-plane ten- processing problems and structurally inferior
sion strength. Resin-dominated strengths are laminates.
both more difficult to measure and more diffi- Figure 29.6 shows a greatly magnified view
cult to analyze, so few such considerations are of the same peel ply shown in Fig. 29.5. The
included in the design process.) There is usu- peeling motion was from right to left. (The cor-
ally barely enough resin to hold the fibers responding picture to the left of the vertical
together. Removing surface resin makes it bundle of peel-ply fibers shows far less fractur-
more difficult to transfer loads through ing of the matrix.) Noteworthy features shown
bonded joints. It also makes the surface more in Fig. 29.6 include (1) the tendrils of resin
prone to impact damage and more difficult to around the vertical filament on the extreme
prepare for repainting in service. This is left, (2) the way the resin has pulled away
acknowledged by the way in which some cleanly from the horizontal filaments on the
Boeing components are now made with right of the adjacent vertical fiber, (3) striations
preimpregnated peel plies despite the added in the fracture surface of the resin between the
cost and storage and handling problems with second and third horizontal filaments from the
respect to dry ones. top, (4) the loose piece of resin between the first
674 Surface preparations for ensuring that the glue will stick in bonded composite structures

in thick test coupons with impractically short

bonded overlaps, provided that the surface
treatment and cure are carried out properly.
The resulting fracture surface is shown greatly
magnified in Fig. 29.7. The randomly oriented
fibers are the carrier in the adhesive layer and
quite distinct from the regular patterns of
woven composite laminate and peel plies.
There is no possibility of mistaking this cohe-
sive fracture surface for either the interfacial or
interlaminar failures shown in Fig. 29.1.
Apart from the grossly dissimilar surface
textures, most adhesives are distinctively col-
ored. Carbon-epoxy is invariably black. It is
very easy to distinguish between:
1. Cohesive failure of the adhesive with com-
plete coverage of both surfaces by the
adhesive layer.
2. Adhesion failure with all of the adhesive on
Fig. 29.6 Highly magnified view of peel ply one surface, although the adhesive may
removed from carbon--epoxy laminate cured at fracture and end up as small segments
180°C (350°F) (mag x41).
adhering to both surfaces, but only to one in
any given area (Fig. 29.8). The figure shows
and second horizontal filaments from the bot-
predominantly the surface of the adhesive
tom, and (5) the white markings on the nylon
to which the overlaying piece of composite
filaments themselves. At still higher magnifica-
tion, these markings appear to be crazing that
occurs within the nylon filaments as they were
bent while the peel ply was being stripped off.
The marks do not appear to be resin extracted
from the surface of the laminate. There was no
indication of any matching roughness in the
grooves left in the laminate for the samples that
did match this particular piece of peel ply,
although, as discussed later, there is some indi-
cation at very high magnification that ultrathin
layers of nylon transferred to some areas (but
not all) of the composite laminate as it cured.
(There was also no evidence of nylon transfer
when nominally the same peel ply was used
with a different resin matrix and cured at
120°C (250°F).
Had the adhesive in Fig. 29.1 stuck properly,
failure would have occurred interlaminarly
within the thin carbon~oxy composite lami-
nate, with no adhesive visible anywhere. One Fig. 29.7 Cohesive fracture within adhesive bond
can enforce a cohesive failure of adhesive layers (mag x83).
 Samples of differently prepared surfaces for bonding 675

3. Interlaminar failure of the composite lami-

nate, with fibers on both surfaces. (Failure
close to the surface of the composite may
leave a transparent layer of resin matrix
covering the adhesive, but microscopic
examination will reveal a very different
fracture surface from that associated with
the adhesion failure shown in Fig. 29.1.)
Naturally, there can be no photograph equiva-
lent to Fig. 29.1 for a grit-blasted surface
because the interface will not fail unless it was
deliberately contaminated prior to bonding.

29.4 SAMPLES OF DIFFERENTLY PREPARED

SURFACES FOR BONDING

The remainder of this chapter will provide a

comparison between photomicrographs of dif-
ferent surface preparations, good and bad.
Figure 29.9 shows the surface left by light
grit-blasting with alumina grit at 140 kPa
(20 psi). Although resin has been removed
from the entire surface, so little has been
removed that the texture of the peel ply is still
evident. The uniformity of the surface treat-
ment, prepared under less than ideal
conditions with no handling aids, is impres-
sive and suggests that the procedure is not
unreasonably demanding, particularly when
done with suitable equipment. A very highly
magnified photo (Fig. 29.10) confirms that no
damage was done to the fibers in the car-
bon-epoxy laminate. The grooves in this
figure are not carbon fibers; they are the fur-
rows left by removing the peel ply.
The same grit-blast machine used to pro-
duce the sample illustrated in Figs. 29.9 and
29.10 had many years earlier burnt holes in
Fig. 29.8 Predominantly adhesion failures in peel- 6.3 mm (0.25 in)-thick laminates in only 20 s
ply bonded joint (mag X5). when operated at 698 kPa (100 psi). Restricting
the blast pressure is critical when using this
failed to adhere and local dark exposed technique. Because the negatives of the photos
areas of the underlaying composite with the used in the earlier article 2 had been lost, the
peel-ply imprint made clearly visible as blast pressure was turned up to provide a
those portions of the adhesive lifted off comparative illustration of what happens
with the upper piece of composite. when the blasting is overdone. Despite these
676 Surface preparations for ensuring that the glue will stick in bonded composite structures

Fig. 29.9 Lightly grit blasted composite surface, Fig. 29.10 Highly magnified grit-blasted epoxy sur-
retaining imprint of peel ply (mag XSO). face, showing no damage to underlying fibers (mag
xlOOO).

explicit instructions, the process did not seri-

ously damage the laminate. The new
overblasted surface is shown in Figs. 29.11 and
29.12 at low and high magnifications. The
greatest difference with respect to Figs. 29.9
and 29.11 is that the texture visible in Fig. 29.11
is that of the weave of the carbon fabric. The
imprint of the peel ply has been totally
removed. (Figure 29.11 is at a far lower magni-
fication than Fig. 29.9.) The vertical bands in
Fig. 29.12 are the fibers. Only a tiny fraction of
the surface fibers is damaged and the surface
is rough enough for the adhesive to stick.
The difference in size between the weaves
of peel plies and typical composite fabrics is
made very clear in the peculiar failure surfaces
shown in Fig. 29.13. The uppermost areas are
of the imprint of the fine-weave peel ply, to
which the adhesive refused to bond. The lower
coarser weave shows exposed carbon fibers Fig. 29.11 Deliberately over-blasted epoxy compos-
left after very local interlaminar failures in the ite surface, showing weave of carbon fibers with no
trace of the fine-weave peel-ply surface (mag x9).
uppermost ply. The reason for such rapidly
alternating failure modes is unknown,
although it is probably significantly related to since the exposed carbon fibers are all oriented
the direction of propagation of the disbond in the same direction.
 Samples of differently prepared surfaces for bonding 677

The lightly done and deliberately overdone

grit-blasted surfaces in Figs. 29.9 to 29.12 are
consistent with the ease and reliability with
which the grit-blasting was carried out on the
Lear Fan all-composite aircraft. A portable
low-pressure machine (Fig. 29.14) was used
then. It had a vacuum collector to minimize
the spread of debris, and the central nozzle
was surrounded by a bristle comb, which
ensured that the nozzle was always held at the
same distance from the surfaces. The samples
shown in Figs. 29.9 to 29.12 were made under
far more difficult circumstances - with cum-
bersome gloves to hold the part and the nozzle
inside a chamber. Production parts as good as
those in Figs. 29.9 and 29.10 should be far eas-
ier to produce.
Figures 29.15 to 29.17 show low, intermedi-
ate, and high magnifications of the surface left
Fig. 29.12 Highly magnified over-blasted compos- by removal of a nylon peel ply from a car-
ite surface still showing minimal damage to fibers bon-epoxy laminate cured at 180°C (350°F)
(mag XSOO). The imprint is consistent with the intermedi-
ate magnification shown in Fig. 29.4. The
imprint left by each fiber in the peel ply is so
smooth and slick that the adhesive will not

Fig. 29.13 Disbonded surface, showing peel-ply

imprint where adhesive failed to bond and local
interlaminar failures in underlying composite (mag Fig. 29.14 Portable low-pressure grit-blasting
XIS). machine.
678 Surface preparations for ensuring that the glue will stick in bonded composite structures

Fig. 29.17 Same sample at higher magnification

Fig. 29.15 Peel-ply imprint on epoxy surface to (mag x750).
which the glue will not adhere (mag x38).

adhere there. It can adhere only to the exceed-

ingly narrow strips of fractured resin between
each fiber depression. The adhesive can
adhere to the entire composite surface shown
in Fig. 29.9, a bondable area that is an order of
magnitude greater, at least.
Figure 29.17 shows something not evident
in the companion photo of the peel ply surface
itself (Fig. 29.6). At a magnification of x750,
some of the furrows appear to be coated in
places by ultrathin coatings of nylon from the
peel ply. The lesser magnification of x215 (Fig.
29.16) of the same area on the surface indicates
that some but not all other areas are similarly
coated.
Figure 29.18 shows what some regard as
adequate sanding. Yet, on close examination, it
is evident that the peel-ply imprint is clearly
visible except for a miniscule fraction of the
total area. The sanded area lies outside the
area meant to be bonded, on the left, but there
Fig. 29.16 Magnified peel-ply imprints showing is no reason to suspect any different degree of
transfer of ultra-thin layer of nylon to otherwise sanding. Unfortunately, many specifications
smooth furrows in composite surface (mag ><225). call for light (or scuff) sanding because of
 Samples of differently prepared surfaces for bonding 679

stiffener as in the unbonded area to the right.

Figure 29.19 shows slight damage to the
uppermost fibers as the result of thorough
sanding of another sample, this time a unidi-
rectional carbon-epoxy laminate, prepared at
a different facility. In this case, sanding paral-
lel to the surface fibers would minimize any
damage. However, that option would not be
available for woven fabrics because the sur-
face fibers do not all run in the same direction.
The number of fibers damaged in Fig. 29.19 is
an extremely small fraction of even the surface
ply. Carbon fibers are so small that each tow
contains 3000-12 000 individual fibers in a
layer 0.13 mm-D.33 mm (0.005-0.013 in) thick.
The loss of strength from damaging a few sur-
face fibers through too energetic sanding is far
less than from a bonded structure coming
apart in service because of inadequate sand-
ing.
Figures 29.20 and 29.21 show sanded sam-
ples. The woven texture is that of the peel ply,
not the carbon fabric. The matt sanded areas in
the figure are the high spots in the laminate
surface. The peel-ply imprints still visible are
in depressions. Clearly, the operator never
came close to damaging the carbon fibers.
Sanding should have continued for far longer
Fig. 29.18 Inadequate abrasion achieved by scuff or a more abrasive grade of emery paper
sanding (mag x12). should have been used. It is clear that some
30% of the bond area was sanded, a huge
greater concern that the fibers in the compos-
ite laminate not be damaged than for the need
for sufficient mechanical abrasion to ensure
that the adhesive will stick properly. As long
as fifteen years ago, the McDonnell (St. Louis)
specifications explicitly called for sufficient
sanding to eliminate the imprint of the peel
ply. In other words, the entire surface of the
composite laminate was to be stripped back,
just as in Fig. 29.9. (One wonders what moti-
vated such a precise specification. Something
taught them a lesson they planned not to for-
get.) The results of the inadequate level of
sanding are evident in Fig. 29.18. The imprint Fig. 29.19 Slight fiber damage caused by thorough
of the peel ply is just as clear to the left of the sanding of unidirectional carbon-epoxy laminate
broken glue fillet at the edge of the disbanded (mag X460).
680 Surface preparations for ensuring that the glue will stick in bonded composite structures

Fig. 29.20 Moderately sanded peel-ply imprint on Fig. 29.21 Highly magnified surface, showing how
surface of woven composite laminate, showing how most of the peel-ply imprint remains after hand
sanding does not abrade the entire surface (mag sanding (mag xlOOO).
xlOO).
improvement over the few percent removed in
the sample shown in Fig. 29.18.
While it must be acknowledged that both
grit-blasting and hand-sanding can be over-
done, doing so takes time and effort if one is
using the right abrasives and equipment. A
significant loss of strength from such actions is
far less likely than from either simply remov-
ing a peel ply or sanding the composite
surfaces far too lightly.
To cover the possibility that the bad experi-
ences with peel plies at Douglas might have
been associated exclusively with the break-
down of nylon at too high a curing
temperature, samples were obtained from
Oxford Brookes University, which used a dif-
ferent laminating resin and both a polyester
peel ply and what is probably the same nylon
peel ply. The polyester peel ply was noticeably
more difficult to remove than the nylon peel
ply, but far easier than what were referred to
as tear plies in Reference 2. The surface created Fig. 29.22 Imprint of polyester peel-ply, showing
by removing the polyester peel ply from a more fractured resin than with nylon peel-plies
120°C (250 0 P) cured carbon-epoxy laminate (mag X50).
 Samples of differently prepared surfaces for bonding 681

(Fig. 29.22 mag x37.5) is generally very similar

to the surface shown in Fig. 29.4 after a nylon
peel ply had been removed, except for the ran-
domly oriented tendrils, which appear to be
lengths of attached polyester filaments. The
university's nylon peel ply left slick imprints
almost exactly like the one used at Douglas. A
comparison of the SEM photographs suggests
that the polyester peel ply performed better
because the filaments were smaller and there
was proportionately a greater length of the
very fine strips of fractured resin between the
filaments. The improvement in strength from
this source alone should be about 25%.
However, there was another possible differ-
ence: the higher magnification of the polyester
peel-ply imprint (Fig. 29.23) shows faint
streaks in the grooves left by removal of the
peel ply. If these could be traced to surface
roughness in the resin, rather than to smooth Fig. 29.24 Highly magnified image of polyester
irregularities in the polyester filaments, it peel-ply after removal from 120°C (250 0 P) cured
would be an indication that the peel ply carbon-epoxy laminate (mag x830).
stripped off some surface resin and promoted
adhesion of the glue. Unfortunately, a microscopic examination of
the removed peel ply itself suggests that the
imprint remains smooth at the molecular level.
The polyester peel ply seems to be better than
the nylon one but not in the same class as the
surface created by light grit-blasting. Figure
29.24 shows a x830 enlargement of the surface
of the removed polyester peel ply. The fibers
are generally smooth and, where they are not,
the visible strands represent tearing of the
fibers rather than a buildup on the surface as
the result of tearing the matrix resin.
Nevertheless, it should be noted that the tear-
ing shown could not have occurred if there
were no adhesion at all between the peel ply
and the resin matrix. The interface definitely
seems more bondable than the much smoother
one associated with nylon peel plies.
What may be a far more significant differ-
ence is that the surface created from peeling off
a nylon peel ply from the carbon-epoxy lami-
nate cured at 1200 e (250°F) shows no sign of
Fig. 29.23 Highly magnified imprint of polyester nylon on the surface at either low or high mag-
peel-ply in epoxy composite (mag x500). nification (Figs. 29.25 and 29.26). Nevertheless,
682 Surface preparations for ensuring that the glue will stick in bonded composite structures

the surfaces are so slick that one would have no

confidence that a reliable bond could be made
other than to the thin strips of fractured resin
between the grooves left by the peel ply.
However, it might well be far less weak than
would result from trying to bond to a nylon
interface and, indeed, lap-shear testing of such
bonded joints has been reported in Reference 6
as resulting in cohesive failures at the expected
bond strength. The fractured surfaces look
excessively porous, but so do the companion
coupons made with grit-blasted surfaces,
which developed essentially the same strength.
These bonds definitely do not look as strong as
that shown in Fig. 29.7, but they are ever so
much stronger than that shown in Fig. 29.l.
Figure 29.27 shows the fractured surface of
an interlaminar failure within the
carbon-epoxy composite. The fibers shown
Fig. 29.25 Slick surface left by removal of nylon are carbon and are noteworthy for the failure
peel-ply from 120°C (250 0 P) cured carbon-epoxy of the resin to adhere to them over most of
laminate (mag x46). their surface. The fractured surface is that of
the resin matrix. One would expect that the
same kind of surface would be created by the

Fig. 29.26 Highly magnified image of nylon

peel-ply imprint taken from 120°C (250 0 P) cured Fig. 29.27 Extreme enlargement of interlaminar
laminate, showing slick surface with no transfer of fracture in top layer of carbon-epoxy laminate (mag
nylon (mag X450). XllOO).
 Commentary 683

removal of a tear ply that adhered to the the production level in over 20 years. Because
matrix as tenaciously as conventional peel the metal-bond problem was resolved so
plies refuse to adhere. This, in tum, raises quickly, very few panels were involved. There
questions about the feasibility of tear plies as a were no in-service failures because every
surface-preparation technique since the lami- affected panel was identified before delivery
nate is damaged far more than by light and reinforced by rivets.
grit-blasting, as a comparison between Figs. Peel-ply 'surface preparation' for bonding
29.10 and 29.27 will attest. of composites has not been as thoroughly
explained. There have been many instances of
such weak bonds not being detected until they
29.5 COMMENTARY
had split apart in service, even though there is
In the mid-1980s, a major effort was aimed at no reason to believe that the bonds wore out
the resolution of a problem associated with under mechanical loads. The observed modes
adhesive bonding of aluminum alloy struc- of failure are consistent with a manufacturing
tures. In that case, as in the problem discussed problem. There is little doubt that gentle grit-
here, the glue failed to stick and interfacial fail- blasting is the most reliable method of
ures resulted. However, there remains a very preparing thermoset composite surfaces for
great difference between the two cases. In the bonding. Sanding can work only on fairly fine-
earlier case, every factor associated with the weave cloths in composite laminates.
weak bonds that passed all ultrasonic inspec- Otherwise, it is not possible to sand to the bot-
tions was identified in a matter of weeks. The tom of all depressions left by the peel ply
conditions had been replicated in a laboratory without also sanding significantly into the
and the investigation closed within a few structural fibers. A few other manufacturers
months. Significantly, there were no loose have used tear plies, which leave a completely
ends and the problem has not recurred. The fractured resin surface, rather than the more
primary cause of the problem was condensate commonly used easily removed peel plies dis-
on adhesive film that had been removed from cussed here. More use should be made of tear
storage before it had thawed out. plies. However, if a tear ply were used on a
However, there was a second factor lightweight honeycomb or foam sandwich
involved as well. The first violation of proper panel, there is a good chance that the core
processing procedures would not cause a would fail instead.
defective bond unless the moisture was There can also be no doubt that when prob-
trapped at the interface between the details. lems have occurred as a result of bonding
This condition happened consistently with directly to the surface left by removal of a peel
one kind of bonding tool, while similar parts ply, or one with totally inadequate sanding,
made concurrently on a different kind of tool they are widespread and serious. These prob-
showed no such problems. The second kind of lems have been experienced at many places.
tool permitted complete ventilation of any Yet, if they happened as repeatedly as the
trapped volatiles as well as of any generated defects caused by the combination of the two
during the cure. There were no large area factors cited above, one would have expected
defective bonds. Both tools produced local such a backlash against the procedure that
bond defects where a tool or the parts were out every umeliable peel ply would have been
of contour. These problems were eliminated withdrawn from the marketplace, preventing
by correcting the tools and by better straight- any recurrence of the problems. That has not
ening of the stiffeners. This problem was happened so, presumably, some organizations
resolved very quickly, whereas the present are able to bond successfully to composite
problem with peel plies has not been solved at structures by simply removing a peel ply.
684 Surface preparations for ensuring that the glue will stick in bonded composite structures

Regrettably, it seems as if anyone doing this for bonding has been found to be extremely
process successfully has no reason to investi- reliable in service. While one needs to buy
gate why he is successful, so the differences appropriate equipment, the cost of doing so is
between his techniques and those that lead to a small fraction of the typical cost of even one
trouble have remained unidentified. composite detail. With the right equipment,
It seems that there is something else training of technicians is straightforward.
involved as well possibly prebond moisture or There can be no valid argument in favor of
something associated with venting during the not mastering the art of grit-blasting. While
cure or possibly the breakdown and transfer of there may be choices for the surface prepara-
nylon peel plies cured at too high a tempera- tion during initial fabrication, peel plies
ture or the transfer of a release agent on some cannot possibly be stripped off a second time
peel plies. Weak interfacial bonds between the to create a new clean' surface during repairs.
I

adhesive layer and both composite adherends It may be that some manufacturers make
should never have been strong enough to rip a successful bonds to peel-ply composite sur-
properly cured adhesive back and forth from faces already. It may also be possible that
one interface to the other in the manner that is coupling agents may be found to enable others
so evident in Fig. 29.8. And, more significantly, to do the same. However, it is undeniably true
attempts to replicate the weak bonds in the that, for at least two decades, some manufac-
laboratory have been inconclusive and, at turers of composite components who have
times, inconsistent. This merely reinforces the relied on peel plies alone as surface prepara-
assumption that the problem is not yet fully tion for bonding have created weak bonds that
understood without, in any degree, diminish- have fallen apart in service. The cost of unan-
ing the conviction that the problem is serious ticipated repairs and investigations has greatly
and needs to be resolved. exceeded any expected initial cost savings.
It would be helpful if there were a reliable Given the widespread nature of these prob-
a peel-dominated quality-control test for com- lems, it is appropriate to recommend that the
posite surface preparation that was use of peel-ply surface preparation alone be
equivalent to the wedge-crack test used for discontinued unless it can be shown that its
metal bonding. Unfortunately, the experi- use never results in interfacial failures
ments performed to date have been between the composite and adhesive layers. It
bedevilled with extraneous influences that are is quite clear, from experience, that the cost of
not yet understood but have a much greater even one in-service bond separation exceeds
effect on the strength of the coupon than the by far the savings derived during an entire
variations in surface treatment. Neverless the production run by not thoroughly abrading
goal remains. the surface.
Scanning electron microscope images,
such as those presented here, show easily dis-
29.6 CONCLUSIONS
tinguishable differences between the matt
As acknowledged above, more technical infor- rough surface created by mechanical abrasion
mation needs to be uncovered about adhesive and the microscopically smooth furrows
bonding to fiber-polymer composite surfaces associated with peel-ply removal alone. It is
created by simply removing a peel ply. The not difficult to ensure that the surface prepa-
weak bonds associated with this technique are ration is adequate for bonding or painting
a financial burden that is both serious and easy before the bond has been made. Conversely, it
to avoid. is extremely difficult to detect a weak bond
The use of low-pressure grit-blasting as the nondestructively once the error has been
final step in the preparation of these surfaces made.
 References 685

REFERENCES 5. Pocius, A.V. and Wenz, RP., Mechanical surface

preparation of graphite-epoxy composite for
1. Schliekelmann, R.J., Adhesive bonding and
adhesive bonding, SAMPE J., 1985, Sept/Oct,
composites, Progress in Science and Engineering of 50-58.
Composites. In Vol. I, Proc. 4th Inti Conf. 6. Wingfield, J.R.J., Treatment of composi~e sur-
Composite Materials (Tokyo), 22-28 Oct., 1982.
faces for adhesive bonding, Int. J. AdheslOn and
Japan Society for Composite Materials and The Adhesives, 1993, 13(3),151-156.
Metallurgical Society (TMS) of AIMEi (T.
7. Hart-Smith, L.J., Joining of organic-matrix com-
Hayashi, K. Kawata and S. Umekawa, eds)
posites. In ASM Handbook, Vol. 6: Welding,
Amsterdam: North-Holland, 1982, pp. 53-78.
Brazing, and Soldering, Ohio: ASM Inti, Dec.
2. Hart-Smith, L.J., Ochsner, RW. and Radecky, 1993, pp. 1026-1036.
RL., Surface preparation of fibrous composites
8. Mahoney, CL., Fundamental factors influenc-
for adhesive bonding or painting, Douglas
ing the performance of structural adhesives,
Service Magazine, 1984, First Quarter, pp. 12-22. Internal Report, Dexter Adhesives & Structural
3. Hart-Smith, L.J., Ochsner, RW. and Radecky,
Materials Division, The Dexter Corporation.
RL., Surface preparation of fibrous composit~s 9. Hart-Smith, L.J., Design and development of
for adhesive bonding or painting, Canada!r
the first Lear Fan all-composite aircraft,
Service News, 1958, 14(2), 2-8. Douglas Paper 8184, presented to Institution of
4. Parker, H.M. and Waghom, RM., Surface pre-
Mechanical Engineers Conference on Advanced
treatment of carbon fibre-reinforced composites
Composites, London, England, March 7-8,
for adhesive bonding, Composites, 1982, 13, 1989.
280-288.
LAMINATE DESIGN 30
Jocelyn M. Seng

30.1 INTRODUCTION birch or spruce, laid over balsa core or fir

stringers to form a sandwich structurel • This
An early example of laminated composite
successful production aircraft (7781 units built)
materials is the de Havilland Mosquito
was designed without the analytical techniques
fighter /bomber used by the British Royal Air
described in this chapter and without fancy
Force during World War II (Fig. 30.1). This air-
computer tools. With current laminate design
craft was built entirely out of wood because of
and analysis techniques, today's higher-perfor-
limited metal supplies and the need for quick
mance composite aircraft are made possible;
delivery. The wings, for example, were made as
with the increase in speed and accuracy of com-
three-ply skins (each 1.5 mm (0.060 in) thick) of
putation results, designer confidence in

Fig. 30.1 The de Havilland Mosquito, an all-wood

Handbook of Composites. Edited by S.T. Peters. Published WWII production aircraft. Printed with permission,
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 Zokeisha Publications, Inc.
 Laminated plate theory 687

composites structures is increased. This chapter 30.2 LAMINATED PLATE THEORY

presents the basic mathematical tools used to
30.2.1 LAMINA
design laminates and provides insight on the
many options for optimizing the material for Advanced composite materials are typically
particular needs. supplied as a thin layer, called a ply or lamina,
Three distinct levels of benefit can be which is subsequently stacked into a thin
derived when using composites and laminate plate, called a laminate. A unidirectional ply or
design. With equal fiber distribution in multi- lamina is a flat or curved layer of fibers ori-
ple directions, rendering an effectively ented in one direction and held together by
quasi-isotropic material, composites can matrix material that serves to support the
approximate metals while providing a weight fibers. The stresses perpendicular to the planar
savings due to the difference in material den- surface are assumed to be zero.
sities. In addition, however, designers opting While the behavior of isotropic materials can
to use composite parts enjoy the advantage of be descnbed with two elastic constants (typi-
being able to tailor the properties of their cally the Young's modulus and the Poisson's
material by orienting load-carrying fibers in ratio) and one strength value, a composite ply
the directions that there are loads. The result is that is transversely isotropic is characterized by
an anisotropic material, which by definition is a four elastic (stiffness) constants and five
material with different properties in different strength parameters in two-dimensional analy-
directions. Ultimately, the composites industry sis. The material properties are defined along
is finally beginning to see the development of the fiber (x--direction ) and perpendicular to the
unique structures that have never before been fiber (y--direction ). For each unidirectional ply
attempted and with material behavior that is in its own axes, the four orthotropic elastic con-
only possible with distinct laminate designs. stants are the longitudinal tensile modulus, E~
This is a result of coupled behavior, for exam- the transverse tensile modulus, E( the major
ple, an extensional load on an anisotropic Poisson's ratio, v; and the shear modulus, E..
material can yield extension coupled with Only one Poisson's ratio is necessary since v =
bending and twisting deformations. vx(E/E). The five strength parameters for elch
The objective of this chapter is to outline a unidirectional ply are the longitudinal tensile
method to design ply layouts which achieve strength, X; the longitudinal compressive
structural design goals for composite parts. strength, X'; the transverse tensile strength, Y;
This method is based on laminated plate the- the transverse compressive strength, Y'; and the
ory used with the quadratic failure criterion2 • shear strength, S.
This general discussion, which assumes famil- The five initial coupon tests to experimen-
iarity with undergraduate mechanical tally determine the nine material constants are
engineering fundamentals, shows how the shown in Fig. 30.2. In the stress, (J, versus
principles for isotropic materials (such as met- strain, £, plot, the material is characterized by
als) are extended to the analysis of advanced the slope of the line, which represents material
composites. The basic equations are presented stiffness, and by the failure point, which
and the analysis procedure is outlined. defines the maximum stress that the material
Simplifying concepts are introduced and dis- can sustain, i.e. its strength, and by the corre-
cussed. Simple computer codes that embody sponding maximum strain. Use of the strength
these equations are now widely available, parameters is discussed in the section on fail-
making it unnecessary to ever have to solve ure criteria.
these equations by hand. Examples and sam- During three of the coupon tests, the
ple problems are included to demonstrate nonzero strains are monitored and the rela-
concepts. tionships in eqn (30.1) are determined. The
688 Laminate design

Stress, 0"
I
Longitudinal

§ ,
Tension
0"
Transverse
Tensjon
cr -
Shear

~
+-

Ex ': •

v = - £y
x Ex

X'
'~I --{ill]-
Long~udinal Transverse
Compression Compression

Fig. 30.2 Coupon tests to determine the nine material constants used to characterize an anisotropic material.

material is assumed to be linear and elastic; H a unidirectional specimen was simultane-

thus, the stiffness of a material is the same in ously tested under the three load cases,
tension as in compression. Based on the four longitudinal tension, transverse tension, and
elastic constants, infinitely different lami- shear, then superposition of the strains results
nates can be designed using laminated plate in
theory.
fx
= ~a
E x -~a
E Y (30.2)
x y
Longitudinal tension test
1 v
1
f -a __xo
y EYE x
ex = -Eax (30.1) Y x
x

Vx
ey = -vex x = - -E0x
x Written conveniently in matrix notation,

Transverse tension test

1 _2:L 0
1 Ex Ey

{::l
ey = -Eay _22. 1
0
t:}=
Y
Ex Ey
= - _E
V 1
fx = -vyf y Ya
y 0 0
y E.

nh}
Shear test Sry

-0
1
rn
= Sry
0
Syy
0
(30.3)
f.
E• •
 Laminated plate theory 689

Defining the plane stress stiffness matrix [Q] = laminate, x, y, sand 1,2, 6 are interchangeable.
[5]-\ another form of eqn (30.3) is Material properties are specified with
respect to the on-axis coordinates. The proper-
ax} = {Qxx Qxy flex}
ey
{a y Qxy Qyy
00 ties of an off-axis ply, anything other than 0
degrees, can be calculated by transforming the
as 0 0 Qss es properties of the O-degree ply. The angle of
transformation, 0, is equal to the ply angle
shown in Fig. 30.3, where 1 and 2 are the lam-
Ex vxEy inate axes and x and yare the rotated ply axes.
0
(I-vxv y) (1- v xV y) ois positive counterclockwise from the I-axis
to 90 0 , and negative clockwise to -90 0 •
{::}
vyEx Ey
= (I-vxv y) (l-v x v y)
0 (30.4)

0 0 Es

This calculation of the plane stress stiffness

matrix [Q] for a single ply is the starting point
of laminated plate theory, once the engineer-
ing constants have been experimentally
determined.

30.2.2 COORDINATE TRANSFORMATIONS

Fig. 30.3 Definition of ply axes (x,y) and laminate
Two coordinate systems are used in laminated axes (1,2), where the lines indicate the fiber direc-
plate theory. The local, or on-axis, coordinate tion.
frame is defined by x and y, also referred to as
the ply axes. The x-axis is along the longitudi- The laminate off-axis stiffness matrix is com-
nal direction of the ply (along the fiber); the puted from the ply on-axis stiffness matrix by
y-axis is in the same plane, but in the trans- using the following transformation relation:
verse direction (perpendicular to the fiber
direction). The subscript s is used with expres-
sions for shear, and is a contraction for the
subscript xy. The ply's material properties are
defined in this axis system. Since not all plies
are aligned in the same direction along the
principal loading axis in a laminate, a second
set of coordinates is necessary to analyze com-
posite laminates. The global, or off-axis,
coordinate frame is defined by 1 and 2, also m4 n4 2m 2n 2 4m Zn 2
referred to as the laminate axes. The 1 direc-
m4

l~:)
tion is along the principal orientation of the n4 2m 2 n Z 4m 2n 2
laminate; the 2 direction is perpendicular to it. m 2n 2 mZn z m4+n4 -4m zn z
The subscript 6 is used with expressions for m2nZ m Zn 2 _2mznz (mZ _nZ)2
shear, and is a contraction for the subscript 12. m 3n -mn 3 mn 3 -m 3 n 2(mn 3 -m 3 n) Qss
The loads on the laminate and the boundary
conditions are usually defined in the global mn 3 -m 3 n m 3 n-mn 3 2(m 3 n-mn 3 )
system. In the case of a O-degree unidirectional (30.5)
690 Laminate design

where m = cos (), n = sin (). A summary of some with the [Q] matrix, the 512 (and 521 ) terms
other useful transformation relations is given reflect the deformation in the direction per-
in Section 30.7. It is clear from these relations pendicular to the direction of loading, and are
that when () = 0, then Q16 = Q26 = 0, which commonly referred to as the 'component due
means shear and extension are uncoupled, i.e. to Poisson's effect' (implying the major
shear loading only causes shear deformation, Poisson's ratio). The 516 and 526 (and 561 and
and extensional loading causes extensional 562 ) terms reflect the amount of shear deforma-
deformation with Poisson's effect but no shear tion under extensional loading, and are
deformation. commonly referred to as 'coupling terms'.
Using the off-axis plane stress stiffness coef- Unlike the off-axis unidirectional ply shown in
ficients, the constitutive relations of eqns (30.4) Fig. 30.4(b), the O-degree unidirectional ply
and (30.3) can be generalized to a ply of any shown in Fig. 30.4(a) does not exhibit any
orientation: shear deformation under extensional loading.

30.2.3 KINEMATICS
(30.6) Kinematics is the study of movement and
depends solely on geometry, not on material
properties. Since composite laminates are
and inversely often thin two-dimensional structures, plate
theory is used to simplify the three-dimen-
sional behavior. Plate theory tries to account
for stretching and bending behavior relative to
the midplane of the laminate.
The key assumption of plate theory is that
normals remain normal, straight and
For example, a physical interpretation is unstretched. In practical terms, the plies in the
shown in Fig. 30.4. In general, the terms on the laminate are assumed to be completely
diagonal (5 u ' 522, 566) reflect the amount of bonded to each other, allowing no interlami-
deformation in the direction of loading. As nar shear. Some other assumptions are that the

xLI
y,2 2
(a)
Lx,l (b) y

-{t--- au

---
__ ~~
...
r~'
.-\~~-~
0".
0'1

Sxx Sxy 0 crx S11 S12y S16 cr 1

Sxy Syy 0 0 = S12 S22 S26 0
o 0 Sss 0 S16 S26 S66 0
Fi.g. 30.4 Exte~ionalloading of (a) a unidirectional ply and (b) an off-axis ply and their associated off-axis
stiffness matriX.
 Laminated plate theory 691

material exhibits perfectly linear elastic behav- The total mechanical strains are rewritten as a
ior and that there is a perfect bond between function of the in-plane and flexural compo-
the fiber and matrix. The out-of-plane dis- nents of strain
placement, w, can be described by a function
of the in-plane coordinates such that
(30.11)
W =Wo (x,y) (30.7)
Based on the Kirchhoff assumptions, Fig. 30.5
shows the deformation of a cross section of the where eO is the strain of the midplane, z is the
plate in the x-z plane, relative to the x-direc- distance from the midplane, and K is the lami-
tion nate curvature.

(30.8)
30.2.4 STRESS RESULTANTS
Similarly, the displacement along the y-axis is Just as beam theory defines net tensile force,
awO shear force, and moment, relating everything
v = vo - z (Jy (30.9) to the neutral axis, plate theory defines stress
resultants and moment resultants to eliminate
Based on the definition of strains, any z-direction dependence and to relate
au auo ()2wo everything to the midplane, as shown in Fig.
e1 = ax = ax - Z ax2 (30.10) 30.6. Ply stresses along each loading direction
are summed for the laminate:
av avo ()2wo
e2 -- ay = ay -z (Jy2
(30.12)

~------,-~ x-----+----~

Undeformed Deformed
Cross Section Cross Section
Fig. 30.6 Force and moment resultants acting on a
Fig. 30.5 Extension and bending deformation. plate.
692 Laminate design

These resultants can be rewritten in terms of An indicates the relationship between longi-
strain by substituting in the constitutive rela- tudinal in-plane load, N 1, and the
tions. Putting the two-part expression for longitudinal extension, et;
strains, eqn (30.11), into the constitutive rela- A12 indicates the coupling between longitu-
tions, eqn (30.6), and substituting the resulting dinal in-plane load, N 1, and the extension
stress expression into the definitions for the in the transverse direction, e;, (the tradi-
resultants, eqn (30.12), tional Poisson's effect);
A 16 indicates the coupling between longitudi-
{N} = [A] {eO} + [B]{K} (30.13)
nal in-plane load, Ny and the in-plane
shear, eZ;
{M) = [B] {eO} + [D]{K)
A66 indicates the relationship between in-
plane shear load, N 6, and the in-plane
N1 All Al2 A 16 Bll B12 B16 eO1
shear distortion, e:;
N2 A21 A22 A 26 B21 B22 B26 eO2 B26 indicates the coupling between transverse
N6 A61 A62 A66 B61 B62 B66 eO6 in-plane load, N 2, and the twist, K 6;
= B66 indicates the relationship between in-
M1 B11 B12 B16 Dll °12 D16 K1
plane shear load, N 6, and the twist, K 6;
M2 B21 B22 B26 °21 D22 D26 K2 D12 indicates the coupling between longitudi-
M6 B61 B62 B66 D61 °62 D66 K6 nal bending load, MI , and the transverse
bending curvature, K2;
(30.14) D66 indicates the relationship between twisting
moment load, M6, and the twist angle, K 6•
where
The ply stacking sequence has no effect on the
[A] = f[Q]dz (30.15) A matrix coefficients, which reflect in-plane
behavior. However, since the B and D matrix
coefficients are a function of z, they are depen-
[B] = J[Q]zdz dent on the stackIDg sequence.
There are two unique physical situations
that deserve mention. When the laminate is
[D] = f[Q]z 2dz symmetric about its midplane, the B coeffi-
cients are zero, which means that there is no
It can be shown that A12 == A21' B12 = B21 , D12= coupling between in-plane loads and curva-
D21' etc. Equation (30.14) represents the funda- tures, nor between bending loads and
mental relationships in laminated plate theory. in-plane deformations. Another common sit-
The 6 x 6 matrix is the laminate stiffness uation is when the A 16 and A 26 coefficients are
matrix. zero (usually in the presence also of all the B
A composite with unidirectional plies lami- coefficients being zero): this arises when a
nated in different directions (a generally laminate is balanced, i.e. there are an equal
anisotropic material) under an inplane load number of off-axis plies in the +0 and -8
may stretch, bend and twist, as a result of directions and they have equal thickness. In
extensional! shear coupling. By comparison, a this case, there is no coupling between exten-
metal structure will stretch only under an sion loads and shear strain. If, in addition,
inplane load, bend only under flexure, and those +0 and -8 plies are effectively the same
twist only under torque. Each matrix coeffi- distance from the midplane, then the corre-
cient in eqn (30.14) relates a particular sponding Band D matrix coefficients tend
resultant to a strain expression. For example, toward zero.
 Laminated plate theory 693

Most laminates used today are symmetric a flexural contribution, then it must be added
to eliminate or reduce any tendency of the
structure to warp unexpectedly. Most lami-
nates are also balanced, often because it is (30.18)
erroneously thought to be necessary to pre-
vent the structure from warping. A balanced
laminate is really only necessary in situations
with reversible shear loading conditions. The off-axis ply strains can be transformed to
on-axis ply strains for each ply and their signif-
icance can be evaluated per a failure criterion
30.2.5 RESULTING STRAIN STATE (refer to the relations given in Section 30.7).
Knowing the laminate stiffness matrix and the
applied loads, the resulting strains can be
computed. The strains are obtained by invert-
ing the stiffness matrix and multiplying by the
input load. Instead of inverting the 6 x 6 stiff-
ness matrix, however, it is sometimes possible where, as before, m = cos e, n = sin e. One step
to simplify the analysis even further. If the further, the on-axis ply stresses can be obtained
laminate is symmetric about the midplane so by multiplying the on-axis ply strains by the
that the B coefficients are identically zero, then ply stiffness matrix [Q] as shown in eqn (30.6).
the in-plane (described by N, [A], e) and bend- The laminate engineering constants, which
ing problems (described by M, [D], /C) become have meaning with symmetric laminates only,
uncoupled. In this case, it is much easier to are calculated from the compliance matrix and
invert two 3 x 3 stiffness matrices to get the are useful for comparison to the properties of
compliance matrices (see also Section 30.7 for other materials, such as metals
the explicit terms to invert a matrix)
111
[a] = [A]-l EO = - EO = EO -
1 a*'
11
2 7'
22
6 - a*66
(30.20)
[d] = [D]-l (30.16)
111
Ef - -
Then, the compliance matrix is multiplied 1 - d* , Ef2 -- -d* , Ef6 -- d*
11 D 66
by the appropriate input load conditions to
compute the laminate strains where the compliance terms have been nor-
malized to have the necessary units of
[length2 / force]
(30.17) [a*] = [a]h (30.21)

[d*] = [d]h 3 /12

In summary, the mathematical process of ana-
lyzing composite laminates is indicated in Fig.
30.7. A laminate's stiffness is calculated as the
summation of its individual ply properties.
Load on the laminate is described in terms of
Since all the plies are bonded together, the the laminate coordinates. Calculated nom the
strains in each ply, in the laminate axes, (1,2), applied load and the known material stiffness
are equivalent to the laminate strain. If there is properties, the response of the laminate is
694 Laminate design

expressed as laminate strain. In order to apply 30.3.2 HYGROTHERMAL EFFECTS

a failure criterion, the laminate strain is com- Most structure is exposed to a variety of envi-
monly transformed into ply strains and each ronmental effects. Of particular concern are
ply is individually evaluated. heat and moisture. The design must account
for the hygrothermal effects (hygrothermal
30.3 ENHANCEMENTS TO THE BASIC means water and temperature). A laminate
LAMINATED PLATE EQUATIONS that is stress-free when curing at an elevated
temperature will have residual stresses when
30.3.1 SANDWICH CONSTRUCTION brought back to room temperature. It has been
thought that some of the apparent improve-
In composite structures, sandwich construc- ment in toughness of 250°F resins over 350°F
tions are commonly used. By increasing the resins is simply due to the reduction in resid-
distance between the load-carrying laminate
ual stresses (~T =75 - 250 =-175°F versus ~T
skins, a core can provide increased bending = 75 - 350 = -275°F). In addition, moisture is
stiffness without a significant weight penalty.
absorbed by the laminate, usually into the
The core is often idealized in laminate design: resin. The negative effect on the mechanical
it is assumed that the core does not contribute
properties is particularly pronounced at both
to laminate strength or in-plane stiffness, and high temperature and high humidity.
that the shear bonds between the skins and
Assuming that the primary effect on resid-
core are perfect. The parallel axis theorem can ual stresses is due to different thermal
be used to account for the increased moment expansions and moisture expansions along
of inertia that the core creates by offsetting the and transverse to the fiber direction, two addi-
laminate load-carrying skins from the mid- tional strains on the laminate result:
plane. Once the skin laminates have been
sized, further calculations can be performed to {e} = {a}~T+ {Il}&: (30.22)
confirm that the core assumptions are valid3 •

Measured Plane Laminate Inverted Given Loads;

Ply Stress Stiffness Stiffness Resulting Ply
Stiffnesses Coeff. Matrix Matrix Laminate Strains
Strains
Fig. 30.7 Logical flow of calculations involved in analyzing a symmetric composite laminate that is loaded
axially and/or in bending.
 Failure criteria 695

The summation process can be used to deter- X' = transverse tensile strength,
mine the effective laminate expansion Y = longitudinal compressive strength,
coefficients. Y' = transverse compressive strength,
The hygrothermal load, sometimes called S = shear strength. (30.24)
non-mechanical load, can be computed by
The maximum ply strain values can be inter-
multiplying the laminate stiffness by the
preted by dividing the above strengths by the
hygrothermal strain (i.e. laminate thermal
appropriate ply stiffness coefficient.
expansion coefficients multiplied by the
change in temperature). The stress induced by Ex * = max longitudinal tensile strain,
moisture absorption can be accounted for sim- Ex'* = max longitudinal compressive strength,
ilarly by using {3Ac in place of aAT. Thus, the Ey * = max transverse tensile strength,
non-mechanical loading in the laminate can be E/* = max transverse compressive strength,
expressed as Es * = max shear strength. (30.25)
Laminate strength is function of material (ply)
strength and the constraints on the ply within
(30.23)
the laminate. Thus, failure is best assessed at
the ply level. The proper interpretation, how-
ever, of the significance of the applied stress
relative to the material strength is still
debated. Maximum stress and maximum
strain failure criteria are common wherein the
applied stress or strain value is compared
directly to the strength value. A review of fail-
ure criteria has been published4•
The mechanical and non-mechanical loads, N Early laminate failure theories fail to
and W, can be added together to determine account for Poisson's effects and interaction
the total load experienced by the laminate. between loads in orthogonal directions (a
complex load condition). For example, the
major weakness of both the maximum stress
30.4 FAILURE CRITERIA
and the maximum strain failure criteria are
The ultimate objective in any structural design their inability to couple stress, or strain, com-
is to create a structure able to withstand ponents in determining the ultimate failure of
deflections or loads without failing. The initial a ply. It is important to understand that the
concern is to remain below a prescribed deflec- longitudinal tensile failure of a ply is affected
tion as part of stiffness criteria. Once these not only by the longitudinal load, but also by
criteria are satisfied, the focus shifts to a the magnitude of applied transverse loads. As
strength criterion, such that applied stress a result, stress interaction criteria are widely
must not exceed laminate strength. used throughout the industry to determine ply
Composite materials normally possess dif- failure in a laminate.
ferent strengths when loaded in either tension
or compression. The following represent the
minimum number of strength properties nec- 30.4.1 QUADRATIC FAILURE CRITERION
essary to characterize a unidirectional or fabric
Tsai developed a two-dimensional stress inter-
ply. They are determined using material
action failure criterion and predicted the
coupon tests, previously outlined in Fig. 30.2.
strength of an orthotropic ply subjected to
X = longitudinal tensile strength, combined stresses or strains. This analysis
696 Laminate design

takes into account the effects of other stress before failure occurs);
components on the strength in anyone direc- R < 1 failure has already occurred (i.e. occurs
tion. Tsai postulated a criterion in stress space prematurely at some point below the
consisting of the sum of linear and quadratic applied stress or strain) and the applied
scalar products as follows: stress or strain level can not be attained
(e.g. if R = 0.5, then only half the
F;p,oj + F,o; ~ 1 i, j = X, y, s (30.26) applied stress can be sustained).
or, in expanded form,
Equation (30.28) is substituted into eqn (30.26)
(F a +2Fxyxy
xxx
2 aa +Fyyy
a +FSSg
a 2 2) and the solution of this quadratic equation can
be obtained.
+(Fa
x x +Fa)~l
y y

Note that Fxs = Fys = F. = 0, and the six strength

parameters are interpreted from the ply i, j = X, y, s
strength values (reviewed in Section 30.8):
The positive and negative roots of the qua-
1 1
dratic equation can be found and represent
Fxx = XX' Fyy = YY'
failure of the laminate in tension and compres-
sion (where the absolute value of the negative
1
F. = 52 Fxy = Fxy*--/(F~yy
P } root is used), respectively.
Failure envelopes can be plotted to show
laminate strength for any combination of
1 1 1 1
Fx = X - X' Fy = Y- Y' (30.27) loads. Instead of the stress space representa-
tion, however, the examination of failure
The stress interaction term, F *, can have a envelopes in strain space is a useful alterna-
tive. The representation of failure envelopes in
value of -1 ~ F:: ~ 1 although is1ecommended
to be -1/2. When F; = -1/2, the quadratic fail- strain space is preferred because strain is usu-
ure criterion is a general case of the von Mises ally specified in laminated plate theory. Strain,
criterion (Section 30.8.2). unlike stress, is at most a linear function of the
Instead of simply evaluating the failure cri- thickness. Furthermore, failure envelopes are
terion to determine if the laminate failed, it is fixed in strain space, and are independent of
useful to consider a nondimensional ratio to other plies with different angles which may
exist in a laminate. Thus, they can be regarded
provide a perspective of the significance of the
as material properties. Another additional
applied stress relative to material strength.
Tsai defines the strength ratio, R, such that advantage of strain space is that the axes are
dimensionless.
{amax } = R {aapplied}
30.4.2 STRENGTH OF LAMINATES
(30.28)
Traditional failure criteria based on strength of
The strength ratio is always a positive number materials are limited to the prediction of the
with the following physical implications: FPF, the point beyond which the continuous
R = 1 failure occurs (at the applied stress or and homogeneous material assumptions are
strain level); no longer valid. The use of a simple method
R > 1 failure has not occurred and R repre- for modeling of degraded plies is recom-
sents a factor of safety (e.g. if R = 2, then mended, from which the FPF can be estimated.
the applied stress can be safely doubled The load-carrying capability of a laminate
 Laminate design 697

beyond the FPF can be formulated using a ply of the preselected orientations results in a
degradation model. Two possible methods are quasi-isotropic laminate. This is the perfor-
recommended: first, the simplified microme- mance baseline, because load-carrying fiber is
chanics model based on the modified in effectively all directions. Laminate perfor-
rule-of-mixtures relations can be used. Plies mance can only be improved beyond that of a
with transverse cracks are replaced by plies quasi-isotropic laminate as fiber is biased into
with reduced matrix modulus, Em. load directions, since, of course, fiber would
Micromechanics translates the effect of the never be put in unnecessary directions.
altered constituent material properties to the Heretofore, quasi-isotropic laminates have
ply level, e.g. how a change in the matrix mod- been used because they give properties like
ulus affects the shear and transverse modulus those of metals, and predictable responses that
of the unidirectional ply. Degraded plies are are familiar, although they are not optimal in
modeled by quasi-homogeneous plies so that strength-to-weight or stiffness-to-weight
laminated plate theory can be reapplied to ratios. Many laminates used today on aircraft
determine the ply stresses and ply strains. structures tend to be of this type. In general,
Another approach for the prediction of however, the more directional the loading, the
post-FPF strength can be based on macrome- bigger the payoff possible with anisotropic tai-
chanics, without resorting to micromechanics. loring.
The degradation factor (DF) is applied directly To improve on the performance obtained
to the transverse and shear modulus, as well with a quasi-isotropic laminate, the cost to
as the major Poisson's ratio. The exact value design and analyze the anisotropic part (using
for the degradation factor must be determined the tools like those discussed in this chapter) is
empirically. A value between 0.1 and 0.3 is rec- unfortunately often thought not to be worth
ommended. If the degradation factor is given the additional weight savings. This attitude is
a value close to zero, the quadratic failure cri- commonly rationalized by worry about holes,
terion can be made to resemble the maximum increase in work associated with more compli-
strain criterion and results in a generally con- cated fiber placement (preform assembly), etc.
servative estimation of laminate strength. In practice, laminate designs, if not quasi-
isotropic, are certainly still symmetric about
30.5 LAMINATE DESIGN
the midplane, balanced (equal quantity of -8
and +0 plies), and orthotropic. Capitalizing on
To simplify the analysis, it is commonly initially the benefits of anisotropy will probably occur
specified that a laminate will be constructed of in other industries first before being adopted
plies oriented with fibers in a few preselected by the more conservative aircraft industry.
directions, where only the percentage distribu- An exception to traditional aircraft laminate
tion in each orientation must then be design is the X-29 experimental aircraft, which
determined. Laminates with plies distributed demonstrated a unique attribute of anisotropy
every 45° are called 1tI4laminates (plies can be (Fig. 30.8). The basis for this design lies in the
in the 0, 45, 90 or -45 directions. Ply orienta- important assumption that the 1,2,6 axes are
tions are usually specified as a value between usually the primary load directions for the
-90 and 90°. For example, instead of identifying laminate. With the coordinate system for load-
the orientation as 135, the laminate orientation ing changed to be 20° off a designated
is more commonly called -45°, although they laminate system, it can be shown that the lam-
are the same). Another class of laminates are inate behavior in flexure and torsion is
called n/3, where plies are placed every 60° coupled. In fact, twisting will result with flex-
(plies can be in the 0, 60 or -60 directions). In ural loading, even though the material would
both cases, an equal percentage of plies in each normally behave as most metals. This is the
698 Laminate design

principle used on the X-29 6. The normal ten- the laminate. Composite materials are not
dency for forward swept wings to diverge at merely a light-weight substitute for heavy-
increasing speeds was counteracted by this weight metals. Structural performances which
laminate design: the increase in lift creates a are not possible with metals are easily achiev-
decrease in angle of attack, as the laminate able. Examples of such unique properties
twists in the direction opposing the forces. include Poisson's ratios greater than unity or
H is conceivable that in the future the even negative, bending-twisting coupling, and
graphite golf shafts currently gaining in popu- zero or negative coefficients of thermal expan-
larity could be tailored to the individual golfer. sion (CrE). The problems and examples below
The same coupling principle could be applied. illustrate the engineering constants of angle-
A golfer's tendency to consistently slice the ply and related laminates. Examples of large
ball might allow the designer to customize a and negative Poisson's ratios and examples of
golf shaft which not only bends, but also bend-twist coupling are also given.
twists slightly under the bending load of the
bad swing.
30.5.2 UNUSUAL POISSON'S RATIOS
Personal computer software based on a com-
30.5.1 UNIQUE BEHAVIOR
puter spreadsheet allows rapid sensitivity
The most unique features of composite materi- studies and parametric analysis of the behav-
als are the highly direction-dependent ior of laminates. Laminated plate theory with
properties. Highly coupled deformation and micromechanics is programmed into 'Mic-
load-carrying capability can be designed into Mac/In-Plane'2. A companion charting toot

Fig. 30.8 Top view of the Grumman X-29 aircraft with wings that twist under flexure to counteract the
detrimental aerodynamic effects.(© NASA)
 Laminate design 699

'Chart-quick', can be used to plot variation of Table 30.1 Material property data for three differ-
CTE as a function of independent variables (0, ent carbon fiber systems: IM6/Epoxy, T300/5208
Ell' E22, vf ' etc.). For the following problems and M40J /F584
and examples, the carbon fiber reinforced
IM6/ T300/ M4DJ/
polymer material data used are shown in 5208 F854
Epoxy
Table 30.1.
Figure 30.9(a) shows the engineering con- Longitudinal tensile
stants for a unidirectional laminate as it is modulus, Ex (Msi) 29.44 26.27 32.8
rotated from the on-axis. The Poisson's ratio, Vx' Transverse modulus,
of a 0° laminate is approximately 0.3. With Ey (Msi) 1.62 1.49 1.2
increasing angle of the off-axis laminate, the Poisson's ratio 0.32 0.28 0.26
Poisson's ratio decreases. The Poisson's ratio of
a 90° laminate is effectively zero, because con- Shear modulus, Es (Msi) 1.22 1.04 0.66
traction in the transverse direction is Longitudinal CTE, a 1 -0.14
constrained by the fibers. Transverse CTE, a 2 15
Figure 30.9(b) shows the engineering con-
stants for an angle-ply laminate. It is interesting Volume fraction V f(%) 66 70 62
to observe the very large Poisson's ratio of 1.32
for a [±30] laminate. A value of greater than one
implies that the transverse dimensional change image' of the longitudinal modulus, Ex'
is more than in the dimensional change in the Figure 30.10(a) shows the engineering con-
longitudinal direction of loading. stants for cross-ply laminates. For any given
When the ply angle is either 0 or 90°, the laminate, the longitudinal modulus, Ex' and
laminates (and consequently the values for the the transverse modulus, E, are equal. The
engineering constants) in Figs. 30.9(a) and Poisson's ratio, v x' of a [0/90] laminate is
30.9(b) are the same. In both Figs. 30.9(a) and approximately zero, because of the presence of
30.9(b), the transverse modulus, Ey is a 'mirror fibers in the transverse direction. The largest

Modulus Poisson's
Modulus
(Msi) [e]as Poisson's
Ratio (Mai) [± e]4s Ratio
30.00 2 30.00 2
-0- Ex
22.50 1.5 22.50 -<>- Ey 1.5

15.00 15.00

7.50 0.5 7.50 0.5

0.00 0 0.00 0
15 30 45 60 75 15 30 45 60 75
-0.5 -0.5
Ply Angle, e (degrees) Ply Angle, e (degrees)
(a) (b)

Fig.30.9 Engineering constants of IM6/ epoxy laminates as a function of (J for (a) off-axis unidirectional [(J]s.i
and (b) angle-ply [±O]4s'
700 Laminate design

[8, (8 + 90}]4S [04, ± 82]s

Nodulus Poisson's Modulus Poisson's
(Msi) Ratio ( Msi) Ratio
30.00 2 30.00 <J- Ex 2
<J- Ex -o-Ey
22.50 -o-Ey 1.5 22.50 -0- Es 1.5
-0- Es
15.00 15.00

7.50 0.5 7.50 0.5

0.00 0 0.00 0
15 30 45 60
Ply Angle, 0 (degrees) -0.5 Ply Angle, 0 (degrees) -0.5
(a) (b)

Fig. 30.10 Engineering constants of IM6/epoxy laminates as a function of 0 for (a) cross-ply [0,(0 + 90)]4.0;
and (b) [04, ± 021..

Poisson's ratio is 0.55 for a [±45] laminate. The laminate exhibits a very large Poisson's ratio
shear modulus, E , is a maximum, of course, of 1.32, when compared with that of an
•
for the [±45] laminate. isotropic material (0.3). Besides the unique
Figure 30.1O(b) shows the engineering con- Poisson's ratio behavior, it is also important to
stants for laminates with 50% 0° plies and 50% examine the values of the other coupling coef-
angle-plies. With the exception of the trans- ficients.
verse modulus, the results are similar to those
for the angle-ply laminate shown in Fig. EXAMPLE
30.9(b). When the ply angle is 90°, the values
for the engineering constants in Figs. 30.10(a) Table 30.2 considers the resulting deforma-
and 30.10(b) are the same. tions on coupon specimens under load, and
Figures 30.11(a) and 30. 11 (b) show the engi- Fig. 30.12 indicates the relative magnitude of
neering constants for some unusual laminates. deformation due to large and negative
When the ply angle, 0, is 15°, Fig. 30.11(a) Poisson's ratios.
shows an off-axis unidirectional laminate and
Fig. 30.11(b) shows an angle-ply. For all other 30.5.3 STIFFNESS AND COUPLING
ply angles, the laminates are unbalanced.
From Fig. 30.11(a), it can be observed that the It is useful to look at the A, B, D stiffness matri-
[15/60]. laminate exhibits an extremely large ces of some simple laminates. For ease of
negative Poisson's ratio of -0.32, meaning the comparison, the stiffness matrices can be nor-
laminate will expand in the transverse direc- malized to have units of [force/length2] by
tion under longitudinal tension loading and defining
compress in the transverse direction under [A*] = [A]!h, [B*] = 2[B]!h2,
longitudinal compressive loading. From Fig.
30.11(b), it can be observed that the [-15/30] [D*] = 12[D]/h3
•
 Laminate design 701

Modulus [15/8 ]4s Poisson's Modulus [-15/8 ]4s Poisson's

30.0b
(Msi) Ratio Msi ) Ratio
30.00 2 2
-0- Ex
-0- Ex
22.50 1.5 22.50

15.00 15.00

7.50 0.5 7.50

0.00 0 0.00

Ply Angle, e (degrees) -0.5 Ply Angle, e (degrees) -0.5

(a) (b)

Fig. 30.11 Engineering constants of IM6/epoxy laminates as a function of 8 for (a) [15/8]45; and (b)
[-15/8]45'

Table 30.2 Strains, deformations and strength ratio (based on first-ply-failure) of 10 in x I in X 0.1 in spec-
imens under 1000lb longitudinal load, N]

Longitudinal Transverse Longitudinal Transverse Strength

strain strain displacement displacement ratio
[1 [2 Xl X2 R
Material Layup (10-3 in/in) (10-3 in/in) (10- 3 in) (10- 3 in)
40ksi Steel 0.34 -0.10 3.4 -1.0 3.9
IM6/Epoxy [0]20 0.34 -0.11 3.4 -1.1 50.0
IM6/Ep [0, 90]Ss 0.65 -0.02 6.5 -0.2 7.9
IM6/Ep [+30/ -30]Ss 0.95 -1.29 9.5 -12.9 6.8
30ksi Aluminum 1.02 -0.30 10.2 -3.0 2.9
IM6/Ep [15/60]5s 1.09 0.35 10.9 3.5 3.1
IM6/Ep [+45/ -45]5, 2.40 -1.77 24.0 -17.7 2.7
E-glass/Ep [+30/-30]5S 3.20 -2.07 32.0 -20.7 4.5
IM6/Ep [90]20 6.25 -0.11 62.5 -1.1 0.8

[15/6O ls s

[±45]5s
aluminum [+30] [±30]5s(E-gl/ep)
- 5s
[0/90 ls s
[90]105
steel, [O]lOs

undeformed
Fig. 30.12 Relative deformation of 10 x 1 x 0.1 in specimens under 1000 lb load along the centerline (lami-
nates are IM6/epoxy, unless otherwise indicated).
702 Laminate design

A four-ply laminate consisting of two 0° and B* matrix with nonzero terms. The first and
two 90° plies can be combined into four differ- fourth laminates are balanced and so the A \6
ent laminates. From Table 30.3 it can be and A \6 coefficients are zero. For the second
observed that while the A * matrix remains and third laminates which differ by the sign of
unchanged through varied ply stacking the off-axis plies, the stiffness behavior differs
sequences, large differences arise in the B* and only in that the A*16' A*26' D*16 and D*26 coeffi-
D* matrices. cients are of opposite signs.
From Table 30.4 it can be observed that only Table 30.5 displays different quasi-isotropic
the fourth laminate is unsymmetric and has a laminates. Note that the normalized A * matrix

Table 30.3 Normalized stiffness coefficients for four IM6/epoxy laminates, in units of Msi

Layup [0/90/90/01 [90/0/0/901 [0/90/0/901 [0/0/90/901

15.624 0.525 15.624 0.525 0 15.624 0.525 o 15.624 0.525 o

[A *] [ 0.525 15.634 [ 0.525 15.634 0 [ 0.525 15.634 o 0.525 15.634 o
o 0 o 0 1.220 o 0 1.220 o 0 1.220

o o o o o o 6.997 a o
[B*] o
o
o
o ~ 1[ o
o
o
o
o
o
o
o
o
o
-6.997
0
o
o
26.129 0.525 5.138 0.525 o 15.634 0.525 15.634 0.525
[D*] [ 0.525 5.138 [ 0.525 26.129 o [ 0.525 15.634 0.525 15.634
o 0 o 0 1.220 o 0 a a

Table 30.4 Normalized stiffness coefficients for four IM6/epoxy laminates, in units of Msi

L11yup [0/0/+45/-451, [0/0/+45/+451, [0/0/-45/-451, [+45/-4541

3.692 19.463 3.692 3.499 [ 92.97 6.859 0

[19.~ [19.~ 3.692 ~3.499
[A"] 3.692 5.469 3.692 5.469 3.499 3.692 5.469 -3.499 6.859 9.297 0
0 0 4.Ll 3.499 3.499 4.387 -3.499 -3.499 4.387 0 a 7.555
0 a a 0 0 0 0 a 0 0 3.499

1 [3L
[B*] 0 0 0 0 0 0 0 0 0 0 3.499
0 a a 1 0 a 0 a a 0 3.499 0

27.087 1.316
o.~ 1 [27.087 1.316 0.875 [27.087 1.316 -<l.875 [ 9.2~ 6.859 0
[D*] 1.316 2.597 0.656 1.316 2.597 0.875 1.316 2.597 -0.875 6.859 9.297 0
0.656 0.656 2.012 0.875 0.875 2.012 -0.875 -0.875 2.012 0 0 7.555
 Laminate design 703

Table 30.5 Normalized stiffness coefficients for four IM6/epoxy laminates, in units of Msi

Layup [0/45/-45/90]5 [0/45/90/-45]5 [10/55/100/-35]5 [0/60/-60]5

['2.466 3.692 12.466 3.692 ['2.466 3.692 ['2.466 3.692

[A"] 3.692 12.466 3.692 12.466 3.692 12.466 3.692 12.466
o 0 .l71 0 0 .l71 o 0 4.L1 o 0 4,L1

[20.932 3.098 1.312 [20.2115 2504 968 ['8.,94 2.782 3.735 [22001
1.932 787
1
1.932 6.451 0.
1.956 1
,.
[D"] 3.098 5.188 1.312 2.504 7.094 1.968 2.782 8.558 2.208
1.312 1.312 3.793 1.968 1.968 3.200 3.735 2.208 3.477 0.737 1.956 2.628

terms are equivalent for all quasi-isotropic Multiplying out the strains, 8 1 =anN 1, 82 =
laminates. This means all have the same stiff- a12N2, 8 6 = a16N1
ness to weight ratio. (iii) Evaluate the displacements by integrating
The differences between these laminates the strains. (Note that for the tube,
thus manifests themselves only in how they
behave in bending. d<P
86 = r dx.)
PROBLEM
Find the amount that an anisotropic 20-layer
[0/30]5s T300/5208, 3 in diameter, 12 in tube

I
tube will extend, change in circumference, and
twist under an in-plane load, Nl = 100Ib/in.
E,dy ~ r :~dC ~ r·,,N,dC ~ v ~ ',,N,21tR
SOLUTION
(i) Compute the laminate stiffness matrix and
invert to get the compliance matrix:

21.159 2.567 3.935 1 (iv) Evaluate the displacements numerically:

[A*] = 12.567 2.476 1.458, U = a*n N L = 62.563 x 100 x 12
3.935 1.458 3.195 h 1 0.1
= 0.75 x 10-3 in
62.563 -26.647 -64.8931
[a*] = [ -26.647 563.836 -224.518
v = a*12 N 27tR = 26.647 x 100 x 2 x 3.14 x 1.5
-64.893 -224.518495.374 h 1 0.1

(ii) Evaluate the strains: = -0.25 x 10-3 in

<P = a\6 N 1::. = 64.893 x 100 x 12/1.5

h 1R 0.1
= -0.52 x 10-3 radian = -0.03 degree
704 Laminate design

30.5.4 CTE BEHAVIOR the a 1 is -2 to -2.5 (versus -0.14 for a 1 of the

The following four figures show the coefficient unidirectional tape, as shown in Fig. 30.13(a».
of thermal expansion (CTE) in two principal As the number of ply angles increases, the
directions (referred to as a 1 and a 2 ) for CTE behavior becomes less intuitive. Figure
M40J /F584 carbon fiber laminates. Figure 30.14(a) shows the CTE of a laminate with 50%
30.13(a) shows the CTE of an off-axis unidirec- oplies and 50% angle-ply; Fig. 30.13(b) shows
tional ply; Fig. 30.13(b) shows the CTE of an the CTE of a laminate with 25% 0 plies, 25% 90
angle-ply laminate. From Fig. 30.13(b), it can plies and 50% angle-ply. Note that when ±1}
be observed that, due to the Poisson coupling equals 45°, the resulting laminate is quasi-
effect, laminate CTE values less than that of a isotropic [0, 90, ±45] as confirmed by a 1
unidirectional material are possible for specific equaling a 2• Examination of the fundamental
ply angles. In Fig. 30.13(b), for () of ±20 to ±40°, trends in Figs. 30.13 and 30.14 indicates

....
:ac:
16.00 16.00

1400
[68] 1....00
0
'iii 12.00 12.00
c:
'IS
C. 10.00 10.00
!Ii &00 &00
iU
E
.... 6.00 6.00
Q)
;; ....00 ....00
'0
E(I) 2.00 2.00

'0 0.00 0.00

:e
Q)
0 -2.00 -2.00
U Ply Angle, 0 (degrees)
4.00 Ply Anile, 8 (degrees)
(a) (b)

Fig. 30.13 Coefficient of thermal expansion of M40J IF584 laminates as a function of () for (a) off-axis undi-
rectional [(}s]'; and (b) angle-ply [±64 ]s'

.... 16.00

Ic:
16.00

1400 [04, ±6 2] 1.... 00 [02,902,±6 2]

0 12.00
"iii 12.00
c:
'IS 10.00
C. 10.00
!Ii
~
&00
&00
iU a2
E (I)
6.00
.go
6.00

;; 400
400
'0
.--------
'-
2.00
EQ) 2.00
"0 0.00
:e Q)
0.00
10 40 so 60 70 80 90 -2.00
0 -2.00
U Ply Angle, 0 (degrees) Ply Angle, 8 (degrees)
(a) (b)

Fig. 30.14 Coefficient of thermal expansion M40J/F584 laminates as a function of () in the following lami-
nates (a)[04' ±62].; and (b) [02,902' ±62]•.
 Laminate design 705

potential near-zero CTE laminates, particu- dominantly in the longitudinal direction to

larly useful in spacecraft applications to accommodate flexural loading (like a mast or
minimize deformation due to the large cyclic golf club), Fig. 30.14(a) indicates that a [04,
thermal loading. t
±302 with a steel mandrel would be a prob-
lematic choice, resulting in a composite shaft
locked on to the mandrel as shown in Fig.
EXAMPLE
30.1S(b). There is a need for sufficient fibers in
To remove a composite shaft from a metal the hoop direction (90°) to result in a laminate
mandrel after elevated temperature cure, the CTE less than that of the mandrel material.
laminate CTE in the hoop direction of the The CTE of the metal materials given in Table
cylindrical section has to be less than that of 30.6 indicates that it is easier to remove a com-
the mandrel material to prevent lock-on. The posite shaft from an aluminum mandrel than
composite is considered to be stress-free at from a steel mandrel.
cure-temperature, and thus the temperature
loading is associated with the temperature
decrease to room-temperature. Table 30.6 Average coefficients of linear thermal
AT = Toperaling - Tstressfree = Troom - Tcure < 0 expansion of selected materials

(30.30)
Figure 30.1S(a) illustrates that it is preferable Aluminum alloy 12.8
to have the metal mandrel contract more than Concrete 6.7
the composite during cool-down, which Invar 0.39
Steel 6.5
means that the metal CTE must be more than
Titanium alloy 4.9
the composite CTE. For a shaft with fibers pre-

a composite < a mandrel

(1lgnll1clnt mount of I1bers
a composite > a mandrel
(predomlnlntly longitudlnll fibers,
In hoop dlrectJon)

~T>O
(dl
Q
Fig. 30.15 Result of different coefficients of thermal expansion (CTE), a., for metal and composite cylindri-
cal sections under two thermal load cases.
706 Laminate design

EXAMPLE freedom to design the laminated material is at

the price of more complicated analysis. This
Carbon fiber room temperature curing materi-
chapter presented an explanation of the basic
als are often used for composite bicycle
mathematical tools used to design laminates
frames. To prevent a metal-composite joint
including laminated plate theory and the qua-
from coming apart under in-service thermal
dratic failure criterion. Some examples were
loading, the metal lug is generally on the
provided to lend insight on the many options
inside. Ideall~ under an in-service tempera-
for optimizing the material to particular needs.
ture increase due to sun shining on a part
(which can be up to 82°C(180°F», or friction,
tile lug will expand more than the composites.
REFERENCES
This 'seizing' action will prevent the lug from
'breaking loose. If there are mostly longitudinal 1. Harper, R. The Fifth Halford Memorial Lecture,
fibers due to flexural design considerations, ,. Roy. Aeronaut. Soc., Apr 1966, 70,477-486.
the CTE in the hoop direction may be greater 2. Tsai, SOW. 1990 Composites Design, 4th Edn.,
Think CompOSites, Dayton, OH, 1988.
than that of the metal; the metal fitting should 3. MIL-HDBK-23A, Military Handbook
be then on the outside. More commonly, how- Structural Sandwich Composites, Department
ever, the composite has more fiber wound of Defense, Washington., 1968.
around the lug (in the hoop direction) and the 4. Quinn B.J. and Sun, c.r. A critical evaluation of
CTE is less than that of the metal. failure analysis methods for composite lami-
nates, Proceedings of the Ninth
These are only two of the challenges in cop- DoD/NASA/FAA Conference on Fibrous
ing with different thermal expansion Composites in Structural Design, Lake Tahoe,
coefficients of dissimilar materials. In general, Nevada, Nov 4-7,1991, pp. V21-V37.
because of the low CTE of the fiber, multidirec- 5. Tsai, S.W. and Hahn, H.T. Introduction to
tionallaminates typically have a laminate CTE Composite Materials. Technomic Publishing Co.,
1980.
less than metals. This can cause difficulties in 6. Hadcock, R.N. X-29 Composite Wing, AIAA
part manufacture, in cases of elevated cure. It is Evaluation of Aircraft/ Aerospace Structure and
generally desirable to match the CTE of a com- Materials Symposium, Air Force Museum,
posite part with the tooling to minimize Wright-PattAFB, Apr 24-25, 1985.
springback and residual stresses. A tool with 7. Popov, E.P. Mechanics of Materials, 2nd Edn.,
zero CTE results in a dimensionally more Prentice-Hall, Inc., 1976, p. 295.
accurate part; a tool with eqUivalent CTE
results in no chanee of lock-on. Composite tool-
ing systems may be used in order to match OTHER READING
CTE; however, such tools uSually do not have MIL~J:IDBK-17-2C, Military Handbook - Polymer
the necessary durability over long production Matrix Composites, Vol. II: Material Properties,
runs. Steel tooling is often preferred over alu- Department of Defense, Washington D.C., Feb
minum tooling because of the lesser CTE. 25,1994.
MIL-HDBK-17-3D, Military Handbook - Polymer
Matrix Composites, Vol. 1lI: Utilization of Data,
30.6 CONCLUSIONS Department of Defense, Washington DC, Feb
28,1992.
Composite materials have many well-known Hyer, M.W. Cal<:ulations of the Room-Temperature
advantages over other . structural materials Shapes of Unsymmetric Laminates, f. Composite
including increased strength-ta-weight ratio, Milteria}s, 1981, 15, 296-310.
stiffness-ta-weight ratio, increased fatigue life, Hyer, M.W. Some observations on the cured shape
of thin unsymmetric laminates, ,. Composite
corrosion resistance, aI)d the ability to tailor the
Materials, 1981, 15,175-194.
properties of the material. The extraordinary
 Appendix B 707

Garfinkle, M. Twisting Smartly in the Wind,

Aerospace America, American Institute of m2 n2 -mn
Aeronautics and Astronautics, Reston, VA, July
1994, 18-20.
[f] = n2 m 2 mn (30.AS)
Smith, E.C. Vibration and Flutter of Stiff-Inplane 2mn -2mn m2 - n2
Elastically Tailored Composite Rotor Blades,
34th AIAA/ ASME/ ASCE/ AHS/ ASC
Structures, Structural Dynamics and Materials
Conferece, La Jolla, CA, AIAA-93-1302-Cp, Apr
19-21, 1993, pgs. 26-37. (30.A9)

NOTE
To invert a 3 x 3 stiffness matrix [A] to obtain
MIL-HDBK-23 is obsolete and no longer available. the compliance matrix, [a] = [A]-l
The data is being included in MIL-HDBK-17, which
can be obtained from OODSSP, BLDG 4D, 700
Robbins Ave., Philadelphia, PA 19111-5094, or (30.AlO)
http://www.dodssp.daps.mil.

30.7 APPENDIX A: TRANSFORMATION

RELATIONS

There are several useful transformation rela-

tions (prime refers to the off-axis coordinates
1, 2, 6 and no prime refers to the on-axis coor- where the determinant is IAI =(AllA.22 - Ai2)A66
dinates x, y, s as in Fig. 30.3). + 2AI :zA26A16 - AllA~ - A~i6' The flexural stiff-
ness matrix [D] can be inverted similarly to
{a'} = U]{a} (30.A1)
obtain the compliance matrix [d].
{a} = m-l{a'} (30.A2)
30.8 APPENDIX B: MORE DETAILS ON THE
{E} = [f]{E'} (30.A3) QUADRATIC FAILURE CRlTERION5

{E'} = [f)-I{E} (30.A4) 30.8.1 DETERMINING THE COEFFICIENTS

A useful exercise is to review how the F coeffi-
{Q'} = m{Q}[f] (3D.AS)
cients in the quadratic failure criterion
where m = cos 0, n = sin O.
Fxxa; + 2,Fx,pPy + Fyy~ + F~sa;+ fpx + F,py =1
(30.B1)
m= [ m'2
n
n' 2mn
m2 -2mn
1 (30.A6) are mathematically determined from the ply
-mn mn m2 -n2 strength properties. During a uniaxial tension
test of a 0 degree unidirectional laminate along
the longitudinal axis (pulling in the direction
of the fibers); a = X and d = a = O.
m- l = [~n n' -2mn
m2 2mn
2
1 (30.A7) Substituting into :qn (30.B1), theYfailur~ crite-
rion reduces to
mn -mnm2 -n2 .
(30.B2)
708 Laminate design

During a uniaxial compression test of a 0 familiar from isotropic materials. The most
degree unidirectional laminate, (Jr = X' and (J common form of the von Mises criterion is
=(Js = 0 and the failure criterion in eqn (30.Bt) probably the equation written in terms of the
reduces to principal stresses in three dimensions7•
FrrX'2-Fr X' =1 (30.B3) «(Jl - (J2)2 + «(J2 - (J3)2 + «(J3 - (Jl)2 =2cfyield
Solving the two equations given by eqns
(30.B7)
(30.B2) and (30.B3) for two unknowns, Frr and
Fr' yields For plane stress (J3 = 0, eqn (30.B7) becomes
1 (Jl2 - (Jl(J2 + (Jl2 = (J2yield (30.B8)
Frr = XX'
Instead of computing the principal stresses,
(30.B4)
1 1 the von Mises criterion in eqn (30.B8) can also
Fr = X - X' be written in terms of a general stress state that
includes a shear stress contribution
Similarly for the transverse direction, a 90
degree unidirectional laminate can be experi-
~- (JPy +~ + 3q; = tr yield (30.B9)
mentally evaulated under uniaxial tension «(J Now, for comparison, the quadratic failure cri-
=¥, (Jr = (Js =0) and compression «(Jy = ¥', (Jr ;, terion is repeated here
(Js = 0).
Substituting each test situation into the qua- Frr~ + 2Fry(Jpy + Fyy(J~ + Fssa: + Fpr + Fpy =1
dratic failure criterion and solving the two
(30.BI0)
equations, the resulting two unknown coeffi-
cients, Fyy and Fy' would be where
1 1 1
Fyy = W' (30.BS) Frr = XX' Fyy = W' (30.Bll)

1 1 1
Fy = ¥ - ¥' Fss = 52 Fry =F*ry...J(Frr'"J:yy)

For a shear test, (Jr = 5 and the resulting coeffi- 1 1 1 1

cient is F" = X - X' Fy = ¥ - ¥'
(30.B6) We can recover the von Mises failure criterion
for isotropic materials when F lfrry = -1/2 and
ai ,
The sixth coefficient, F must be determined
empirically from biaxi tests such that
by setting the strengths to be the same in all
directions
Fry =F*ry...J(F:xlw> X = X' = ¥ = ¥', 5 = X/...J3 (30.BI2)
To have a closed failure envelope, F lfrry must be where the ...J3 factor is the result of the von
between -1 and 1. Mises invariant. Substituting expressions in
eqn (30.B12) into those of eqn (30.Bll) and
then rewriting eqn (30.BI0) results in
30.8.2 RECOVERING VON MISES CRITERION
(J2 - (J (J +try + 302s = X?- (30.B13)
It may be useful to review the quadratic failure " "y
criterion and make it less of a mathematical which matches eqn (30.B9).
abstraction by comparing it with what is
DESIGN OF STRUCTURE WITH COMPOSITES 31
F.J. Schwan

to avoid confusion in this chapter. The term

31.1 INTRODUCTION
'material properties' is used generally to
Composite materials are finding more uses include extensional and compressive moduli
each year across a wide spectrum of applica- of elasticity, Poisson ratios, inplane shear mod-
tions. Designers of structural components, in ulus, coefficients of thermal expansion and
particular, continue to find new applications for coefficients of moisture expansion. 'Design
reinforced composite materials. The most suc- allowables' refer to material strengths, specifi-
cessful structural applications are those where cally tension, compression, inplane shear,
the innate advantages of reinforced materials interlaminar shear and bearing. This distinc-
can be translated into performance advantages tion between material properties and strengths
for the manufactured part. These advantages is established because of the significant differ-
include stiffness-to-density ratios, strength-to- ences which exist in our ability to predict and
density ratios, low thermal expansion measure each group, and because of the differ-
characteristics, and occasionally others, such as ent treatment which properties and strengths
resistance to specific environments, thermal receive in all phases of the design process.
conductivity, and fatigue characteristics.
This section focuses on the design of com-
31.2 DESIGN PROCESS
ponents determined by mechanical
performance requirements such as stiffness, A general comment concerning the contents of
durability, or strength. The importance of this section, and the chapter in general, is that
establishing and articulating design require- the perspective represented throughout the
ments cannot be overstated. The extent to chapter is that of a designer within a large
which design requirements are isolated and organization, operating as part of a large
quantified determines the degree of certainty design team, a situation typical of aerospace
associated with trade studies and material projects. However, the process is the same for
selection decisions. Unsuccessful applications all sizes of projects involved with design of
are quite often the result of unclear or poorly composite structure. The primary differences
defined design requirements. are in the formality associated with each step in
Two terms which will be used extensively the process. In large organizations, extensive
throughout this chapter are 'material proper- documentation tends to occur at each step,
ties', and 'design allowables'. Since there are while in one-man teams little informal, and
no universally accepted definitions of each of usually no formal documentation is required.
these terms, it is necessary to adopt definitions The process itself, however, is the same.
A typical conceptualization of the design
Handbook of Composites. Edited by S.T. Peters. Published process is presented in Fig. 31.1. While the
in 1998 by Chapman & Hall, London. ISBN 041254020 7 process is generically the same for all engineer-
710 Design of structure with composites

Peliminary design I Definition of requirements I

, I
Material data
I Trade studies I

I Manufacturing data
Tooling concept
I
I
I
Preliminary material selection
,
I

•,
I Cost trades I
I Establish configuration; select materials I
w
Member sizing - laminate design

-
Engineering - materals
- joint design ~ development - critical elements
Detailed design - tooling design tests - inspection techniques

I Material specs j.-[ ,

Manufacturing development I
Process specs
Design -analyses
verification - prototypes
- testing
Fig. 31.1 The design process. - non-destructive evaluation

ing materials, the content of each box in the Table 31.1 Design requirements for an aircraft sta-
chart varies with many factors, including the bilizer
materials of construction. The intent of this dis-
cussion is to provide some guidelines on what Typeo! Description
information is required to make key decisions requirement
and eventually converge on a design when Geometric Maximum length
composite materials are under consideration. constraints Maximum root and tip dimensions
The complex series of decisions which we Production Number of units per month
refer to as the design process is commonly rate
considered to comprise two phases: prelimi- Cost/ Maximum weight; $ per kg of
nary design and detailed design. Each phase benefits weight saved
requires a different level of requirements defi- Minimum service life; + $ per
nition and application. additional hour of service life
Figure 31.1 indicates the information and Environmental Service temperature extremes
data required to proceed in each phase. This Chemicals - cleaning solvents,
data typically includes many kinds of infor- fuels, oils
- concentrations
mation about the materials of construction and - exposure periods
the fabrication processes to be considered.
Design characteristics for each material are an Mechanical Stiffness
important subset of this information. - maximum tip deflections
- aeroelastic performance
The objective of the preliminary design Strength
process is to develop a design concept to a - imposed strains acceptable
level that includes configuration, materials, under design loads with factors
and preliminary assessments of performance, of safety included
cost and manufacturing approaches. To sup- - fatigue life derived from
port these determinations, a set of material required service life
characteristics data is required. As an example,
key design requirements and a sample of what
data is required to support preliminary design
 Design process 711

Table 31.2 Preliminary materials data for aircraft stabilizer study

Type of data Form of data

Available material forms Prepreg
- tape widths
- thicknesses
Pultruded forms
- shapes
Resin systems
Specification details
- fiber volume
- void content
-density
- etc.
Material availability
Manufacturing characteristics Handling characteristics of material in automated tape laying process
Cure characteristics
- gel time
- resin flow
Prepreg storage conditions
Cost data Material price
Material development tests
Manufacturing hours
Tooling costs
Environmental resistance Service temperature limits
Moisture absorption characteristics
Chemical resistance
Mechanical characteristics Material properties
- extensional and shear modulus
Allowable strains
-tension
- compression
-shear
Fatigue data
- strain against number of cycles to failure

of an aircraft component are shown in Tables The data required to support preliminary
31.1 and 31.2 respectively. The component design investigations include stiffness, den-
selected for this example is a vertical stabilizer. sity, strengths, fatigue characteristics, data at
The design requirements define maximum tol- elevated and low temperature, forms in which
erable deflections, imposed loads, a weight the material is available and associated mater-
goal, production rate, cost goal, service life ial prices. In addition, some experience with
and operating temperature extremes. In real- the material in the fabrication process selected
ity, the set of preliminary design requirements is required to support decisions concerning
would be much larger including, perhaps, production rates and costs. Generally, reliable
maintenance and inspection characteristics, published data can be used for preliminary
descriptions of qualification tests, center of design. Only when no data is available is it
gravity characteristics, and numerous others. necessary to generate data in this phase of the
design.
712 Design of structure with composites

The sequence of decisions in this design materials incur increased engineering develop-
case would typically be to make a preliminary ment, testing, and increased raw material costs,
decision on a manufacturing technique, fol- data from previous applications is quite valu-
lowed by selection of one or more materials able in performing cost trades and making
based on cost and availability. Once this has preliminary design decisions.
been done, the component is 'sized' for stiff- Table 31.3 suggests the ever-increasing
ness, strength and service life in each scope of structural applications for reinforced
candidate material. Material properties and composites. The requirements identified in
strengths are used for these calculations. this table are as diverse and varied as the
Material selection may depend on weight components. The result of the wide range of
minimization or life cycle cost minimization requirements imposed upon composite mate-
or a combination of the two criteria. rials, compounded by the ingenuity of
The result of this process is a material selec- material developers, has led to an over-
tion and a preliminary design. This design'
I whelming array of constituent materials.
possesses enough detail to support confidence Many of these materials are very specific sys-
that (1) the material selection is justified, and tems with a narrow range of applications.
(2) that the design concept is feasible in this Some materials are suited to numerous uses.
material. The set of mechanical properties and This multiplicity of materials and features has
strengths was, by no means, a complete or final produced a plethora of design data. In order
set of design values, but rather contains those to make material comparisons and selections
characteristics needed to support preliminary on a rational basis, the designer needs to
design calculations. These values are generally establish a clear set of requirements for the
the 'best available' from material supplier data, material. These requirements are sometimes
literature publications, data from other projects captured in a document which governs mate-
and data from material specifications. rial procurement. The aerospace industry
While the dominant design goal of aerospace typically refers to these documents as mater-
designers is usually to reduce weight, the dri- ial specifications.
ving requirements vary widely from one The situation in which the structural
product to another. Table 31.3 provides a syn- designer usually finds himself at the initiation
opsis of the diversity of these requirements of a design includes a broad and usually
across defense and commercial industries. This incomplete set of requirements and a moun-
table indicates that requirements are numerous tain of literature on a vast array of materials. A
and, therefore, the derived design objectives are design checklist, such as that contained in
diverse and numerous. Payoffs are achieved in Table 31.4, may be of some help in getting
the form of weight savings or greater service started in this situation. This checklist offers a
life. Sometimes, as in the case of carbon-carbon systematic approach to determining what
rocket nozzles, the use of composite materials information is needed at each step and sug-
results in a new level of performance which is gests which design team members need to be
difficult to quantify due to a lack of economic involved at various points in the process.
alternatives. Table 31.3 shows a very large The structural design team will employ a
range in quantified payoffs. Each of these prod- process which resembles this one in order to
ucts reflect a history of successive applications develop a preliminary design which specifies
and successive generations of the same applica- configuration, materials, manufacturing
tion. With each successive development effort, process, and establishes a basis of confidence
the industry has become better at understand- to proceed with final design. This basis com-
ing and defining requirements, material prises having answers or a plan to develop
characteristics, and payoffs. Since composite answers to all key questions.
Table 31.3 Design requirements and objectives for various composite products
Sector Product Structural item Primary structural Primary design Payoff
requirements objectives
Defense Aircraft Airframe Compression strength Minimum weight $l03/kg
Damage tolerance
Joint strength Maximum service
Durability life $/h
Helicopter Rotor blades Tensile strength Minimum weight $l03/kg
Stiffness
Fatigue life Maximum life $l03/h
Understructure Stiffness Minimum weight $l()2/kg
Energy absorption Crashworthiness
Rocket motors Motor cases Tensile strength Minimum weight $l<J4/kg
Nozzles Resistance to Survivability
elevated temperature at 2000°C (3600°F)
Satellite Primary structure Stiffness Minimum weight $lOS/kg
Low thermal expansion Dimensional
stability
Marine Hull structure Compression strength
(submersibles) and stability Minimum weight $l03/kg
Joint integrity Maximum depth
Commercial Sailboats Hull structure Bending strength Minimum weight $10/kg
Stiffness
Corrosion resistance Minimum maintenance
costs
Automotive
-Race cars Bodies Stiffness Minimum weight $l03/kg
Durability Maximum life $l03/h
- Production cars Body structure Stiffness Minimum weight $/kg
Energy absorption Crashworthiness
Suspension Strength Minimum weight $lO/kg
Durability Maximum life $1O-2 /mi ~
Recreational Golf shafts Stiffness Minimum weight $10/kg OQ"
;:s
and tennis rackets ~
Industrial Water tanks Tensile strength Maximum life $lD2/yr ~
pressure Air tanks Corrosion resistance Minimum weight $/kg ~
<:I>
<:I>
vessels
'l
....
VJ
714 Design of structure with composites

Table 31.4 Design checklist for composite structures

What do I need to proceed with design of a structural component in composites?

1. A clear design objective What am I trying to achieve?
2. Some quantitative measure of what
that improvement is worth
3. A manufacturing approach Are the tooling concepts feasible?
Are the fabrication processes well defined?
Do I need tool try units?
4. Material properties and allowables Which characteristics are critical?
How will I establish and verify these?
5. Material procurement Will materials be procured to specifications?
Do all specifications already exist?
Will materials be available in suitable time and quantities?
Is material· development or tailoring necessary?
6. Joint concepts How is the component to be attached to adjacent parts?
7. Design verification Do I need special tests because I am using composites
e.g. exposure to temperature or moisture?
What kind of prototypes do I need?
What tests will I subject the component to?
8. Inspection What am I looking for?
What are my accept/reject criteria?
What techniques will I use?

Material
procurement

Fig. 31.2 Design team.

 Preliminary design 715

Figure 31.2 shows the makeup of a typical preliminary design. The primary assumption
design team. For designs with composite mate- made in this section is that the materials of
rials, the team may be larger than design teams construction are laminated composites with
using other engineering materials. The focus of continuous fiber reinforcement. These materi-
the team is the individual with design responsi- als are typical of structural applications.
bility. Essential members are representatives of
manufacturing, tooling, and materials procure-
31.3.1 MATERIALS
ment. Others might include materials
suppliers, and specialists in testing and analy- The subject of material design data, and the
sis. In some product development activities, the closely related subjects of material characteri-
team includes subcontract managers or mater- zation and specifications, assum~ greater
ial procurement specialists who interact both importance for composites than for conven-
with other team members and with other com- tional materials. There are several primary
panies who also need to be considered as team reasons for this: lack of a single, comprehen-
members. More important than the list itself, is sive design database or source; greater
the early and continued involvement of. all variability in measured values from one mate-
members. Interaction among these team mem- rial to another and from one lot of the same
bers produces the design. As shown in Fig. 31.1, material to another; differences between pre-
preliminary design begins with requirements dicted and measured component performance
definition and produces a design complete to introduced by variables associated with
the point of materials identification, fabrication 'workmanship' in the manufacture of the part.
approaches, size and shape, and preliminary The most basic reason, however, is the simple
lay-ups and joint designs. The development of fact that the engineering development process
this information requires the constant input encompasses design of both the material and
and review of all key team members. the structure. For these reasons, the designer
One feature of the design team which is of composite structure needs to focus more
unusual is the inclusion of the material sup- attention, and usually more resources, onthe
plier in the design team. There are several subject of materials, than his counterpart
reasons for doing this. Sometimes, existing using only conventional materials.
materials do not 'fit the bill'. In these instances, Material data is required in order to make
further material development or modification key design decisions, first in the preliminary
may become part of the design process. The design phase and later in detailed design. The
material specification can be looked .at as a type and extent of the data depends on the
blueprint for the material supplier to use in application, but, typically, data on the stiff-
design of the desired material. It is important nesses, strengths, and densities of candidate
to recognize the material supplier as a member materials is needed.
of the design team so that his inputs and Figure 31.3 shows relative ranges of specific
development efforts can be integrated into the tensile modulus for composites with various
design. Otherwise, some beneficial options fiber reinforcements. Specific modulus is a
may be excluded, both in terms of existing term which refers to the ratio of composite
materials and near term material development modulus to composite density. Specifi,c modu.-
efforts which may be of value to the project. Ius is a measure of stiffness per kilogram of
material, and is commonly used as an aid in
material comparisons. Figure 31.4 shows the
31.3 PRELIMINARY DESIGN
ranges of material costs for these same families
The discussions which follow concemthe por- of materials .. Both Qfthese figur~s were con-
.tion of the design process characterized as structed considering the entire range .of
716 Design of structure with composites

300.00

III 250 .00

~
"3
-g
:E;;""'
200.00
E
f
Gl
~ 150 .00
~o.
~~ 100.00
'u
l 50 .00

0 .00 + - - - - - t - - - - - t - - - - - j - - - - - t - - - - - - l
E-<lLASS ARAMID T300 M I-fJI
Fl3ER Fl3ER GRAPHITE GRAPHITE GRAPHITE
Fl3ER

Material
Fig. 31.3 Specific tensile modulus of various composite materials.

properties available from each specific mater- materials is usually made on the basis of per-
ial. These ranges are the result of including formance. Figure 31.4 considers only material
different forms of the material and the entire cost, which in some applications, is the major
range of laminate values, including quasi- portion of product cost. In other applications,
isotropic values on one end and unidirectional material cost is completely overwhelmed by
values on the other. manufacturing costs or other life-cycle costs.
Figure 31.4 suggests some general conclu- Material cost is typically about 30% of compo-
sions. First, there are three more or less distinct nent cost in aircraft parts, less than 10% in
regions of composites. The first group is glass- satellite components, and generally higher than
reinforced composites. The second is 30% in most commercial applications. It is
intermediate modulus graphite and aramid therefore critical that product cost analysis
fiber-reinforced composites. The third group include all important cost contributors.
includes high and ultra high modulus graphite Generally, 'you get what you pay for' or,
fiber composites. Each of these types is suitable more precisely, 'you must pay for what you
for a distinct range of applications. A compari- need.' However, the range of properties and
son of where the common metals fall on this costs within a family of materials is an indica-
figure provides a pretty good explanation of tion of the latitude which a structural designer
whyaramid and graphite have replaced metals has when employing reinforced composites.
in applications where the market can 'pay for The upper end of each range of specific
performance'. Glass has replaced metals in modulus in Fig. 31.3 is established by values
those applications where some performance for a unidirectional laminate. While this is one
can be sacrificed for lower cost, increased envi- equitable basis for comparing reinforced mate-
ronmental resistance or longer life. rials to each other, it is not a good basis to use
The second conclusion suggested by Fig 31.4 for design calculations because so many
is that there is extensive overlap in the price of requirements drive the design towards multi-
aramid and intermediate modulus graphite directional reinforcement. Foremost among
fiber composites. Selection from this group of these is the need for properties in more than
 Preliminary design 717

1000

900

800

'iii 700
~ 600 .1-N SYt.B:lUZES HIGH t.Q)ULUS FIBERS
I-
(J)
0 INCLUDING GY70, M6OJ, Pl00 & Pl20
(.)
500
..J
:$ 1M SYMBOU2ES I/IlTERMEDIATE MODl.UJS

.......
a: 400 FIBERS INCLUDING T40, TSO, T650 AND
M40J
e
~ 300

200

- -
100

...~
~
~ I ~
J:
c..
~
C)

~
I I
~ ~

Fig. 31.4 Costs of composite materials.

140

120

100 ~
'1
~
80
...
<II
::::I

~
~ 60

~
40

20

II) 0 Ii) 0 II) 0

o '" <0 "'" ..... co CD 0)

LAMINATE ANGLE,. (cieg.)

Fig. 31.5 Mechanical properties of ±a laminates - T300/epoxy (0.60 fiber volumes).

718 Design of structure with composites

one direction. Figure 31.5 shows the relation- forced composite materials. This table shows
ship of tensile modulus in x and y directions mechanical properties and strengths for four
and inplane shear modulus (xy) for a varying fiber-reinforced materials. These material
lay-up angle, (). The same type of relationship properties and strengths are the basic data
exists for strengths and thermal expansion from which laminate values are calculated. AB
coefficients. A design which maximizes modu- explained in Chapter 30, lamination theory
lus in oriedirection at the same time simply uses layer values to determine engi-
minimizes modulus in the direction transverse neering constants and to estimate strengths for
to this and minimizes inplane shear modulus any desired laminate. Materials testing is
as well. Strength in the transverse direction therefore conducted primarily at the layer
and inplane shear strength will be low in mag- level. Laminate test coupons are employed to
nitude also. These low properties and confirm predictions and 'workmanship'. The
strengths in the 'secondary directions' will exceptions to this statement are strengths
severely limit the ability of the laminate to which are laminate-specific, such as bearing
resist load and deformation in these direc- strength and interlaminar shear strength.
tions. Highly directional laminates are However, these are values which are impor-
therefore not appropriate for structures which tant to only specific regions of the component,
act as plates or shells, and for beams with sig- and generally do not enter into material selec-
nificant secondary loads. Highly directional tion decisions.
laminates also present severe design con- For preliminary design, laminate values are
straints in joint regions where loads are typically predicted from available ply-level
multi-directional in nature. values. Design calculations and trades are con-
Multi-directional reinforcement offers a ducted with these predicted laminate values.
more robust design because strengths and Later, in detailed design, predictions are con-
properties are dominated by fiber properties firmed with selective tests of design laminates.
in all inplane directions. Furthermore, lami-
nates with multi-directional reinforcement
Carpet plots
offer more desirable cure characteristics and
usually avoid high levels of residual stress[ The infinite number of combinations of ply
internal damage or warpage in thin parts, thickness, ply angle and material reinforce-
induced by cooldown from cure temperatures. ment is both a curse and a blessing in
Successful instances of unidirectionally rein~ preliminary design. Once the design require-
forced parts are nearly non-existent. . ments have been identified and quantified, the
Table 31.5 provides a typical list of materials question of how to meet these requirements
data required for structural design in rein- with any given material presents itself. An

Table 31.5 Typical materials data for preliminary design

Material Mechanical properties Strengths

Fiber volume Density Modulus Tensile Compressive
fraction (g/cm3)
Long. Trans. Long. Trans. Long. Trans.
(GPa) (GPa) (MPa) (MPa) (MPa) (MPa)
T300Fabric 0,6 1.55 68.9 68.9 620 620 414 414
M60JTape 0.6 1.69 248 5.52 993 27.6 496 172
Pl00Tape 0.6 1.74 414 4.14 827 27,6 414 138
Ararnid fabric 0,6 1.36 37,2 37.2 827 827 138 13$
 Preliminary design 719

extremely useful tool in addressing this ques- forcement in each of three specified directions.
tion is the simple carpet plot. This graphical While 0, 45 and 90° plies are most commonly
representation provides a way to present the used, the three angles can be any selected val-
entire range of properties, or strengths, avail- ues. Another common combination is
able from a particular family of laminates. 0/30/60. Typically, values are calculated using
Figure 31.6 shows a carpet plot of tensile mod- a difference of 10% between successive points,
ulus for laminates comprised of layers of which requires about 1000 data points per set
T300-epoxy with orientations of 0,45 and 90°. of curves. For this reason, creation of carpet
The plot contains all possible proportions of plots is done with the aid of computers.
the three directions, and is read as the legend Development of carpet plots of strength
indicates. This type of plot treats laminates as values is less straightforward. The primary
homogeneous combinations of various per- complication is the definition of failure of the
centages of layers. Carpet plots do not laminate. An extensive discussion of laminate
recognize lay-up sequence. Generally, this pre- failure modes is contained in Chapter 30.
sents no limitation, but for laminates Calculation of laminate strength involves con-
composed of very few plies, or laminates sideration of progressive failure of each layer
employing significantly unbalanced stacking and construction of strength envelopes for
sequences, the assumption can lead to incor- combined states of loading. Construction of a
rect strengths. carpet plot requires an approximation of ply
Carpet plots of material properties are con- strength which is adequate for preliminary
structed directly from a table of values for trade studies and initial member sizing. The
many specific laminates. Values of moduli, recommended approach is to define laminate
Poisson ratios, and expansion coefficients are strength as the product of allowable fiber
calculated from lamination theory for a series strain and the laminate modulus in the appro-
of laminates with varying percentages of rein- priate direction. This calculation assumes that

140

120

_ 100 -0--0% 0 LAYERS

r. --10% oLAYERS
!2. -0- 20% 0 LAYERS
~
...I
80 --A- 30% 0 LAYERS
§ --40% 0 LAYERS
....... 50% 0 LAYERS
:IE
60
~ -+-60% 0 LAYERS
~ .......... 70% 0 LAYERS
J!! 40
-IB-8O% 0 LAYERS
--®-90% o LAYERS
--+- 100% 0 LAYERS
20

0
0 10 20 30 40 50 60 70 80 90 100
PEACENTAGE GoOEGREE LAYERS

Fig 31.6 T300/epoxy x-direction modulus.

720 Design of structure with composites

the laminate contains fibers oriented in at least 427 MPa and 15.5 GPa, respectively. If this set
two directions, and further ignores the Poisson of properties were not suitable, the three carpet
effect in the calculation of strength. While this plots would be used to iterate to an optimum
approximation tends to overpredict strength, set of values and an accompanying laminate.
it has proven to be suitable for preliminary The use of carpet plots can greatly facilitate
laminate selection in strength critical struc- the laminate selection process in the early
ture. Using this definition of strength, carpet stages of design. It should be noted that these
plots can be directly constructed from allow- plots are simply visual representations of the
able tensile and compressive strain values, results of lamination theory calculations, using
and predicted modulus values. ply level input values. Carpet plots of all
Carpet plots provide a quick method for inplane material properties: elastic moduli,
selecting a candidate laminate for the prelimi- Poisson ratios, thermal expansion coefficients
nary design process. For example, let us and moisture expansion coefficients can be
assume that design calculations indicate that a developed from the appropriate equations.
modulus of about 90 GPa is required. The car-
pet plot of Fig. 31.6 offers a large number of
laminates which meet this single requirement. Design values
One is a laminate comprised of about 55% Appropriate values for properties and
0° layers, 35% 45° layers and 10% 90° layers. If strengths are required in both of the primary
tensile strength and inplane shear stiffness design phases - preliminary and detailed. For
were important to the design, it would be nec- preliminary design, values are typically
essary to consider the values associated with acquired from existing sources. These can
the selected laminate. Related carpet plots pro- include data published or provided by mater-
vide ranges of tensile strength (Fig. 31.7) and ial suppliers in the form of brochures, design
shear modulus (Fig. 31.8). These are about handbooks, data sheets, data published in

900

800

70 0 ;t::....:::-::.::m::.:.:.;...=\II..- -.. -... -

! ~ O% 0 LAYERS~
-;-
~ 600
~10% o LAYERS
-0- 20% 0 LAYERS
I
~ 500 L_- ._.-_..._.. ._-..- _-._-4_....-_._
--1.J-_.- __....:
_ :::".
. _._ -6-30% 0 LAYERS
z -1- 40% 0 LAYERS 1
LLI 1
a: -B- 50% 0 LAYERS
IiiUJ 400 1....... 60% 0 LAYERS
o LAYERS
I
I
...I ........ 70%
in
~ 300 · - ' -lD- "'" 0 LAYERS
-®-90% 0 LAYERS
200 i ::,+:-_100~~.P_~YERS

100

0 +---~---4----+_--~---+----~--~--_+----~~
o 10 20 30 40 50 60 70 80 90 100
PERCENTAGE 45-DEGREE LAYERS

Fig. 31.7 T300/epoxy x-direction strength.

 Preliminary design 721

40

35

30
'i
~ 25
(J)

3
5 20
o:;;
a: 15
~
~ 10
~
5

o +----r--~~--~--_+----r_--~--_+----+_--~--~
o o o o o o o o o
co o
C\J M '<t LO
'" CI> o
PERC£NTAGE~DEGREeLAYERS

Fig. 31.8 T300/epoxy x-direction shear modulus.

professional society proceedings or publica- Test methods also tend to vary from one lab-
tions, data from textbooks such as this one, or oratory to another. In general, it is best to be
data developed on previous projects within wary of specimens other than the standard
the organization. An abbreviated list of ones. Standard coupons and test methods are
sources of preliminary design data is pre- defined by the American Society for Testing
sented at the end of this chapter. and Materials (ASTM), and other organiza-
As might be expected, this available data tions such as Suppliers of Advanced
comes in various forms, so comparison of data Composite Materials Association (SACMA).
from different sources may require some judg- The primary value of employing standard
ment and mathematical manipulation. Fiber coupons and tests is the confidence established
volume fraction usually varies from one with time and existing databases. While each
source to another. The data is sometimes nor- standard coupon may possess some shortcom-
malized to a 'standard' value, typically about ing, it does offer consistency and a basis of
60%. This normalization process is rarely comparison with existing data.
explained in the literature, but most often con- Fiber volume and coupon configuration are
sists of simply multiplying measured values only two variants present among published
by the ratio of 0.60 to whatever fiber volume data. Others include the resin system, material
was measured. The values most frequently form, and method of fabricating panels from
normalized include tensile and compressive which specimens are taken. These variants
modulus, and tensile and compressive underscore the reasons why existing data can
strength. The validity of normalizing compres- be useful for preliminary design, but almost
sive strength is debatable because always needs to be augmented or replaced by
compression strength of reinforced composites data specific to the design application for final
is dependent, to some extent, on the matrix. design. Two different approaches are
The validity of normalizing values other than employed in the development of design val-
the four mentioned above is dubious. ues. The first of these is to buy existing
722 Design of structure with composites

commercial grade composite material, charac- displacement and rotation at both ends, and
terize it, and design to those values. The loaded by a uniform load along the entire
second is to include the material supplier in the length, I, as shown in Fig. 31.9. Design require-
process by jointly developing a specification to ments are summarized below:
which the delivered material must comply.
While the second approach may increase mate- w=2.0 ntlmm
rial costs, it guarantees the suitability of the
material and usually results in the use of
higher or more specific design values.
-d'******'***~ ~ Ih
;::1~-------i~ D
A material specification typically defines I... 1= 1500 mm ~I
acceptable values of all critical characteristics of
the material to be purchased. The specification
may be a simple one page document specifying Fig. 31.9 Example: trade study of beam with a uni-
reinforcement, resin system, and limiting val- form load.
ues of fiber volume, or a multiple page
document including minimum values of mod-
1. Maximum beam weight is not to exceed
ulus, strengths, cure characteristics, and other
5.0 kg or 3.33 g/mm excluding end fittings.
items. Material procurement cost will tend to
2. Maximum beam deflection (at midspan)
follow the number and stringency of the criteria
must be less than 2.5 mm under the load
specified, since the material supplier will need
shown in Fig. 31.9.
more quality assurance activity before delivery
3. Maximum stresses in the beam must not
and may need to perform additional material
exceed allowables, including a factor of
development. The more demanding a specifica-
safety of 2.0.
tion becomes, the greater the need to identify
4. The cross section of the beam must be
and work with one (or more) material suppliers
closed and rectangular as shown.
in its development. One reason for this is sim-
ply to ensure the development of a specification Four materials will be considered for this
which can be met by at least one supplier. A application: aluminum (7075 alloy), E-
good specification needs to recognize the nor- glass/ epoxy composite, T300 graphite/epoxy
mal variation in material characteristics from composite, and M60J graphite/epoxy compos-
one manufactured lot of material to another. ite. Formulas for deflection, maximum
Data illustrating this lot-ta-Iot variation will be compressive stress, and beam weight, in
shown and discussed in a later section of this pounds per inch, are presented below:
chapter.
Deflection: d = (wl4/384EI)
where EI = (Ebth2)/2
31.3.2 DESIGN EXAMPLE
Max. stress: fe = 2Mc / I
Key aspects of the preliminary design process fc = (2M/hbt)
have been described in this chapter. An exam- = (wU/4hbt)
ple, consisting of the design of a composite
Weight: W = v(2h+2b)t < 3.33 g/mm
beam, is included here to illustrate the
methodology typically employed in prelimi- One additional constraint must be introduced to
nary design, and to provide a demonstration the design. Buckling of one side of the rectangu-
of a trade study performed for the purpose of lar section is governed by elastic moduli and by
material selection. the width-ta-thickness ratio of that side. The
The structural element selected is a beam of crippling strength of a side can be significantly
constant cross section, fixed against lower than the laminate compressive strength or
 Detailed design 723

inplane shear strength. In order to prevent com- However, the cost of a beam fabricated in T300-
pressive or shear buckling of the sides and epoxy may still exceed the cost of an aluminum
flanges, a rule of thumb is to limit the side beam, despite the fact that less T300 material is
dimension, h, to no more than twenty times the required. The next calculation might be to esti-
thickness, and to limit the flange dimension, b, mate the fabricated costs, including materials,
to no more than fifteen times the thickness. of the T300 and aluminum designs, to ascertain
Using these relations, h and b are eliminated the cost differential. A final comparison can be
from the equations and replaced by multiples of made between T300 and M60J fiber reinforced
t. The three equations are then solved by itera- beams. As presented in Table 31.6, the addi-
tion for each of the four materials using tional weight savings associated with an M60J
properties and strengths shown in Table 31.6. beam is relatively small. The cost differential is
The summary table presents a comparison of going to be substantial because M60J prepreg is
the calculated design thicknesses and corre- over $200 per kilogram while T300 prepreg is
sponding beam weights for each of the four about $20 per kilogram. Only a substantial
candidate materials. The deflection requirement payoff in increased component life ot overall
proved to be the dominant requirement for all system performance would justify the selection
four materials, i.e. the thickness required. to sat- of M60J over T300.
isfy the deflection requirement was greater, in In a real design situation, the trade would
each case, than that required to limit imposed probably be more complex. Perhaps, the, load
stresses to acceptable values. The estimated pre- would be a cyclic load or an intermittent load,
liminary design weight of each design is less leading to consideration of fatigue allowables.
than 5 kg, as required.. Or, perhaps the natural frequency of the beam
Several conclusions can be drawn from this is another important design consideration in
comparison. First, the aluminum and E-glass the structural system. The trade methodology
designs are comparable on a weight basis due remains the same, but additional characteris-
to the fact that the lower density of fiberglass tics would need to be considered.
offsets its lower stiffness. End fittings have
been ignored, along with numerous other
31.4 DETAILED DESIGN
details, in this preliminary study, but it is
worth noting that the additional weight intro- Many of the activities which constitute
duced by metallic end fittings would probably detailed design are extensions or iterations of
cause the E-glass detailed design to be heavier activities already performed in preliminary
than the aluminum design. The T300 / epoxy design. However, a few topics require special
design offerS a significant weight savings over attention. This section will focus on three areas
aluminum and E-glass, deriving primarily specifically. These include a discussion of the
from its higher modulus-to-density ratio. development and use of material design data,

Table 31.6 Results of design trade study of composite beam

Material' Modulus Compressive strength Density Wall thickness Weight

(CPa) (MPa) (g/cm 3) (mm) (g/mm)
Alum (7075) 68.9 482 2.68 2.7 1.37
E-glass 34.5 475 1.93 3.2 1.38
T300 99.3 572 1.50 2.5 0.66
M60J 248 379 1.69 2.0 0.47
• Design values for each composite material are developed assuming a laminate with the following proportions: 65% 0°
plies; 25% 45° plies; and 10% 90° plies.
724 Design of structure with composites

the approach to joining and the development measurement of this property, and the
of joining details, and the construction of expected variation in this characteristic.
drawings. Differences in design values and measured
strengths are larger. However, preliminary
design values are all exceeded by average mea-
31.4.1 USE OF MATERIAL DESIGN DATA
sured values, confirming the fact that strength
The first of these three areas is an extension of critical regions are conservatively designed.
the discussion of material data introduced in There are some differences in values mea-
the section on preliminary design. The point sured from one lot of material to another. The
was made in that section that material data kind and amount of variation is described in
relies heavily on ply-level values. A typical set more detail in Figs. 31.10 and 31.11. Specified
of mechanical material data is shown in Table 'acceptance values' are shown on each figure
31.7. The material of interest is TSO graphite for reference. In this case, these acceptance val-
fiber in an epoxy resin system. The material ues are lower than those used for preliminary
was ordered in significant quantity to an aero- design. The figures also show resu1ts of tests
space specification and was manufactured in conducted by the material supplier and by the
ten lots. Table 31.7 is subdivided into material receiving company. Differences in these resu1ts
properties and strength values. The values provide some insight into the influence of test
shown as 'measured' are average for all speci- methods, facilities, and personnel. The first
mens from all ten lots of material. The table figure compares tensile modulus and the sec-
also includes preliminary design values for the ond tensile strength. Figure 31.10 indicates
sake of comparison. that, while values are different from lot to lot,
This comparison shows excellent agreement all lots exceed the specification value for ten-
between measured and preliminary design sile modu1us and are therefore acceptable. The
material properties. The largest differences are measured ply-level data confirms the design
in coefficient of thermal expansion (CTE) val- values of tensile modu1us. The measured val-
ues, and these are on the order of 20-25%. This ues for tensile strength exhibit larger
is excellent agreement considering the difficu1- variations, but are also higher than the speci-
ties and uncertainties associated with the fied value. Once again, all measured values

Table 31.7 Mechanical design values for TSO/epoxy tape: preliminary and measured layer properties and
strengths

Material cluzracteristic Preliminary design value Measured value

Tensile modulus, El,(GPa) 209 228.2
Tensile modulus, E2,(GPa) 6.89 7.03
Poisson ratio, v 0.27 0.245
Inplane shear modulus, G12, (GPa) 5.52 5.44
Longitudinal CTE (x l~;oC) -0.63 -0.79
Transverse CTE (x l~;oC) 28.8 25.0
Tensile strength, F/, (MPa) 1034 1136
Tensile strength, F2t, (MPa) 24.1 28.06
Compressive strength, Flc, (MPa) 758 821.2
Compressive strength, F2c, (MPa) 110 171.6
Inplane shear strength, Fl2, (MPa) 62
Subscript 1 denotes fiber direction.
Subscript 2 denotes direction transverse to fiber.
Subscript 12 denotes shear in the plane of the fiber.
 Detailed design 725

300

l250
~
I/)
::;)
..J
::;) 200
c
o
~
~ _LMSCDATA
~ 150 _ HEXCEL DATA
~ ___ LMSC SPECIFICATION VALUE
..J
«
~ 100
E
(!)
z
9 50

o
2 3 4 5 6 7 8 9 10
LOT NUMBEA

Fig. 31.10 Lot-to-Iot variation in tensile modulus - T50/F384 unidirectional tape (manufactured by
Hexcel).

1400

.. 1200
0..
~
~ 1000
Z
w
~
~. 800
w _ HEXCEL DATA
..J
iii _ LMSCDATA
z
~ 600 ___ LMSC SPECIFICATION VALUE
..J
«
z
is
E 400
(!)
Z
o
..J 200

o
2 3 4 5 6 7 8 9 10
LOT NUMBER

Fig. 31.11 Lot-to-Iot variation in tensile strength - T50/F384 unidirectional tape (manufactured by Hexcel).
726 Design of structure with composites

exceed specification levels and all lots are database for establishing preliminary design
acceptable. values for the next application.
This data is typical of aerospace applica- Table 31.8 contains a further comparison of
tions. The types and amounts of data will, of predicted and measured laminate values for
course, depend on the specific application. TSO-epoxy tape, developed in the same appli-
The major difference between this data and cation. The values in this table correspond to
preliminary design data is that the data sup- specific design laminates. Three sets of values
porting final design will be specific to the are shown for each of the design laminates -
material form, resin system, and fiber volume the first developed from preliminary design
selected. Final design data is an extension of layer values, the second from measured layer
data available for preliminary design. It will values, and the third directly from tests per-
generally focus on ply level values since this is formed on laminate coupons. Using measured
the most cost-effective way to obtain data for layer properties produces a better correspon-
all of the laminates which will be employed in dence between predictions and measurements,
final design of the part or assembly. These val- but not an exact overlay. The primary differ-
ues now become an important part of the ence is workmanship, i.e. changes introduced

Table 31.8 Comparison of measured and predicted material design values for T50/F584 graphite epoxy
tape laminates

Design value lAminate description

(60/O/-60/0)s (45/-45/02)s
Preliminary Prediction Measurement Preliminary Prediction Measurement
design design
Tensile modulus, 129.5 141.1 148.9 (60) 115.4 125.2 129.6 (10)
Ex' (GPa)
Tensile modulus, 51.4 55.6 51.1 (60) 28.8 30.8 26.8 (10)
E ,(GPa)
pY. .
Olsson ratio, v 0.303 0.303 0.289 (60) 0.762 0.776 0.777 (10)
Inplane shear 19.8 21.2 29.3 31.8
modulus, G,;r (GPa)
Longitudin CTE -0.58 -0.49 - 0.41 (5) -0.67 -1.08 -1.17 (3)
(x1~rC)
Transverse CTE -0.29 +1.55 +2.02 (6) -0.01 +3.51 +4.10 (3)
(x l~rC)
Tensile strength, 641 703 693 (60) 565 578 781 (10)
Fxl, (MPa)
Tensile strength, 324 341 170 (60) 255 284 116 (10)
F I, (MPa)
Compressive strength, 469 507 623 (60) 414 446 605 (10)
F.c, (MPa)
Compressive strength, 255 278 165 (60) 255 235 128 (10)
F/, (MPa)
hi.plane shear strength, 165 134 214 209
Fxy' (MPa)
Number of test specimens shown in parentheses alongside measured values.
Subscript x denotes the 0° direction.
Subscript y denotes the 90° direction.
 Detailed design 727

by handling and fabrication that simply are allow ables, and so provides maximum lever-
not accounted for by theory. These differences age in terms of applying testing resources to
are larger for strengths than for properties, development of design confidence.
and account for the knockdown factors nor- A less extensive series of laminate tests
mally used in developing material design may be required in addition to the unidirec-
strengths. tional coupon testing. The purpose of these
One final point to be made concerning these tests would be to provide some selective con-
comparisons is that differences exist, but are firmation of key design properties and
not typically an inhibitor to designing with strengths for comparison with values pre-
composites. These variations must be recog- dicted from measured layer values. In
nized in the design process, and can be addition, some very specific strength charac-
handled with statistically-derived knockdown teristics, such as bearing strength, are specific
factors, and conservative but realistic design to laminates, and cannot be reliably predicted
approaches. Key design values can also be from layer values. If this allowable is critical
controlled through development of intelligent to the design, tests of specific design lami-
material specifications. In most designs, the nates may be necessary.
usual variations are easy to accommodate. The
important point is that they need to be quanti-
31.4.2 JOINING
fied and included in the design process.
This leads to a discussion of the role of test- Joining of composites normally presents some
ing in the development of design allowables. It challenges and tends to influence the configu-
was previously pointed out that normally little ration of the part. The two recognized
or no testing is performed in support of pre- methods for joining composite structure to
liminary design. In fact, the only testing which other composite or to metallic parts are adhe-
is essential is that performed to determine crit- sive bonding and mechanical fastening. The
ical values which are just not available. advantages and disadvantages of each method
In the process of detailed design, however, will be discussed, followed by some guide-
the picture changes. Once a material is lines for selection between the two. The
selected and a design acquires some maturity, discussion is not an exhaustive one by any
the values critical to establishing sufficient means, but is intended to highlight key con-
confidence in the final design can be identified siderations which determine joint designs and
and addressed through test. The emphasis is provide some initial direction to the designer.
usually placed upon layer level coupons for Numerous factors need to be considered in
two reasons. The first is that the information the selection process. Figure 31.12 offers a list
obtained from unidirectional, standard of the most common requirements. It should
coupons is the most direct information avail- be kept in mind that some of these 'require-
able about the material. Failure modes are ments' may become design variables in the
predictable and repeatable, and results can be course of the process. Geometry of the mem-
directly converted to allowables. An addi- bers being joined, for instance, could be
tional benefit deriving from the use of altered locally to facilitate joint design.
unidirectional coupons is that direct compar- Reliability potentially includes an array of
isons to other materials and different lots of requirements, one of which is the implications
the same material can easily be made. of joint failure on system performance. Is it
The second reason for allocating the major- catastrophic or benign? Is it easily detectable
ity of testing resources to ply-level testing is prior to failure or insidious?
that this data can be used for prediction of a Determination of this list of requirements
wide range of laminate properties and will lead directly to the formulation of key
728 Design of structure with composites

Joint design criteria

Static strength
Durability (fatigue)
Avaliable
area for
load transfer
How important
Length is weight?
How critical
.....-_--'-_ _-, is cost?

Achieve uniform pressure

throughout cure
Minimization of peel Implications of
and tensile stress failure

Fig. 31.12 Joint design.

design questions possibly including, but usu-

ally not limited to the following: lIB \

• Will the joint be disassembled and reassem- Mechanically fastened

bled? - fatigue life
• What is the nature and magnitude of the - insensitive to temperature
- ease of inspection
loads? - reliability
• Are the loads cyclic or static or a combina-

,
tion?
• What are the operating temperature and I
moisture ranges?
• What are the cost limitations? Adhesively bonded
- no local reinforcement
Some of the major advantages of bolted joints (load spreading)
and bonded joints are shown in Fig. 31.13. - fewer parts
High tolerance to repeated loads, good resis- -lower cost
- corrosion resistance
tance to most environments, ease of
inspection, and high reliability are primary Fig. 31.13 Advantages of bolted and bonded joints.
advantages offered by mechanical fasteners
in composite structure. Fewer pieces, lower
weight, good load distribution, and lower selection of the adhesive system, and process-
cost are advantages offered by adhesive ing. Extensive information on each of these
joints. subjects is available in other sections of this
The major considerations in design of a text. The following design procedure is offered
bonded joint can be grouped into five cate- as a facilitator to the design process.
gories. These include joint strength, Generally, component configurations are
environmental resistance, joint geometry, predetermined by other design requirements,
 Detailed design 729

and so the joint geometry is established prior other sections, and so this discussion is aimed
to finalization of the adhesive joint design. at providing a brief summary of the process.
Factors including how well the mating parts The process begins with the determination
will 'fit up' to each other after allowances for of a configuration for the joint. Single lap joints
all specified dimensional tolerances, and the are normally adequate for thin laminates (up
quality of the adherend surfaces need to be to about 5 mm (0.2 in) in thickness). Fastener
recognized and defined. bending and initial bearing failure are primary
The first step in the process is to determine concerns. Double lap joints are better for cyclic
a dimensional configuration which minimizes (fatigue) loads, and generally stronger.
tensile and peel stresses in the bonded joint. The next step is to select the type of fastener
Once this is accomplished, the next task is to required. Fastener selection usually raises
select an adhesive system which best satisfies issues requiring decisions concerning laminate
static strength, fatigue life, and environmental reinforcement, hole sizes and locations,
requirements. This selection is sometimes the drilling, fastener installation, and inspection.
result of a simple trade study among several Table 31.9 identifies issues and proven design
candidates. The third step is the development approaches to each issue. The table reveals
of formal or informal process specifications for that the complexity of designing bolted joints
the joint. This specification will include details derives from two primary sources: (1) com-
for surface preparation, curing the joint, and posite laminates cannot redistribute high local
maintaining pressure during cure if necessary. loads by yielding and plasticity; (2) compos-
The joint strength is typically verified analyti- ites are more easily damaged by drilling and
cally or by structural test or both. fastener installation than metals.
Similarly, a design process can be outlined Design of local reinforcement of the lami-
for mechanical joints. The primary design con- nate to resist local stresses is an important step
siderations for bolted joints include joint in the design of bolted joints. If reinforcement
strength, fastener type, local reinforcement, is required, a proven approach is to increase
joint configuration, holes, and preload. All of laminate thickness by the addition of plies
these subjects are discussed extensively in placed at ±45 and 90° to the primary load

Table 31.9 Design issues: fasteners in composite laminates

Issue Solution approach

Drilling damage 1. Closely controlled manufacturing operations
2. Inspection of drilled holes
High local stresses 1. Larger fastener diameter
2. Insert (bushing)
3. Increased laminate thickness (locally)
Preload relaxation 1. Large fastener head
2. Washers (one or both sides)
3. Limit on installation torque
Countersunk head 1. Avoid, if possible
2. Increased laminate thickness (locally)
Damage induced by installation of blind fasteners 1. Specially-designed blind fasteners
and drive rivets 2. Development of modified installation techniques
3. Verify joint strength with tests
730 Design of structure with composites

direction. A quasi-isotropic laminate provides

the best bearing strength in any continuous 4xO.25
fiber polymer composite.
~b
~
LocalreinforceEnent
~
As in metal structures, local reinforcement is (1.0)
Scale: 1/8 S ' B View B
generally required where any hole or cutout is eevlew
placed in a structural part. Analogous to metal L 0.10 (reference) / ~ - ~ L
0.20 (reference)
parts, reinforcement can be bonded or fas- I : J
tened to the structure. Reinforcement can often
T \ T
A-A
be incorporated into the design of the part in Scale 1:2
the form of additional plies extending across a
limited region of the structure. Local reinforce- Fig. 31.14 Local reinforcement.
ment may also be required in the vicinity of
joints, either bonded or bolted, and in loca-
tions where concentrated loads are introduced
into the structure. in the stiffnesses of the two plies. Therefore,
The design of the reinforced region will load transfer is most effective between two
derive primarily from the orientation and plies with the same fiber angle and least effi-
magnitudes of the stresses in the structure, cient between two plies with a 900 difference
and so no generalizations can be made con- in fiber angle. The use of 450 plies between 0
cerning laminate design. However, some and 900 plies is a good compromise for rein-
generalizations concerning transition in these forcement areas.
regions and the behavior of laminates are AnalysiS methods are limited by underly-
appropriate. ing assumptions, and are most reliable for
Laminates behave as a series of layers, each regions where the assumptions are simple
stressed in membrane action, and each capable and few. This is not the case for regions in
of transferring stresses to adjacent layers which the laminate changes and the stress
through shear. When laminates bend, adjacent gradients are steep. Therefore, complete
layers remain attached and experience states reliance on analysis in the regions of joints
of strain which are similar. Therefore, in a rein- and local reinforcement is not prudent. In
forced region, such as that shown in Fig. 31.14, addition, these regions sometimes present
load is introduced into the reinforcing plies manufacturing challenges which raise ques-
through interlaminar shear. Since the level of tions concerning the quality of the finished
interlaminar shear stress required to induce part. Therefore, testing is very often an inte-
delamination in any composite will not exceed gral part of engineering development and of
70-80 MPa, a design must aim to maintain the design verification process.
these stresses to acceptable levels. This is typi-
cally done through the gradual introduction of
31.4.3 DRAWING CONSTRUCTION
reinforcing plies, as shown in Fig. 31.14. This
gradual transition precludes a sudden discon- A drawing of a component or assembly in
tinuity in thickness which would produce composite materials has the same objective as
sharp peaks in interlaminar shear stress. this type of drawing in any material: furnish-
The selection of ply orientation is also a ing a blueprint for manufacture of the item.
consideration. Transfer of load from ply to ply However, some of the details and the manner
is facilitated by the minimization of differences in which information is presented on the
 Detailed design 731

drawing are unique. There are countless varia- table defining the material: fiber reinforce-
tions on the format of a drawing, so the ment, resin system, material form, resin
discussion will be more or less limited to the content, material source or sources, and sur-
content, with a specific format shown only for face treatments. The second is to invoke a
the purpose of providing an example. document, such as a specification, which
A drawing detailing a composite part needs defines the material required. If the second
to specify geometry, materials, sources of approach is used, it may still be a good idea to
those materials, ply sequence for each unique provide a brief identification of the material
laminate, details of transition regions, process- on the face of the drawing to facilitate inter-
ing definition, post fabrication instructions for pretation by analysts, tool designers,
drilling, sawing, etc. and any inspection manufacturing and inspection personnel.
requirements. Other details, such as individ- Table 31.10 shows an example of material
ual ply patterns, may be necessary for a description employing both methods.
specific part, but this list is suggested as one of Specifying a fabric requires the designer to be
general applicability. familiar with all of the variable characteristics
Dimensions are defined in the same manner of the material form - weave style, surface
as for metal parts with the exception of thick- treatment, and end count. This is necessary
nesses. Thickness is really defined by laminate because fabrics are available in numerous
sequence meaning that the part will be as thick weave styles, with various surface treatments,
as the cured laminate dictates. A thickness and can be woven with any of several differ-
dimension is normally shown as a reference ent size yarns.
dimension and provided for tooling design. It Each unique laminate in the part requires
should be recognized that specifying a thick- definition sufficient to allow a fabricator to lay-
ness as a fixed dimension demands up the part. Figure 31.15 presents a typical
compliance to that dimension. Unless manu- example of the usage of a ply sequence table on
facturing 'tool try' units have been fabricated the face of a drawing. The figure identifies the
to determine cured laminate thickness with surface that ply number 1 is placed upon and
the particular manufacturing process speci- then details the ply stacking sequence to be fol-
fied, there is no guarantee that the thickness lowed. It is desirable, but not always practical,
can be achieved with the number of plies and that the first ply shown in the table be the first
resin content specified. ply physically laid up on the tool. The ply table
Dimensional tolerances are generally deter- contains a column identifying the material of
mined by tooling design and the amount of each ply since sometimes a single laminate
material shrinkage. Experience with the contains plies of different materials.
process and material is the only reliable guide Areas of transition are handled as shown
in setting tolerances. in Fig. 31.16. The two adjacent regions of
Material definition can be accomplished in 'constant' thickness are defined by appropri-
two ways. The first is to create a note or a ate tables as shown. The tables, however,

Table 31.10 Format for material definition on drawings

Material designation Characteristics Source

M1 Fabric prep reg, 120 style weave, Company name and address
T300 fiber, 10K bundle size,
934 resin system, sizing
M2 M40Jtape Specification number
732 Design of structure with composites

Material Fiber orientation

Ply number designation e)
1 M1 90
2 M1 o
3 M2 45
4 M1 90
5 M1 o
6 M1 o
7 M1 o
8 M2 45
9 M1 o
10 M1 90

Fig. 31.15 Ply sequence table.

Ply Material Fiber angle

(C) Ply Material Fiber angle
(C)

1 M2 90 1 M2 90
3 M2 0 2 M2 0
4 M1 45 3 M2 0
5 M1 -45 I 4 M1 45
12 M1 ·45 5 M1 -45
13 M1 45 6 M2 90
14 M2 0 7 M2 0
16 M2 90 8 M1 45
9 M1 ·45
10 M2 0
11 M2 90
12 M1 ·45
13 M1 45
14 M2 0
15 M2 0
16 M2 90

Fig. 31.16 Specifying transition regions.

employ a common ply numbering sequence specifications typically include definition of

so that the continued and terminating plies cure cycle (temperature and pressure against
can be distinguished. The termination details time), pre-cure instructions for vacuum bag,
are specified on the face of the drawing as bleeder plies, resin preheating for pultrusion,
shown in Fig. 31.16. tool preparation, removal of the part from the
Processing instructions are supplied in one tool, in-process inspection requirements, and
of two ways, either as notes, or by reference to many other items.
a document such as a process specification. Post-fabrication operations may include
The second alternative is by far the most drilling holes, sawing to length, removing
common because of the amount of informa- edges by grinding, secondary bonding of
tion required to define the process. Process items, painting, light sanding, and almost
 Detailed design 733

anything else. These operations are normally The advantage of documenting each NDE
specified on the drawing by means of notes. activity is traceability. If a problem arises in
Notes need to identify items such as the type development or production, NOE records can
of drill bit and cutting speed. usually identify potential sources and trends.
The disadvantage is cost. Each company per-
forms its own cost/payoff determination, and
31.4.4 DESIGN VERIFICATION PROCESS
develops its own verification processes from
A few words about design verification are that determination. One point for the designer
appropriate to a section dealing with compo- to keep in mind is the necessity and value of
nent design. Design verification encompasses this verification for composite parts. Variability
more than the -nondestructive evaluation in raw materials, processes, and workmanship
(NOE) procedures specified on the compo- can and will affect the final part. The designer
nent drawing. Design verification is integral needs to be aware of all potential sources of
to the entire design and manufacturing variation, so that he knows where to look when
process. Figure 31.17 identifies verification final inspection reveals defects in his part.
activities in each phase of the component Part inspection methods are usually identi-
development process. Each of these activities fied on the drawing. These requirements
ranges from a formal, documented step in actually form the accept/reject basis for the
some industries, to undocumented, highly finished part, and can range from visual only
integrated steps in other industries. The key to extensive ultrasound methods. Table 31.11
point is that verification is the part of the presents a summary of which defects are
development process which ensures that the detectable with each of several established
developed component fulfills all design NOE techniques. The most commonly used
requirements. methods are X-ray and ultrasonic, particularly

Development Primary Verification

phase Activities actions

r- Development tests
Design
Engineering Malerial
development Procurement Material acceptance tests
Tool design
~ Tool tries
Kitting
Tool prep - In-process inspection
Layup - Processing records
Fabrication
I Debulk
Bagging
Cure
-
- Tag end coupon test
Trim - Witness panels

Bonding - Dimensional verification

Post Drilling holes
fabrication
Sawing Non-destructive inspection
Part
acceptance
'-- Proof test / acoustic
emissions

Fig. 31.17 Component verification process.

734 Design of structure with composites

Table 31.11 Defects and nondestructive test methods

Defect or variable X-ray Gamma Ultrasonic Sonic Microwave Temperature Penetrant

ray differential
Unbond X X X X X
Delamination X X X X X X X
Undercure X X X
Fiber misalignment X X
Damaged fibers X X
Variation in resin
Variation in thickness X X X X X
Variation in density X X X X
Voids X X X X X X X
Porosity X X X X X X
Fracture X X X X X
Contamination X X X
Moisture X

C-scan. These methods require proper equip- SOURCES OF MATERIALS DATA FOR
ment, cumbersome setup in some instances, PRELIMINARY DESIGN
and time. X-ray and ultrasonic methods can
detect a wide range of defects, and are sensi-
tive in ranges where other methods are not. Military Specifications
However, these methods have definite limits.
Mil-R-9300B Resin, Epoxy, Low Pressure
In fact, no single method can detect all types
Laminating
of defects or variations. Moreover, it is impor-
Mil-C-47257C Compound, Epoxy, Filament
tant to understand, in advance, what a defect
Winding
looks like. For this reason, it is common prac-
MiI-R-24719 Resins, Vmyl Ester, Low
tice to perform calibration with intentional
Pressure Laminating
defects built into test panels. Mil-Y-1140 Yarn, Cord, Sleeving, Cloth
Table 31.11 is an oversimplification of an
and Tape, Glass
extensive and complicated subject. It is Mil-R-60346 Roving Glass, Fibrous (for
included here as a handy reference for the
Filament Winding
designer and a starting point for determining
Applications)
which techniques are appropriate for a given
Mil-Y-83371 Yarns, Graphite, High
part. Inspection requirements will, of cours:' Modulus, Continuous
be related to the function of the part, and will
Filament
be focused on the critical aspects of the design.
Mil-Y-87125A Yarn, Graphite, 1000/3000
In some companies, it is common practice to
Filaments
limit part inspection to simple dimensional
Mil-F-87121A Fabric, Graphite Fiber
and visual checks, and to rely on a 'proof' test
Mil-T-29586/1 Thermosetting Polymer
to verify the structural adequacy of the com-
(Epoxy) Matrix, 350°F (177°C)
ponent. Since inspection requirements tend to
Cure, Intermediate Modulus
be unique to each component, it is usually nec-
Carbon Fiber Reinforced
essary to make inspection requirements an
Prepreg Tape
integral part of the design drawing.
 Sources of materials data for preliminary design 735

Aerospace Material Specifications Modern Plastics Encyclopedia, McGraw Hill,

AMS 3828 Glass Roving, Epoxy Resin New York, yearly edition.
Preimpregnated, Type E
Composite Design Encyclopedia, University of
Glass
Delaware, Newark, 1984.
AMS 3892/8-84 Tow or Yam, Carbon
(Graphite) Fibers - For Composite Materials Handbook, (ed. Mel M.
Structural Composites Schwartz,) McGraw Hill, New York, 1984.
AMS 3894/11-83 Carbon (Graphite) Fiber
Tape and Sheet, Epoxy
Resin Impregnated Periodicals
AMS 3822 Ooth - Type fE' Glass, 'B'
Modern Plastics Magazine, 1221 Sixth Avenue,
Stage Epoxy Resin
New York, monthly periodical
Impregnated, Style 181-75DE
SAMPE Quarterly Journal
Federal Journal of Advanced Composites
Mil-Handbook 17C, Polymer Matrix Composites,
Volumes 1 and 3, Standardization Documents
Material Suppliers
Order Desk, Building 4D, 700 Robbins Avenue,
Philadelphia, PA 19111-5094 Plastic Design Guide Owens/ Coming Fiberglas
Corporation, 1974
Reinforced Plastics By Design PPG Industries,
Professional Society Publications
Fiberglass Reinforcements Market Series
Society for the Advancement of Material and
Hexcel Design Guide
Process Engineering (SAMPE) - Conference
Proceedings ICI/Fiberite Materials Handbook
Society of Manufacturing Engineers (SME) - Data Sheets from material suppliers:
Conference Proceedings
ICI/Fiberite
Hexcel Aerospace Products
Handbooks
AMOCO Performance Products
Handbook of Reinforced Plastics. (eds. S.S.
Oleesky, J.G. Mohr). Van Nostrand Reinhold, Nippon Oil Company, Ltd.
New York.
ANALYSIS METHODS 32
v.v. Vasiliev

32.1 INTRODUCTION appropriate model for the structural element

A detailed analysis of composite structures can under consideration.
provide information about the stress-strain To start the analysis (Fig. 32.1), one must
state, strength, stiffness, stability and vibra- know the geometric configuration of the struc-
tional behaviour of structural elements. The ture, material properties, environmental and
analytic methods are those used in solid operational conditions along with the applied
mechanics; the equations are modified to loads. It must be noted that the available data
reflect the structural behaviour of composite on material properties and loading conditions
structures. These equations are based on a sys- affect accuracy, duration and cost of the analy-
tem of assumptions that take into account the sis. As an illustration, consider a cylindrical
key features of a structure and establish the shell loaded with a concentrated radial force.
The solution for this problem can be obtained

ENDIRONMENTRL RNO
OPERRTIONRL RSSUMPTIONS
CONDITIONS RNO MODELS

STRUCTURE
GEOMETRY

MRTERIRL
PROPERTIES
EHPERIMENTRL
RNRLYSIS
SPECIRl
REQUIREMENTS

MICROMECHRNICS

Handbook of Composites. Edited by S.T. Peters. Published Fig. 32.1 Outline of analysis for composite struc-
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 tures.
 Introduction 737

in the form of a Fourier series that has a very the experimental method can not be used
poor convergence at the loaded point. for unique structural designs.
However, a closer inspection of the loading • To minimize the degree of uncertainty, a
conditions often shows that the concentrated second method is applied. According to this
force is the resultant of some distributed load method, the stiffness coefficients are calcu-
over an infinitesimal area; this distributed load lated from data that specify the composite
can then be expressed by the series with a material, ply co-ordinates, ply thickness,
finite number of terms to achieve convergence. ply orientation and mechanical properties
Operational requirements for the composite of individual plies. The latter, in principle,
structures are also of great importance. For can be predicted by analytical or numerical
example, a filament wound composite shell of methods of micro-mechanics of composite
revolution can be a model for either a gas pres- materials; these results, however, have
sure vessel or a solid propellant rocket motor more academic than practical value.
case for which, in contrast to the pressure ves- Idealized micro-mechanical models, that
sel, time dependent material properties need approximate the real structure of the mater-
not be considered. On the basis of this knowl- ial, do not account for material porosity,
edge, assumptions are formulated for the non-uniform fiber distribution, possible
model of the structural component under variation of physical properties in each
study. This model should consider only the batch of resin, variation in strength of the
key factors affecting the capability of the struc- resin/ fiber interface and a number of other
ture and ignore the secondary effects that factors associated with the manufacturing
complicate the analysis without significant process; in many instances, the above varia-
improvement. tions can not be described in formal analytic
The next step implies analytical formula- terms.
tion of the problem, i.e. development of a
For important structures, usually, the mechan-
complete set of the governing equations with
ical ply-properties are experimentally
the pertinent boundary conditions. For struc-
determined by testing specifically fabricated
tural analysis, the governing set of equations
plates, rings, or tubular unidirectionally rein-
consists of equilibrium equations (or equa-
forced specimens produced by the same
tions of motion), strain-displacement and the
process as the structure under study!.
constitutive equations which are material
Once all of the pertinent data are together,
dependent. The constitutive equations include
the governing equations are solved by either
stiffness coefficients which are determined by
the analytic or the numerical methods applied
either of the following two methods:
in solid mechanics. The results by either
• The first method is associated with direct method yield information on deformation,
experimental analysis of test specimens cut displacements, stiffness and strength. To eval-
out of the structure or its excess length. This uate stiffness, the calculated displacements
method is used mainly for structures in due to an applied loading can be used
mass production. For simple laminates (e.g. directly. Evaluation of strength requires trans-
for cross-ply laminates), this method gives formation of the calculated global strains to
rather accurate results while for the more ply strains in terms of principal material co-
complicated (e.g. including angle-ply lay- ordinates, determination of the ply stresses
ers) laminates, separation of a specimen can and application of some strength criterion.
cause a disruption in the material structure One must note that the numerous strength cri-
and induce specific edge-effects with ques- teria, that have been developed to date, are no
tionable mechanical properties. Obviously, more than analytical approximations of the
738 Analysis methods

experimental results with a significant scatter.

Also, the failure of a composite material is a
rather complicated process that defies accu-
rate theoretical description. Therefore, the
strength of important structures should be
always verified by testing either sub-scale or
full-scale models.
Finally, a thorough analysis of a composite
structure, based on macrostructural models,
effective stiffnesses and approximate displace-
ment fields, does not preclude the possibility
of a detailed analysis of inlaminar or interlam-
inar stresses. To accomplish this analysis, the
ply under study is separated out of the struc-
ture, loaded with unknown contact forces,
deformed in accordance with the known Fig. 32.2 Element of a composite structure.
global strains of the structure and described by
the proper microstructural model.
1. Equations of motion
32.2 GENERAL EQUATIONS FOR COMPOSITE
STRUCTURES
Composite structural elements consist of a large
number of plies with different angular orienta-
tion of fibers. The general equations, describing
the behaviour of the composite elements, are
based on the following assumptions:
• the structural element consists of perfectly
bonded anisotropic layers made out of lin-
ear elastic materials;
= 0 (1,2)
• the laminate thickness (h) is much smaller
than the other dimensions;
• the radii of curvatures are assumed to be
the same for each lamina;
a
da (A2Q,.) +
a
ap (AIQ~) -
(Na N~)
AIA2 RI + R2
• loading does not affect laminate thickness.
The differential composite element (Fig. 32.2)
is described in terms of curvilinear coordi-
-A 1A 2( : - fy) - !
(A 2N,p)(J.)

nates a, p and y. The coordinate axes a and P a a

coincide with lines of principal curvatures of - da (A~(J.~w~) - ap (AIN(J.~w(J.)
the basic surface located at a distance e mea-
sured from an outer element surface. The a
geometry of this basic surface is defined by the - ap (AIN~w~) = 0 (32.1)
coefficients of the first quadratic form AI' A2
and the principal radii of curvature RI and R2•
Therefore, the governing equations of com- where
posite structural analysis become as follows:
 General equations for composite structures 739

a aAz a where
La(F) :;:: aa (Ala> - F~ aa + ap (Al ~)
aA l
+ F~ ap (1, 2) The symbol (1, 2) indicates, hereafter, that an
equation written for the a variable yields one
and F :;:: (N, M). more analogous equation for the p variable by
commutating the indices a, p, 1 and 2.
The foregoing set of equations (32.1, 32.2
2. Constitutive equations and 32.3) include 23 equations and the same
number of unknowns, namely:
• the stress resultants N, Q and the couples M
acting on the basic surface shown in Fig.
32.3;
• displacements ua' u~ w of the points on the
basic surface in the a, p and y directions;
• rotations ea' efl of the normal to the basic
surface;
• rotations of the tangents (Oa and (Op to the
basic surface in the ay and py planes;
• strains e and flexural deformations K of the
basic surface;
• transverse shear deformations 1/Ja and 1/Jfl in
the ay and py planes.

(32.2)

3. Strain-displacement equations
1 au
ea :;:: A a: w 1
+ a1u fl + R + TW; (1,2)
1 1
1 aUfl 1 aU a 1
e afl :;:: A
1
aa + A JR - a1u a - a2ufl + TWaW~
2 I--'

Fig. 32.3 External and internal forces and moments.

() a :;:: 1/J a + Wa (1, 2)
U 1 aw The equations of motion (32.1) contain inertia
Wa:;:: Ra - A aa (1,2), (32.3) terms with time derivatives of displacements
1 1
and rotations; the equations reduce to static
740 Analysis methods

equilibrium without these derivatives. The

ma ::: [Fa(t - e)dt + s(qa + q7J) - e(Pa + P~a)
or
coefficients B and Dp characterize the inertia
properties the element in Fig. 32.2 with
respect to displacements and rotations. CpI on (1, 2) (32.5)
the other hand, is the coupling coefficient
describing the interaction between displace- where the Fs are the body forces and P, q are
ments and rotations. the simultaneously applied surface forces as
shown in Fig. 32.3. Functions 7J and ~ reflect
variation of the wall thickness, h (a, p) ::: e + 5,
along the coordinate lines, i.e.

7Ja = ~1 : (1,2)

~~ (1 2)
"'I:a :::
A 1 ua
':I '

It is assumed that functions 5 (a, p) and e (a, p)

vary rather slowly, therefore sin 7J '" 7J, cos 7J '"
Fig. 32.4 Layer coordinates. I, sin ~ '" ~ and cos ~ '" 1. It must be noted that,
for variable thickness, the layer coordinates t;
in Fig. 32.4 can also depend on a and p. For
For a thin walled laminate (Fig. 32.4) these structures with constant thickness, 7J a = 7J~ = 0
coefficients become and ~a ::: ~~ ::: O. The governing set of equations
B ::: 1(0)
(32.1, 32.2 and 32.3) accounts for geometric
p p nonlinearity and is applicable for small strains
C ::: 1(1) - eI(O) and moderate rotations w, (sin w '" w, cos w '" 1)
p p p of the structural element in Fig. 32.2. For thin-
D p ::: 1(2) - 2eIp(1) + e21p(0) (32.4) walled composite structures, the first-order
p nonlinear effects, associated with variation of
where (r ::: I, 2) radii Rl and ~ under loading, can play an
important role; in this case, the underlined
ICr):::
P
Ihpt'dt::: _1_~ p.(t.r+1_F+1)
r + 1 I.J. ' , ,-1 '
rotations ware omissible when compared with
o ;= 1 their derivatives describing the changes in the
P is the material densityi and t; _I' t; are coordi- curvature of the basic surface under loading.
nates of layers. The external loads in (32.1), The linearized buckling equations are formu-
reduced to the basic surface, are specified by lated by modifying eqns (32.1), (32.2) and
the following relations: (32.3) using the following assumptions:
• equation (32.1) excludes the terms corre-
fa ::: [Fadt + Pa + qa + P(~a + W a) + q(7J a -wJ sponding to inertia, body and surface
forces;
(1,2) • in nonlinear terms of eqn (32.1), Na ::: Nao,
N~ = N~o, Na~::: Na~o where Nao, Npo, Na~o are
the stress resultants that correspond to
membrane prebuckling state of the struc-
ture;
• constitutive equations (32.2) should exclude
terms W and MT.
 General equations for composite structures 741

• Strain-displacement equations (32.3) and (32.10) depend on the thermo-elastic con-

exclude non-linear terms. stants of the material. For a composite
layer-referred to principal material coordi-
The membrane (Bmn)' the bending (Dmn)' the
nates 1,2,3 in Fig. 32.5, with axis 1 enclosing
coupling (emn ), and the shear (Kmn ), stiffness
an angle 'Pi or -'Pi with the a axis as in Fig. 32.6.
coefficients in the constitutive equations (32.2)
These coefficients are
are determined with the following equations:

A (i) =A (,) v (E(I) E(') - 2E s

emn = 1(1) -
mn
eI(O)
mn
=:; (i)E(') +
1221.1211212"
+ (i»C 2 2

Dmn = 1(2)
mn
- 2eI (1) + ell (0)
mn mn

K44 K45 K54

Kll = [(' K12 = ~1 = K = K

(32.6)
where (mn = 11, 12, 13, 23, 22, 33; r =0, I, 2)
I (r) = I (r) = fhA trdt
mn nm mn
a
1 k
= -- ~ A (i)(t:+ 1 _ t.r+l) (32.7)
r+1 ~ mn, ,-1
;=1

and (mn = 44, 45, 55)

Fig. 32.5 Element of a unidirectional ply.
K mn = K nm = -h12 fhamndt = -h12 Lk amnJ
(iJ h.

o i= 1
Am
a(iJ = mn
mn AJ) As~) - (A4~) )2 (32.8)
Constitutive equations, (32.2), include also
temperature terms
NT =:; 1(0)
m m

= 1~1) - e/~O)
M~ (32.9)
where (m = 1,2,3; r = 0, 1)
h 1 k
J(l) = J A mt'dt = -r + 1 ~ A (I)(t!+I- F+ 1)(32.10)
m ~ m ,-1 I
o i=1
The stiffness and the temperature coefficients
of the ith layer, Amn' Am' in eqns (32.7), (32.8) Fig. 32.6 Angle-ply laminate.
742 Analysis methods

A 13(il = A31(il = +[£WC

- 1 i
2 thermal strains and stresses are taken to be
zero. The constitutive equations (32.2) also
_£(I)S.2 -
2 I
E
12
(il (C.2 - S.2)] CS.
I I I I
allow for hygrothermal effects if A~) = f31~)Mfj
and A~) =f3J)Mfj where Alii is the change of
A (I)=A (1)=
23 32
moisture concentration in the ith layer of the
material.
For an orthotropic layer, with the
orthotropy axes coinciding with the a and the
f3 axes, the nonzero stiffness and thermal coef-
ficients are as follows:

A (I)
11
=£ a(I)' A (I)
22
= £p(I)

A (I) =A 21(I) = Vall(1)£ a =vIla(1)£ (.) (I)

A 45(I) =A 54(I) =± (G13(I) - G23(I)

12 ~
CS.
..
A
33
(I) =G
all'
(I) A
44
(I) =G"",'
(I) A (I)
55
=GItT
(I)

A
1
(I) = £ a(I) (a aT(I) + V ai3~T
(.) a (i»!l.T.
+1
I
A (!) =·1 (I)E-(I)C 2
11\.12. Ii 1\.21
(I)E-(I)S2
2i

1 (I)E-(I)S2 + 1 (I)E-(I)C2
A 2(I)= 1\.12 1 i 1\.21 2 j
£a,~(I) =Ea,~(') /(1 - vap(l) vlla(l)
(32.11)
In case of an isotropic layer:
where E(i)=E/l(i)=E.
a •

E.
G (i) = G (i) = G (i) = I
all "'" ItT 2(1 + Vi)
=V (1)£(1) + 2G (I)
E1(I)2 121 12
a (i) = a (i) = a
aT ~T i
£ (I) = E (1)/(I-v (11v (I)
1,2 1,2 12 21

A (I) =A(I) + v
121122'212211
(I) A (I) A (I) =A (I) + v (IWI)

where Es (s = I, 2, 3) is the modulus of elastic-

ity in the sth direCtion; Gst is the shear modulus
in the st-plane; v.t and vIs are Poisson's ratios
satisfying the following symmetry condition
E.v.t = Et vt.; a 1T and a 2T are thermal expansion
coefficients; llTi is the temperature change of
the ith layer measured from the value at which Fig. 32.7 A sandwich wall.
 General equations for composite structures 743

For a sandwich wall with a light core and thin

b
facings as in Fig. 32.7, the transverse shear
deformation of the facings can be negligible in
- -
comparison with the corresponding deforma-
tion of the core (A.S,3) = 00, As11,3) = 00) . On the 3 ~
other hand, the in-plane stiffness of the core c j~
can be ignored in comparison with the corre-
sponding stiffnesses of the facings (Alf) = Alf) t3
= Ai} =A3~2) = 0). 2 ~ ~
Therefore, eqns (32.7) and (32.8) yield
k
1 r-- •
I (T) = _1_~ [A (1lt r+1 + A (3)(tr+1_ F+1)] d ~
mn r+1"" mnl mn 3 2
j =1
hz
K
mn
= -a (2)
h2 mn Fig. 32.9 Rib parameters.

LCl,
n
As a first approximation a system of densely
arranged equivalent ribs can be also treated as
A22 = sin41fJj
j=l
a system of continuous layers as depicted in
L
n
Fig. 32.8. For the layers, simulating the rib ele- Al2 = A33 = CjEjsin21fJFOS2tpj
ments in Fig. 32.9, the non-zero inplane j=I
stiffness coefficients are
L
n

A (1) = E!!a A44 = CPFos zlfJj

n j=l

= L
n

Ass cP,sin2tpj
A (2) = E.£
n a j=l
where c. = b./a., E. and G. are the elasticity and
the she~r m6dt'ili for the jth system of ribs. The
A (3)
11
= E.!a stiffness coefficients in the constitutive equa-
tions (32.2) are associated with the following
where it is assumed that the ribs are parallel to deformations of the composite structure:
the a axis, and E is the rib modulus. (a) B11' B12, B22 stretching and contraction of
For a lattice structure in Fig. 32.10: the basic surface due to corre-

LcjEFos tpj
n
sponding loading or Poisson's
An = 4
effect;
j=l (b) B33 shear of the basic surface;

Fig. 32.S A stiffened wall.

744 Analysis methods

Fig. 32.10 Lattice composite structure.

(c) D ll , D 12, D22 bending of the basic surface These equations (note that mn - 11, 12, 22, 13,
due to corresponding loading 23,33) can be satisfied for the following struc-
or Poisson's effect; tures:
twisting of the basic surface;
• homogenous or symmetric with respect to
transverse shear of the wall in
the middle surface of thin laminates for
the ay and Pr planes;
which e =s = h/2 (see Fig. 32.2);
stretching-shearing coupling
• laminates consisting of isotropic layers with
deformation (a) and (b);
different moduli of elasticity (E) and thick-
(g) Cn, C12, C22 bending-stretching coupling
ness (hJ but the same Poisson's ratio, v; =v,
deformation (a) and (c);
for which
(h) C33 shearing-twisting coupling
L L
k k
deformation (b) and (d);
stretching-twisting (a) and (d) e= Ep;(t; + t;_1)/2 E,hj
i=l
and shearing-bending (b) and
• sandwich structures with facings made
(c) coupling deformations;
from the same material but having different
bending-twisting coupling
thicknesses (see Fig. 32.7) for which
deformation (c) and (d);
interaction of transverse shear
deformations in the ay and f3y e = 2(h1: h3) [h 12 + h3(h3 + 2h1 + 2h2)]
planes.
In the general case, eqns (32.12) are incom-
The coordinates of the basic surface, e, in eqns patible, because it is impossible to find a basic
(32.4), (32.6) and (32.9) can be taken arbitrarily surface for an arbitrary stacking sequence
and as a rule, is used to eliminate coupling such that all the coupling stiffnesses simulta-
stiffnesses Cmn in the constitutive equations. neously become zero. However, one of the
Thus, in accordance with eqn (32.6): coupling stiffnesses can always be eliminated
Cmn = Imn(1) - eImn(0) = 0 with the aid of the corresponding equation in
(32.12)
eqns (32.12).
 Composite beams 745

There exists the so-called method of Ua (a, /3, y) = ua (a, /3) + yOa(a, /3) (1,2)
reduced bending stiffnesses according to
which each of eqns (32.12) is used to obtain the
U/a, /3, y) = w(a, /3)
corresponding coordinates
The corresponding strains are
emn = 1mn(1) 11mn(0)
Then, eqns (32.6) yield
Bmn = 1mn(0) eaP (a, /3, y) = cafl (a, /3) + YKa~(a, /3) (1,2)

Strains in the ± CfJ i plies, whose stiffnesses are

Cmn =0
specified by eqn (32.11), have the form
(I (1» 2 e1(i) = ea(i) cos2 itI, + e~ (i) sin2 Y 11'1, ± eap(J) sin Y, cos Y,
11'1, {f1,
- 1 (2)_~
D mn- Y, ,
mn 1(0)
mn
e2(i) = ea(i) sin2 YIfI,, + e~(I) COS2 itI,
't',
± ea~ (i) sin Y,
IfI, COS itI ,
Y,
The method of reduced bending stiffnesses is
not formally established and can give both sat- e12(i) = ± (e ~(i) - ea(I) sin "'"'f',
')"', + ea~(I) COS 2it1,
isfactory results and large errors. Also, it can Y,

be noted that eqns (32.12) can sometimes be where, e.g. ea(l) =ea (a, /3, Y =yJ The stresses, in
satisfied if there exists a possibility to change terms of the principal material coordinates of a
the stacking sequence ot the layers, i.e. to use ply (see Fig. 32.5), are as follows:
coordinates t,1 as unknown factors.
a(i) = E(I)(e (i) + v (I)e(i» (1 2)
The governing set of equations, (32.1), 1 1 1 12 2 '

(32.2), (32.3), is of the tenth order with respect

r = G
to variables a and /3; five boundary conditions
(i) (i)e (i)
12 12 12
need to be formulated at each point of the To evaluate the strength of the ply, these
boundary. If the edge of a composite structural stresses should be substituted into any avail-
element is clamped, then able strength criterion.
ua = u~ = w = 0a = O~ = 0
For a simply supported edge a = constant (1, 2) 32.3 COMPOSITE BEAMS

Na = u~ = w = Ma = O~ = 0 (1, 2) 32.3.1 LINEAR BENDING AND AXIAL

For a free edge a = constant (1, 2) LOADING

Combined linear bending and tension (or

Na = Na~ = Sa =Ma = Ma~ = 0 (1, 2) compression) of a composite beam in Fig. 32.11
where are described by the following equations:
Sa = Qa - NiJJa - Na~())p (1, 2) N' =0
For a linear problem, Sa = Qa.
The solution of the governing set of equa- M'=Q
tions, satisfying the proper boundary
conditions, specifies generalized strains,c, K, 1jJ Q' + fi = 0
and displacements u'" u~, w, ()a' ()~ as functions
of a and /3. Then, the displacements of any N=Bu'
point of the structure can be found in terms of
a, /3 and y by the following expression: M=DO'
746 Analysis methods

PM

tjfu
h
--------~-
L

Fig. 32.11 Composite beam.

k
Q =K(6 + V') (32.13)
where ( ) I =d( ) / dx, P=pb1 - qbk • The axial and
the bending stiffnesses of the beams are
B = 10

D =12 -ell
where e =1/10 is the coordinate of the neutral
axis. For a laminated beam with a cross section Fig. 32.12 Layer coordinates.
shown in Fig. 32.12

I = IhbE yndy = _1_

n " n +1
ti.J b.E' xW(t.n+l_
,
t.n+l)
, -1
o i=1

(n = 0,1,2)

(t
and the transverse shear stiffness is defined by
h
K = h2(I bG
dy )-1 = h2
t. - t.
I.J 'b;G '(~
1)-1
o xy i=1 xy
_ 6_x - MaX 2 _ Q~ + V + V
Successive integration of eqn (32.13) yields the 0"- 2D 6D p R
following general solution:
where quantities with subscript 'a' correspond
N=No to the initial cross section at x =O.
No The following integral terms:
U = uo+Tx
Qp = J: pdx
 Thin-walled beams 747

force, should be taken into account. The corre-

sponding equations describing such a type of
beam behavior have the following form:

Op = b [MpdX
M'=Q

Q' + NV" + P= 0
Vp = fIo dx
p N = B[u' + ~(V')2]
o
account for the distributed loads (see Fig.
32.11) and these next terms M = DO'

(QR' M R, OR' V R) = DQRm, MRm, 0Rm, VRm) Q = K(O + V')

m
describe the action of the concentrated forces. where N is a constant axial force. If the beam
For x < xm' the forces should be QRm =MRm =0Rm has fixed ends, N is determined from the
= VRm = 0 and for x >-
xm boundary conditions. For combined axial
loading and bending, N is a known applied
QRm =Rm
load.
Mm=R
R m
(x-x)
m
32.3.3 BUCKLING
om= Rm(X_X)2 The critical magnitude of an axial compressive
R 2D m
force, causing the column to buckle, can be
determined as
V m = Rm (x _ X )3
R 6D m

Rm =Rm -Fm
where, NE = C1t2 D / U is the critical Euler force,
Displacements and stresses at any point of the L is the column length, and c is the coefficient
beam are defined by dependent on boundary conditions. For a col-
Ux(x,y) = u(x) + y8(x)
umn with simply supported ends c = I, for a
cantilever column c = 1/4, and for a column

ax(x,y) = Ex[ ~ + Mg) y] with damped ends (c = 4).

Txy(X,y) =- ~g) r
o
E)lydy
32.4 THIN-WALLED BEAMS

Composite thin-walled beams are used as

members of trusses, aircraft propellers, heli-
y =y-e copter rotor blades, drive shafts, etc. For an
orthotropic beam whose cross section with
two axes of symmetry (Fig. 32.13), the normal
stress resultant nz (Fig. 32.12) due to combined
32.3.2 NONLINEAR AND LONGITUDINAL
BENDING
axial loading and bending has the following
form:
For nonlinear transverse or longitudinal bend-
ing, the axial strain, due to the large lateral
deflection developed by the applied axial
748 Analysis methods

Fig. 32.13 Thin-walled beam.

where Fig.32.14 Normal and shear stress resultants acting

in a thin-walled beam.
S = pB ds
D= pBy ds2 uy(z,s) = v(z) - x8z(z)

and B is the axial stiffness of the beam wall. uz(z,s) = w(z) + y8x(z) + /(s) (32.15)
For a beam with a non-deformable cross sec-
tion contour B = Bn' For a beam whose where v, ware the displacements in the y and
cross-section contour can be treated as z directions and 8x' 8z are the angles of rotation
absolutely compliant in its plane B = Bll - about the x and z axes of the beam cross sec-
B1i/B22' where Bmn are determined with eqn tion. These displacements are defined as
(32.6). follows:
The shear stress resultant nzs (Fig. 32.14)
developed by a transverse force and a torque
is
M
nzs = QF(s)+-z
yy 2A
where

Fy(s) = .b [Sx(S) + ~ pSx(s)r dS] (32.14)

8z = 8z°+ czIZMzdz
SJs) = /"BY ds where VOl WOl 8xo, 8zo are the°displacements and
o
the rotations of the initial cross section at z = 0,
where A is the area bounded by the contour of 1 ,/; ds
the cross section and r is the length of the nor- Cz = 4A2 :r B
mal to the contour shown in Fig. 32.14. The 33

displacements of an arbitrary point on the and 1/J is the beam shear

beam cross-section contour (Fig. 32.14) along deford;.ation having the following form:
the x, y and z axes can be expressed as
1/J = Q PFy2(s) ds
y Y B33
 Rectangular plates 749

F (s) is specified by eqn (32.6) and B33 is the 32.5.1 LINEAR BENDING OF SYMMETRICALLY
shear stiffness of the beam wall given by the LAMINATED PLATES
corresponding equation in (32.6). If the laminate has a symmetric lay-up, the
Function j(s), entering the last equation in basic surface is at e = h/2 and eqns (32.6) and
(32.15), determines the free warping of the (32.7) yield
beam cross section under torsion and bending k/2
and has the form B = 2 t A (;)(z. - z )
mn ~ mn l t-1

j(s)
Mz JS - ds
=_ -
JS ds + Q JS -j--ds
f}' z r
F (s) i =1
y
2A a B33 a a 33 Cmn =0

1/A~sin /3 ds
k/2
+ Dmn = l3 iJt A mn(;)(Z3 - I
Z3
1-1
)

;=1
Note that the axial (P) and the transverse (Q) where z. are the layer coordinates shown in
forces, as well as the bending moment (MJ - Fig. 32.16. Then, the problem of bending of an
acting on the beam cross sections and entering orthotropic plate under the combined action of
the foregoing equations - are governed by the surface pressures p and q (see Fig. 32.3) is
equilibrium equations for a beam element sim- reduced to the following equation in terms of
ilar to equation (32.13). plate deflection w (note that the transverse
shear deformation is ignored):
32.5 RECTANGULAR PLATES
aw
4 a4w ()4w _-
Composite plates possess high specific D11 ax4 + 2(°12 + 2°33) axZayZ + °22 at - p
strength and stiffness. They are used as struc- where p = p - q. The strains at an arbitrary
tural elements for aircraft, ships and other point of the plate are defined
structures. The governing equations for a plate
(Fig. 32.15) can be obtained in Cartesian .coor- a2w
e = ---z
dinates from eqns (32.1), (32.2), (32.3) If we x ax2
takeA j =A z = 1, I/Rj = l/R z = 0 and replace a,
/3, y, with x, y, z.

-X
b

a/2 a/2

Fig. 32.16 Layer coordinates for a symmetric lami-

Fig. 32.15 Rectangular plate. nate.
750 Analysis methods

a2w
e = ---z
Y at;
where
e
xy
= -2 axay
h z
kiO = Wi (x =a/2)
they allow to find strains and stresses in terms
of principal material coordinates of the com- kit = W/ (x =a/2)
posite plies.
The variational approach, based on the ki2 = W/' (x = a/2)
approximate analytic expression for the i = 1, 2 and 0' = d O/dx. Functions W1(x) and
deflection in the y-direction while satisfying W2(x) are particular solutions that are symmet-
the corresponding boundary conditions, ric with respect to coordinate x and satisfy the
yields the following approximate (but rather following ordinary differential equation:
accurate) solutions for a plate (Fig. 32.15)
loaded with a uniform pressure p = Po = con- Willi - '21PW" + f4W =0
stant. The approximate deflection equations where
for different boundary conditions are as fol-
lows: c2(D 12 + 2D33)
c1D n
(a) Plate with simply supported longitudinal
and transverse edges c3D22
f4= -
w(x,y) = wp [1 - 11(x)]c,01(Y) (32.16) c1D n

(b) Plate whose longitudinal edges (y = 0 and c parameters in these equations have the form
y = b) are simply supported and transverse
edges (x = ± a/2) are clamped c1 = [c,02dY
w(x,y) = wp [1 - fz(X)]c,ol(Y) (32.17)
(c) Plate with clamped longitudinal and sim- c2 = [(c,o/)2dY
ply supported transverse edges
w(x,y) = wp [1- h(X)]c,02(Y) (32.18) c3 = [(c,oll)ZdY
(d) Plate with clamped longitudinal and trans-
verse edges
c= [c,odY
w(x,y) = wp [1 - 12(x)]c,02(Y) (32.19)
The following are used: For solutions in eqns (32.16) and (32.17) it
should be taken c1 = 0.04921 b9, c2 = 0.48571 b7,
Poc c3 =4.8lJ5, c =0.2lJ5. For solution in eqns (32.18)
W =--
P D22C3 and (32.19) c1 = 0.001587 b9, c2 = 0.01905 b7, c3 =
0.8lJ5, c =0.03333lJ5.
32.5.2 BUCKLING AND POST-BUCKLING
c,02(Y) =y2 (b _ y)2 BEHAVIOR OF SYMMETRICALLY LAMINATED
PLATES
k22W1(X) - k12W2(X)
/1 (x) = --=~--:-=:--=--­ In-plane compression or shear (Fig. 32.17)
k1ok22 - k~12 (where the forces Tx' Ty' T xy are uniformly
 Rectangular plates 751

y = 0 is simply supported and

(b) if the edge
the edge y =b is free, then
_ 12(D12 + 2D33) (J!...)2~ Du
x k - 1t'l 22 + a
2:l(DuD) D 22

Under pure shear (Tx =T =0), the critical load

for an orthotropic symketrically laminated
plate is expressed as
xl
TxyC = k-
ab V(D1P22) (32.22)

Fig. 32.17 In-plane loading of a rectangular plate. Coefficient k is given in Table 32.1 for typical
values of the following parameters:
b2 .1 Du
distributed along the plate edges) can result in p = -;q D22
plate buckling. For a simply supported rectan-
gular (a ~ b) plate under uniaxial compression D12 + 2D33
q-
(Ty = Txy = 0), the critical load is expressed as - "J(D1P22)
Since the value of the critical load for an
(32.20) orthotropic plate does not depend on the
direction of shear forces, the parameter p in
where Table 32.1 can be replaced by lip, so that Table
. D12 + 2D33) 32.1 presents coefficients k as p varies from
k = 2(1 + "J(D1P22) (32.21) 0.02 to 50.
Critical combination of compressive and
Note that eqn (32.21) is valid if the longitudi- shear forces can be determined using the fol-
nal edges of the plate (y = 0 and y = b in Fig. lowing equation:
32.17) can experience displacement along the
y-axis. If these edges are fixed in this direction
(which is often the case), then the compressive
Tx
TCx
+( Txy)2
Tc
=1
xy
forces Tx give rise to transverse compressive where TCx and TxyCare specified by eqns (32.20)
forces Ty = vxy Tx due to Poisson's effect. Then and (32.22).
Used as the skin elements of stringer panels
k = 1 + (~xy IA) [Dlf. + 2(D12 + 2D33) + ~221 or shear webs, composite plates Gust as metal
ones) can sustain high compressive or shear-
should be substituted into eqn (32.20) and ing loads after buckling. However, in contrast
minimized with respect to A =(mIl a)2 where m to metal panels whose ultimate loads are usu-
is the number of half waves in the x-direction. ally determined by rib fracture, buckling
The following approximate expressions are failure of composite panels (particularly made
derived for the k coefficients with different from carbon-epoxy composites) is often
plate edge supports: caused by skin fracture due to bending.
(a) if the edges y = 0 and y = b are clamped, Therefore, traditional engineering methods of
then strength analysis, such as the method of
reduced width for compressed panels and the
concept of diagonal stress field for shear webs,
can hardly be used for composite panels; more
752 Analysis methods

Table 32.1 Buckling coefficient k for pure shear

p
q 0.02 0.04 0.06 0.08 0.10 0.14 0.18 0.24 0.32 0040 0.50 0.60 0.80 1.00
0.2 25.5 18.4 15.3 13.5 12.4 10.9 10.4 9.61 8.40 7.50 6.77 6.32 5.87 5.76
0.4 28.9 20.8 17.3 15.2 13.9 122 11.5 10.4 9.27 8.36 7.66 7.20 6.73 6.66
0.6 32.2 23.1 19.2 16.9 15.4 13.4 12.4 11.5 10.2 9.24 8.52 8.08 7.68 7.56
0.8 35.3 25.4 21.0 18.8 16.8 14.6 13.6 12.4 11.0 10.1 9.39 9.05 8.55 8.47
1.0 38.4 27.5 22.8 20.0 18.1 15.8 14.7 13.3 11.9 11.0 10.4 9.93 9.45 9.34
1.2 41.3 29.6 24.4 21.4 19.4 16.9 15.7 14.2 12.8 11.8 11.2 10.8 10.3 10.1
1.4 44.1 31.5 26.0 22.8 20.7 18.0 16.7 15.2 13.6 12.7 12.1 11.6 11.1 11.0
1.6 46.8 33.4 27.6 24.1 21.8 19.0 17.7 16.0 14.5 13.7 12.9 12.4 12.0 11.9
1.8 49.4 35.2 29.0 25.4 23.0 20.1 18.6 16.8 15.4 14.4 13.7 13.3 12.9 12.8
2.0 51.8 36.9 30.3 26.6 24.0 21.1 19.5 17.8 16.4 15.2 14.6 14.1 13.8 13.7
2.2 54.2 38.5 31.6 27.7 25.0 22.0 20.4 18.4 17.1 16.1 15.4 14.9 14.6 14.5
2.4 56.4 40.1 32.8 28.8 26.0 23.0 23.1 19.3 17.9 16.7 16.2 15.8 15.4 15.3
2.6 59.3 42.1 34.4 30.8 27.1 24.1 22.3 20.3 18.8 17.9 17.1 16.7 16.3 16.2
2.8 62.1 44.0 36.7 31.4 28.3 25.3 23.3 21.0 19.5 18.7 17.9 17.4 17.1 17.0
3.2 67.8 48.0 39.3 34.2 30.8 27.4 25.1 22.8 21.3 20.4 19.7 19.1 18.7 18.7
3.4 70.6 50.0 40.9 35.6 32.1 28.5 /26.0 23.7 22.1 21.2 20.4 19.9 19.6 19.5
3.6 73.4 52.0 42.5 37.1 33.4 29.5 26.8 24.5 22.9 22.0 21.2 20.7 20.3 20.2

rigorous solutions of the corresponding non- p, z)is reduced to the following set of equa-
linear problems are required. tions in terms of radial displacement of the
basic surface uri deflection w, and rotation of
32.5.3 NONSYMMETRICALLY LAMINATED the normal to the basic surface Or' i.e.
PLATES r'lu r "" + 6rur III + (7 - np2 - -'b
n. 2)rur " + (1 - np2
In contrast to symmetrically laminated plates,
bending of plates with an arbitrary stacking
sequence of the layers is accompanied by
stretching of the basic surface. Also, the plate
deflections depend on boundary conditions - IfJrdr) - r(rF.)" - (rFJ + ~2FrI (32.23)
imposed on the inplane displacements2 •
For in-plane compression, nonsymmetri-
1
cally laminated plates experience bending Or = C[rB 11(ru r" + u.') - B22ur + rFrl
which should, in general, be described by 22
nonlinear equations. This longitudinal bend-
ing can be unstable and can be usually
accompanied by the so-called mode-jumping.
w = ill (Cll ln r - Jd; JfJrdr) - fOr dr + C2

Here, 0' = dOl dr, fJ =p - q (see Fig. 32.3), Fr(r)

32.5.4 AXISYMMETRIC DEFORMATION OF is a radial body force (e.g. a centrifugal force
CIRCULAR PLATES AND DISKS for a spinning disk), C1, C2 are constants of
integration, and
The problem of axisymmetric bending and in-
plane deformation of orthotropic composite
plates and disks (in cylindrical coordinates r,
 Cylindrical shells 753

32.6 CYLINDRICAL SHELLS

Filament wound composite cylindrical shells
are used as pressure vessels, reservoirs, pipes,
C2
n4 = __22_ aircraft and ship elements. The governing
C BnDn equations for a cylindrical shell (Fig. 32.18) can
Coordinates of the basic surface e = II~I) / II~O) be obtained from eqns (32.1)-(32.3) if we take
provide, in accordance with eqn (32.12), the Al = A2 = I, l/Rl = 0, R2 = R and replace a,,8,y
zero value of the radial coupling stiffness (Cn with x, y, z.
= 0). x
Radial and circumferential strains at an
arbitrary point of the plate are expressed as
er =u'+z()'
r r

1
r r + z()r
epA = -(u

The general solution for eqn (32.23) has the

form
6
Fig. 32.18 Cylindrical shell.
U
r
=~C.r"i+U
I.' p
;=3
where up is a particular solution and Sj are the 32.6.1 AXISYMMETRIC DEFORMATION
roots of the equation One of the most important loading cases for
s4 - (n/ + ~2)S2 + n/~2 - nc = 0 4 cylindrical shells is the axisymmetric loading
with pressures p, q and axial forces N (Fig.
Six constants of integration can be found from 32.18). In this case, the equations account for
the corresponding boundary conditions the first-order nonlinear effects of the axial
according to which ur = w = ()r = 0 for a forces on the curvature of deformed shell
clamped edge, w = Nr = Mr = 0 for a simply meridian. These equations have the following
supported edge, w = ur = Mr = 0 for a hinged form:
edge fixed in the radial direction, and Nr =Mr
= Qr = 0 for a free edge. To write the force
boundary conditions in terms of displace-
N
ments, the following expressions can be used:
Q/ + Nw" - -t + P= 0
N r = Bu u.' +B12+ + C12
U
+()

N=N
x

Qr = ! (C1 - Jprdr)

Mr = C12 + +Dl).' + D +
U
12
()
754 Analysis methods

(32.24)
where ( )' = d( )/dx,p = P- q. Stiffness coeffi-
cients B, C, D, K are specified by eqn (32.6) in Boundary conditions should be written in
which terms of w and force Sx = Nw'.
e =111(1)1111(0)
32.6.2 NONSYMMETRIC DEFORMATION
The foregoing set of equations, (32.24), can be
reduced to the ordinary differential equation In the general case of loading, composite cylin-
Willi - 2s2w" + t4w = kp (32.25) drical shells can be usually described rather
adequately by the. so-called semimembrane
where theory that, in addition to membrane theory,
takes into account the circumferential bending
S2
1 [C
= 2RCD
12 (1 C) RN RB + + + BD11
K 1 moments. The model of a semimembrane shell
. 1 1 1 1 can be represented by a system of rings with
inextensible axes that take only circumferen-
tial bending moments and by a system of
absolutely flexible beams that connect the
rings and sustain axial and shear forces only.
The semimembrane theory assumptions lead
to the following equations:
dNx dNxy
ax-+T = 0
dN dN Q
--y +~+-y +q = 0
dy dX R y

aM =0
dy -Qy
-y

B = BllB22 - B1; .

Constants of integration entering the solution

of eqn (32.25) can be found from the corre-
a; + ;
aQ N
+p= 0
sponding boundary conditions according to au
which w =ex =0 for a clamped edge, w =Mx = Nx = Bn ax
ofor a simply supported edge, andMx= Sx =0,
where Sx =Qx + No/ for a free edge.
Thin-walled composite pressure vessels can
be described by nonlinear membrane equa-
tions using the assumption that Dll = Cl2 = O.
These equations can be reduced to
w"-k2w=kp
where Qy = ~1fJy

Ey =~ + ~ = 0
 Cylindrical shells 755

v
8 =tp + - - -
aw (32.26)
=N n =0 for a free edge. The following expres-
y Y R ay sio~ can be used for the boundary conditions
where p = p - q, q is the circumferential sur- BuR , B11 '" wn '
face traction, and ltiffnesses B, D are specified un = ,FB 2 Wn
n 33
- 1U 4B Wn - 1U 2
n 33 n
by eqn (32.6) in which e = I.i}) /IJO).
Decomposition into Fourier series, i.e. B R B
Nn = _ _ 11_W"
-
_u_ W
B 1U2
P> = 1
x n n
33 n
(u, w, Nx' Ny'My' [un(x), wn(x), N;(x)
n=1

(v, tpy' Nxy' Qy' qy)

-
=1 [vn(x), 1JJ;(X), N';(x)
where
n2 -1 _
Wn = --Dw
R2 n n
-P n
n=1

It must be noted that the semimembrane theory

is not valid for the case of axisymmetric loading
where A. n = n / R, allows reduction of eqn (32.26) . (n =0) and reduces to the membrane theory for
to the following governing equation: n = 1. Proper combination of solutions of eqn
(32.25) for n = 0 and of eqn (32.27) for n ;:?: 1
wn"" - 2s2wn" + t'w n =kp (32.27)
allows consideration a wide range of practical
where problems for composite cylindrical shells.

32.6.3 BUCKLING
Under axial compression by forces N (here, in
contrast to Fig. 32.18, N are compressive
forces), cylindrical shells can experience three
n4p n2 modes of buckling: column-type, axisymmet-
k = B11R2 n
- -B Pn
-" ric, and nonsymmetric buckling. The actual
p 33 critical load is the smallest of the three values.
D22(n 2 -1) For a column-type buckling
Dn = R2(1 + cn)

cn =
where B = B11B22 - Bl~ and m depends on the
character of end fixity. If the end cross sections
are fixed in such a way that they can freely
Four constants of integration entering the rotate (hinged column), then m = 1. If the ends
solution of eqn (32.27) can be found from the are clamped, then m =2.
corresponding boundary conditions according Axisymmetric mode of buckling is typical
to which un = vn = 0 for a fixed edge (it is essen- for thick and sandwich shells. The corre-
tial that inextensibility condition E = 0 yields sponding critical load is specified by the
wn =-RAnvn' so wn = 0 for a fixed eJge) and N: following equation that allows for transverse
756 Analysis methods

shear and radius variation through the shell

thickness:

where
D11 = I11(2) - 1 (I (1»2
1(0) 11
~+ RBB11 (RK11 -C) 11

2
1 - -
(11m)2 Kll DU Ku
BD D22 -- I22(2) _ ~1_(I (1»2
m- L R2B _ Bu RC __1_1 1(0) 22
22
U KU K211
For a homogenous shell

C = C12( 2 - -
C12 )
RKll h2
N--
c- R.J3
Stiffnesses B, C, D are specified by eqns (32.6)
and (32.7) in which e = Il~l) / Il~O); Amn should be
changed for Amn where
A (I)-sA (I) A (J)=A (I) A (I) = A (I)/S. where Ex = E/ (1 - vxyvyx).
"1 - 1 11 '''12 12 "'12 22 I

The critical value of the lateral external pres-

and
sure can be approximated by
1
Sj = 1 + 2R (tj + tj _l -2e) 41t !IB 1Pd
qc = c RLV 27R2

(t t:p~,,' th'
Transverse shear stiffness is
Here, B11 and D22 are specified by eqn (32.6) in
K" ~ which e =I (1)/1 (0) and parameter c depends on
22 22 •
the boundary conditions. For a sunp y sup-
I
ported shell c = 1, for a hinge~ shell w~ose.end
Note that the shell is assumed to be simply
cross sections cannot move ill the axIal drrec-
supported at x =0 and x =L (see Fig. 32.18).
tion c = 1.5, and for a shell with one end
If transverse shear deformation is not taken
hinged and the other end free c = 0.6.
into account, then
Buckling pressure for an infinitely long

N, ~ i W~' )-c,,] shell is

3D22
The critical load, corresponding to the general qc = R.3
mode of buckling of a thin simply supported
Finally, note that the derivation for equations
orthotropic shell, can be found as
presented in this chapter can be found else-
,,\4 ,,\2
where3.
Nc =Du"\~ + D22 ,,\2n +
m R2"t2
("\t,,t2 )
_ n +~
n Bu B33
 References 757

REFERENCES 2. Whitney, J.M. 1987. Structural Analysis of

Laminated Anisotropic Plates. Lancaster,
1. Tamopolskii, Yu.M. and Kincis, T.Ya. 1985. Static
Pensylvania: Technomic Publishing Co., Inc.
Test Methods for Composites. New York: Van 3. Vasiliev, V.v. 1993. Mechanics of Composite
Nostrand Reinhold. Structures. Washington: Taylor & Francis.
DESIGN ALLOWABLES SUBSTANTIATION 33
Christy Kirchner Lapp

33.1 INTRODUCTION Strength properties for compression, tension

Designing with composite materials requires and shear must be determined. Both elastic
knowledge of a significantly greater number and strength ~roperties can be influenced by
numerous vanables such as the fiber matrix
of properties than for conventional isotropic
fiber volume and processing method.' The tes~
metals. The selection of lamina and laminate
method used to determine the allowable can
allowables can be critical in the analysis of a
composite structure. However, composite also affect the property. In addition, some
allowa~les cannot be readily tested, especially
design allowables may not always be obtained
properties through-the-thickness of the com-
from a single source of data. Several references
must often be consulted to determine all the posite. Often, complete characterization of a
fiber /matrix system may not be available and
p.rope~es in the. necessary directions, espe-
~e engineer must estimate or assume proper-
Clally if several fibers or matrices are being
considered in the design. This can be a time- ties. It can be expensive and time-consuming
consuming effort, especially during the initial to completely characterize a fiber/matrix sys-
design phase. In addition, some organizations tem during the initial design phase, so it is
may not have easy access to all the necessary important for the engineer to have allowables
references required to collect the data. This based on reliable data and to understand the
chapter assembles lamina data from numerous limitations.
Thi~ chapter concentrates on providing the
sources so that the engineer may have a single
reference point for initial design and analysis matenal database and techniques for assem-
bling the necessary composite allowables for
~f composite structures. A broad range of
fibers has been included since composites are preliminary design. Final design allowables
increasingly being applied outside the aero- may require additional testing. The allowables
space community. provided in this chapter should be considered
preliminary design values. Allowables for
Design allowables in this section are for
both two dimensional (2-D) and three dimen-
both elastic and strength properties. Elastic
sional (3-D) properties are included. As a
properties are necessary for laminate design or
starting point, lamina allowables for com-
the analysis of composite structures. These
monly used fiber / matrix systems are
properties include elastic moduli and
Poisson's ratios. Strength properties are described and listed. The references for these
allowables are included. Several references
required to predict laminate strengths or per-
form a failure analYSis of the structure. may be listed for a single material. Methods
for estimating properties when data is not
available or testing cannot be readily per-
formed are also defined. The effects of
Handbook of Composites. Edited by S.T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 processing methods on allowables and methods
 Lamina allowables 759

for adjusting lamina properties for these vari- 33.3 LAMINA ALLOWABLES
ables is described. The intent is to provide a
Lamina allowables may be used in a laminated
baseline for design allowables which can then
plate code to predict laminate elastic and
be expanded with additional ~~sting ~at~ as strength properties, or they may be used
required or modified for a specific application.
directly in a finite element analysis code. Table
33.1 defines the lamina properties and the
33.2 NOMENCLATURE FOR DESIGN basis for each property. Most of the properties
ALLOWABLES are based on test data for the 2-D properties.
The 3-D properties typically represent calcu-
The nomenclature used to describe composite
lated values based on the equations shown in
lamina and laminate properties is not consis-
Table 33.1.
tent within the industry. For this chapter, the
Properties for commonly used fiber and
lamina properties parallel to the fiber are spec-
epoxy resin systems are included and the con-
ified with a 1 and the lamina properties
struction is assumed to be continuous fiber
transverse to the fiber are designated with a 2.
reinforcement. A wide variety of reinforcing
Figure 33.1 illustrates this nomenclature for
fibers is included since composites may be
composite lamina and laminate properties.
used in variety of applications, each with very
The properties in the 1-2 plane are often
different design requirements. Glass, polyeth-
referred to as the in-plane lamina properties
ylene, aramid and graphite fibers.are cove~d.
whereas the properties through-the-thickness
Graphite fibers include standard, mtermedlate
(out-of-plane) of the composite are designated
and high modulus types. The fiber/resin sys-
with a 3. The 1, 2 and 3 directions can be
tems have very different properties and the
referred to as the longitudinal, transverse, and
actual composite application will dictate the
through-the-thickness properties respectively.
selection. For example, glass fibers are less
expensive and more impact resistant than
z
property definition

Lamina property definition

Fig. 33.1 Nomenclature for lamina and laminate properties.

760 Design allowables substantiation

Table 33.1 Definition of lamina properties and equations used to calculate material properties

Lamina material Definition Equation used to calculate

properties material property
Elastic
Elastic modulus in the fiber direction Property based on test data
Elastic modulus transverse to the fiber direction Property based on test data
Elastic modulus through-the-thickness Transverse isotropy: E3 = E2
Shear modulus in the 1-2 plane Property based on test data
E3
G23 Shear modulus in the 2-3 plane G23 = 2(1 + v23 )
Shear modulus in the 1-3 plane Transverse isotropy: G13 = G12
Poisson's ratio in 1-3 plane Property based on test data

Poisson's ratio in 2-3 plane

Poisson's ratio in 1-3 plane Transverse isotropy: 1'13 = 1'12

Strength
Tensile strength in the fiber direction Property based on test data
Compressive strength in the fiber direction Property based on test data
Tensile strength transverse to the fiber Property based on test data
Compressive strength transverse to the fiber Property based on test data
Tensile strength through-the-thickness u3 = u2
Compressive strength through-the-thickness -a3 = -a2
Shear strength in 1-2 plane (in-plane) Property based on test data
Shear strength in the 1-3 plane (interlaminar) Property based on test data
Shear strength in the 2-3 plane (interlaminar) 'r23 = 'r13

graphite fibers, but have a higher density and upon the type of analyses being performed or
lower modulus. Since the lamina properties the analysis code being used. If a strength or
are assumed to be used as preliminary design failure analysis is performed, then the follow-
and analysis parameters, the effects of temper- ing strength allowables may be required; 0'1'
ature, environment and fatigue are not (J2' -crl , -(J2 and 'tIl' Tables 33.2 and 33.3 include
considered. However these conditions must be these elastic and strength properties, which
considered during the final design develop- were compiled from various sources, such as
ment. military standards, material supplier data and
published literature. References are included
for each property so that the engineer may
33.3.1 TWO-DIMENSIONAL (2-D) LAMINA
directly consult a particular reference if further
PROPERTIES
information is required. These references are
The elastic lamina properties required for a listed in Tables 33.4 and 33.5.
composite 2-D analysis are typically El , E2, G12, The lamina properties for glass, polyethyl-
and V12 • The specific properties may depend ene and aramid fibers are listed in Table 33.2.
 Lamina allowables 761

Polyethylene and aramid fibers are more com- these properties are difficult to test and data is
monly known by their trade names as Spectra not always readily available. The 3-D lamina
and Kevlar. Kevlar 29 and 49 are included for properties listed in Tables 33.2 and 33.3 are
aramid properties. The polyethylene fibers estimated based on 2-D properties. The fol-
include Spectra 900 and 1000. E-glass and S- lowing section describes the methods for
glass (trade names) are included for glass calculating lamina properties in the 3-direc-
fibers. Table 33.3 lists lamina properties for tion when data is not available.
standard, intermediate and high modulus
graphite fibers. Data for non-USA produced
33.3.3 ESTIMATING LAMINA PROPERTIES
fibers is not included.
WHEN DATA IS NOT AVAILABLE
The type of material system is also indi-
cated in Tables 33.2 and 33.3. Design Lamina properties through-the-thickness (3-
allowables for a prepreg material system ver- direction) are often not readily available,
sus a wet filament wound system may result although they may be required to perform an
in different properties. It is important to con- analysis. When data is not available, these
sider the processing method when selecting properties can be determined by assuming
the properties to be used in a design. One pro- that the lamina is transversely isotropic. For a
cessing method may result in a lower moduli transversely isotropic lamina, the properties in
or strength than another. The effects of differ- the 2- and 3-directions are assumed to be the
ent processing methods on design allowables same. Thus the following equations may be
is discussed in further detail at the end of this used to determine elastic properties in the 3-
chapter. direction1:
The majority of the 2-D properties is based
on test data, not micromechanics equations.
The test data is typically based on 'thin speci-
mens' (typically less than 2.54 mm (0.10 in)
thick). Properties based on thin specimens E3
may not represent those for thick composites G
23
= ~.,----.,--
2(1 + v~
(typically greater than 6.35 mm (0.25 in) thick).
The exact definition of a thick composite is not
V I3 = V 12
consistent within the composites community.
Thick composite structures may have proper- The transverse Poisson's ratio, v23' can be
ties lower than those of thin composites, so the determined from the following relationship 2:
engineer may need to perform additional test-
ing for certain applications.

33.3.2 THREE-DIMENSIONAL (3-D) LAMINA

PROPERTIES
Initially the majority of analyses performed on where vf is the fiber Poisson's ratio, Vf is the
composite structures were two dimensional. fiber volume fraction, vm is the matrix
However with advancement of finite element Poisson's ratio, and Em is the matrix elastic
analysis programs and as new applications for modulus. Therefore to calculate certain prop-
composites arise, more analyses are being per- erties in the 3-direction, the engineer needs to
formed for the 3-~ case. Thus through-the know certain fiber and matrix properties, and
-thickness lamina properties have become nec- these have also been listed in Tables 33.2 and
essary in performing certain analyses. Often 33.3. For initial design purposes, it is simplest
762 Design allowables substantiation

to assume that the fiber and matrix are and lower void content than a wet filament
isotropic, although certain fibers are consid- wound part. This can affect properties such as
ered to be anisotropic. The fiber or resin shear the elastic modulus (E l ) or tensile strength (aJ
modulus can be calculated by: It is important that the engineer understand
E the limitations of the selected processing
G = 2(1 + v) method and adjust the design allow abies
accordingly. Processing parameters which can
The following equations may be used to esti- affect the lamina allowables are fiber volume
mate strength properties in the 3-direction. fraction and void content. Design allowables
should be modified if the process used to
determine the lamina properties deviates from
the intended process for the final composite
-{J3 =-{J2
part; this ensures that unrealistic properties
The equations listed above should be consid- are not being used to design the part.
ered a starting point for estimating 3-D lamina There are numerous fabrication methods
properties when actual test data is not avail- applicable for composites. The methods which
able. The values may need to be verified by will be discussed are those that apply to con-
testing as the design progresses. tinuous fiber reinforcement. These include
filament winding, hand lay-up and resin
transfer molding (RTM). The method of cure
33.4 LAMINATE ALLOWABLES
can also affect the lamina properties. For
Laminate elastic properties and strengths can example, one part can be filament wound with
be determined by testing or by using a lami- prepreg tow and another can be hand layed up
nated plate code. During the initial design with prepreg tape. These parts have different
phase, laminate allowables are typically deter- fabrication methods, but they may be cured in
mined by using a laminated plate code. This is the same manner; in an autoclave with vac-
especially true if the composite lay-up devi- uum and pressure. It is very likely that these
ates from a 'standard' lay-up, such as a two parts would have similar properties and
quasi-isotropic laminate ([90, +45, -45, 0]). require no adjustment of lamina properties.
There is typically more test data available for a However, if the filament wound part was wet
quasi-isotropic lay-up than any other lay-ups. wound and cured in an oven without vacuum
Laminate testing is often performed after com- or pressure, then the lamina properties would
pletion of the initial design, material selection need to be adjusted if they were based on
and composite lay-up has been decided. properties derived from testing using prepreg
Laminate testing would be performed to con- tape. A wet filament wound part typically has
firm predictions and processing effects. a lower fiber volume fraction and higher void
content. Thus in determining if lamina proper-
ties need to be modified, the engineer must
33.5 EFFECTS OF PROCESSING
consider the complete method of processing,
VARIABILITIES ON DESIGN ALLOWABLES
including the raw material and cure method,
The actual fabrication method used to build a not just the automated or manual process
composite structure can have an impact on the which is being used to fabricate the part.
design allowables. Some processing methods In general, if a part is wet filament wound,
can result in a higher fiber volume fraction it will possess a lower fiber volume fraction
and lower void content than others. For exam- and a higher void content than a hand-layed
ple, an autoclave cured part using prepreg up part using prepreg tape and cured in an
tape will typically have a higher fiber volume autoclave. Wet filament wound parts cured in
 References 763

an oven without vacuum and pressure typi- E1 (wet filament wound)

cally have a fiber volume fraction between
0.55 and 0.60 with a void content between 1 Vf (wet filament wound)
= ----=--------E (prepreg tape)
and 5%. Parts fabricated from prepreg tow or Vf (prep reg tape) 1
tape, which are cured in an autoclave with
Adjusting properties by the ratio of fiber vol-
vacuum and pressure will typically contain
ume fraction is applicable for modulus,
fiber volume fractions between 0.60 and 0.65.
tensile and compressive strength, but does not
Parts which are fabricated using the RTM
serve well for Poisson's ratios which would
process will typically possess fiber volume
require micromechanics. The Poisson's ratio
fractions of approximately 0.50. Determining
can be calculated based on the following
lamina properties for RTM parts is particularly
equation1:
difficult since the preform is usually woven
and properties are not readily available. Also V 12 = VfVf + vm(l- Vf)
weaving in some conditions may slightly
where vf is the fiber Poisson's ratio, Vf is the
degrade the properties.
fiber volume fraction and vm is the matrix's
The adjustment of lamina properties should
Poisson's ratio.
focus on the 2-D elastic properties; E1, E2, G12,
V 12 and the 2-D strength properties; 01' --0'1' O 2,
-P2 and i 12 • The simplest method is to adjust REFERENCES
the desired property by multiplying the value 1. Whitney, J., Daniel, 1. and Pipes, B., 1984.
by the ratio of the fiber volume for the selected Experimental Mechanics of Fiber Reinforced
processing method to the fiber volume listed Composite Materials. Brookfield Center,
in Tables 33.2 and 33.3. For example, if one has Connecticut: The Society for Experimental
properties based on prepreg tape which has Mechanics.
been autoclaved cured and wants to adjust 2. VInson, J. and Sierakowski, R. 1987. The Behavior
these properties for a wet filament wound part of Structures Composed of Composite Materials.
Boston: Kluwer Academic Publishers.
the following calculation would be used to
adjust the longitudinal elastic modulus (E 1):
764 Design allowables substantiation

Table 33.2 Lamina properties for glass, aramid and polyethylene fibers in epoxy matrices

Properties E-glass/ 52-glass/ 52-glass/ Kevlar

DER332 DER 332 XP251S 29/934
Material system Wet wound Wet wound Prep reg Prepreg
Fiber type E-Glass S-Glass 52-Glass Aramid
Supplier Owens Owens Owens DuPont
Resin type Epoxy Epoxy Epoxy Epoxy
Supplier Dow Dow 3M Fiberite
Fiber volume 60% 60% 60% 58%
Composite density, g/ em3 2.05 1.98 1.98 1.38
(lb/in3) (0.074) (0.072) (0.072) (0.050)
Lamina elastic properties
E1, GPa 48 54 51 54
(psi x 106) (7.0) (7.9) (7.5) (7.9)
E2' GPa 12 16 17 5
(psi x 106) (1.8) (2.3) (2.5) (0.7)
E3, GPa 12 16 17 5
(psi x 106) (1.8) (2.3) (2.5) (0.7)
G12,GPa 6 7 7 2
(psi x 106) (0.84) (0.98) (0.98) (0.24)
GwGPa 5 6 7 2
(psi x 106) (0.70) (0.89) (0.95) (0.24)
G1y GPa 6 7 7 2
(psi x 106) (0.84) (0.98) (0.98) (0.24)
V 12 0.19 0.25 0.25 0.40
V23 0.26 0.32 0.32 0.47
vl3 0.19 0.25 0.25 0.40
Lamina strength properties
° 1
Tension, MPa 1020 1613 1779 1069
(psi x 103) (148) (234) (258) (155)
Compression, MPa -490 -462 -641 -272
(psi x 103) -(71) -(67) -(93) -(39)
°2
Tension, MPa 8 39 58 9
(psi x 103) (Ll) (5.6) (8.4) (1.3)
Compression, MPa -78 -103 -186 -130
(psi x 103) -(11) -(15) -(27) -(19)
°3
Tension, MPa 8 39 58 9
(psi x 103) (Ll) (5.6) (8.4) (1.3)
Compression, MPa -78 -103 -186 "":130
(psi x 103) -(11) -(15) -(27) -(19)
TI2
MPa 23 28 75 37
(psi x 103) (3.3) (4.0) (10.9) (5.3)
 Tables 765

Kevlar Kevlar 49 K.evlar Spectra Spectra

49/934 /DER332 149/934 900/EPON 826 1000/EPON 826
Pprepreg Wet Wound Prepreg Wet Wound Wet Wound
Aramid Aramid Aramid Polyethylene Polyethylene
DuPont DuPont DuPont Allied Allied
Epoxy Epoxy Epoxy Epoxy Epoxy
Fiberite Texaco Fiberlite Shell Shell
58% 60% 58% 55% 55%
1.38 1.35 1.38 1.12 1.12
(0.050) (0.049) (0.050) (0.040) (0.040)

72 82 106 31 50
(10.5) (11.9) (15.4) (4.5) (7.3)
5 5 6 4 1
(0.7) (0.7) (0.9) (0.5) (0.1)
5 5 6 4 1
(0.7) (0.7) (0.9) (0.5) (0.1)
2 2 2 1 1
(0.24) (0.26) (0.24) (0.21) (0.10)
2 2 2 1 0
(0.24) (0.27) (0.32) (0.21) (0.05)
2 2 2 1 1
(0.24) (0.26) (0.24) (0.21) (0.10)
0.41 0.31 0.34 0.32 0.28
0.48 0.38 0.42 0.40 0.36
0.41 0.31 0.34 0.32 0.28

1151 1848 1400 1110 1200

(167) (268) (203) (161) (174)
-281 -221 -262 -79 -72
-(41) -(32) -(38) -(12) -(11)

12 8 13 8 7
(1.7) (1.1) (1.9) (1.2) (1.0)
-134 -40 -114 -138 -138
-(19) -(06) -(17) -(20) -(20)

12 8 13 08 07
(1.7) (1.1) (1.9) (1.2) (1.0)
-134 -40 -114 -138 -138
-(19) -(06) -(17) -(20) -(20)

43 24 49 24 17
(6.3) (3.5) (7.1) (3.5) (2.5)
Continued on next page
766 Design allowables substantiation

Table 33.2 continued Lamina properties for glass, aramid and polyethylene fibers in epoxy matrices

Properties E-glass/ S2-glass/ S2-glass/ Kevlar

DER 332 DER332 XP251S 29/934
f13
MPa 66 66 77 34
(psi x lOS) (9.5) (9.5) (11.1) (5.0)
f23
MPa 66 66 77 34
(psi x lOS) (9.5) (9.5) (11.1) (5.0)
Constituent properties
Fiber
Ef' GPa 72 87 87 83
(psi x 1()6) (10.50) (12.60) (12.60) (12.00)
vf 0.09 0.18 0.18 0.44
Gf' GPa 33 37 37 29
(psi x 1()6) (4.8) (5.3) (5.3) (4.2)
Tensile strength, MPa 3103 3792 3792 3620
(psi x 1(3) (450) (550) (550) (525)
Density, g/ cm3 2.60 2.49 2.49 1.44
(lb/in3) (0.094) (0.090) (0.090) (0.052)
Resin
E" GPa 3.4 3.4 3.4 4.1
(psi X 1(6) (0.49) (0.49) (0.49) (0.60)
vr 0.35 0.35 0.35 0.35
G" GPa 1.2 1.2 1.2 1.5
(psi x 106) (0.18) (0.18) (0.18) (0.22)
Tensile strength, MPa 64 64 64 83
(psi x 1(3) (9.3) (9.3) (9.3) (12)
Density, g/cm3 1.22 1.22 1.22 1.30
(lb/in3) (0.044) (0.044) (0.044) (0.047)
 Tables 767

Kevlar Kevlar 49 Kevlar Spectra Spectra

49/934 /DER332 149/934 900/EPON 826 1000/EPON 826

50 50 38 23 23
(7.2) (7.2) (5.5) (3.4) (3.4)

50 50 38 23 23
(7.2) (7.2) (5.5) (3.4) (3.4)

124 124 172 117 172

(18.00) (18.00) (25.00) (17.00) (25.00)
0.45 0.28 0.33 0.30 0.22
43 48 65 45 70
(6.2) (7.0) (9.4) (6.6) (10.2)
3620 3620 3448 2586 2992
(525) (525) (500) (375) (434)
1.44 1.44 1.47 0.97 0.97
(0.052) (0.052) (0.053) (0.035) (0.035)

4.1 3.4 4.1 2.8 2.8

(0.60) (0.49) (0.60) (0.40) (0.40)
0.35 0.35 0.35 0.35 0.35
1.5 1.2 1.5 1.0 1.0
(0.22) (0.18) (0.22) (0.15) (0.15)
83 64 83 83 83
(12) (9.3) (12) (12) (12)
1.30 1.22 1.30 1.30 1.30
(0.047) (0.044) (0.047) (0.047) (0.047)
768 Design allowables substantiation

Table 33.3 Lamina properties for graphite fibers in epoxy matrices

Properties AS4/3501-6 IM6/3501-6 IM7/3501-6 IMB/3501-6

Material system Prepreg Prepreg Prepreg Prepreg
Fiber type Graphite Graphite Graphite Graphite
Supplier Hercules Hercules Hercules Hercules
Resin type Epoxy Epoxy Epoxy Epoxy
Supplier Hercules Hercules Hercules Hercules
Fiber volume 60% 60% 60% 60%
Composite density,
g/cm3 1.58 1.55 1.57 1.58
(lb/in3) (0.057) (0.056) (0.057) (0.057)
Lamina elastic properties
EI , GPa 143 159 159 186
(psi X 106) (20.7) (23.0) (23.0) (27.0)
E2' GPa 10 10 10 10
(psi X 106) (1.4) (1.4) (1.4) (1.4)
E31 GPa 10 10 10 10
(psi X 106) (1.4) (1.4) (1.4) (1.4)
GI2,GPa 6 5 5 6
(psi X 106) (0.85) (0.71) (0.72) (0.80)
GwGPa 3 3 3 3
(psi X 106) (0.41) (0.41) (0.41) (0.41)
G13'GPa 5 5 5 5
(psi X 106) (0.68) (0.68) (0.68) (0.68)
V l2 0.30 0.30 0.30 0.30
V23 0.52 0.52 0.52 0.52
V 13 0.30 0.30 0.30 0.30
Lamina strength properties
°1
Tension, MPa 2172 2413 2620 2689
(psi X 103) (315) (350) (380) (390)
Compression, MPa -1558 -1655 -1862 -1931
(psi X 103) -(226) -(240) -(270) -(280)

°2
Tension, MPa 54 59 59 59
(psi X 103) (7.8) (8.6) (8.6) (8.6)
Compression, MPa -186 -186 -186 -186
(psi X 103) -(27) -(27) -(27) -(27)

°3
Tension, MPa 59 59 59 59
(psi X 103) (8.5) (8.6) (8.6) (8.6)
Compression, MPa -186 -186 -186 -186
(psi X 103) -(27) -(27) -(27) -(27)
 Tables 769

T300/520B T40/1962 T50/1962 P55/1962 P75/1962

Prepreg Prepreg Prepreg Prepreg Prepreg
Graphite Graphite Graphite Graphite Graphite
Amoco Amoco Amoco Amoco Amoco
Epoxy Epoxy Epoxy Epoxy Epoxy
Fiberite Amoco Amoco Amoco Amoco
62% 62% 62% 62% 62%

1.60 1.60 1.72 1.72

(0.057) (0.058) (0.058) (0.062) (0.062)

141 172 241 241 338

(20.5) (25.0) (35.0) (35.0) (49.0)
9 10 7 8 7
(1.2) (1.5) (1.1) (1.1) (1.0)
9 10 7 8 7
(1.2) (1.5) (1.1) (1.1) (1.0)
6 7 6 5 6
(0.92) (1.00) (0.84) (0.79) (0.85)
3 4 3 3 3
(0.41) (0.54) (0.39) (0.39) (0.36)
5 7 6 5 6
(0.68) (1.00) (0.84) (0.79) (0.85)
0.30 0.33 0.28 0.34 0.30
0.52 0.40 0.35 0.41 0.37
0.30 0.33 0.28 0.34 0.30

1524 3241 1413 931 965

(221) (470) (205) (135) (140)
-1482 -1724 -965 -510 --441
-(215) -(250) -(140) -(74) -(64)

36 69 37 33 33
(5.2) (10.0) (5.3) (4.8) (4.8)
-159 -159 -159 -159 -159
-(23) -(23) -(23) -(23) -(23)

36 69 37 33 33
(5.2) (10.0) (5.3) (4.8) (4.8)
-159 -159 -159 -159 -159
-(23) -(23) -(23) -(23) -(23)

Continued on next page

770 Design allowables substantiation

Table 33.3 continued Lamina properties for graphite fibers in epoxy matrices

Properties AS4/3501-6 IM6/3501-6 IM7/3501-6 IMB/3501-6

T12
MPa 87 85 96 80
(PSi X 1(3) (12.6) (12.3) (13.9) (11.6)
T13
MPa 124 121 121 131
(psi X 103) (18.0) (17.5) (17.5) (19.0)

T23
MPa 94 94 94 94
(psi X 103) (13.6) (13.6) (13.6) (13.6)
Constituent properties
Fiber
Ef' GPa 234 276 276 303
(psi X 1()6) (34.0) (40.0) (40.0) (44.0)
vf 0.26 0.26 0.26 0.26
Gf' GPa 93 109 109 120
(psi X 1()6) (13.5) (15.9) (15.9) (17.5)
Tensile strength, MFa 3930 5102 5309 5447
(psi X 1(3) (570) (740) (770) (790)
Density, g/ cm3 1.80 1.74 1.77 1.80
(lb/in3) (0.065) (0.063) (0.064) (0.065)
Resin
Er' GPa 4.4 4.4 4.4 4.4
(psi X 1()6) (0.64) (0.64) (0.64) (0.64)
vr 0.36 0.36 0.36 0.36
Gr'GPa 1.6 1.6 1.6 1.6
(psi X 1()6) (0.24) (0.24) (0.24) (0.24)
Tensile strength, MPa 69 69 69 69
(psi X 103) (10) (10) (10) (10)
Density, g/ cm3 1.26 1.26 1.26 1.26
(lb/in3) (0.046) (0.046) (0.046) (0.046)
 Tables 771

T300/520B T40/1962 T50/1962 P55/1962 P75/1962

77 97 63 50 44
(11.2) (14.0) (9.2) (7.2) (6.4)

115 97 69 62 55
(16.7) (14.0) (10.0) (9.0) (8.0)

94 97 69 62 55
(13.6) (14.0) (10.0) (9.0) (8.0)

231 283 393 379 517

(33.5) (41.0) (57.0) (55.0) (75.0)
0.27 0.32 0.24 0.33 0.27
91 107 159 142 204
(13.2) (15.6) (23.0) (20.6) (29.5)
3241 5654 2413 1724 2069
(470) (820) (350) (250) (300)
1.77 1.80 1.80 1.99 1.99
(0.064) (0.065) (0.065) (0.072) (0.072)

3.9 3.7 3.7 3.7 3.7

(0.56) (0.54) (0.54) (0.54) (0.54)
0.35 0.35 0.35 0.35 0.35
1.4 1.4 1.4 1.4 1.4
(0.21) (0.20) (0.20) (0.20) (0.20)
50 66 66 66 66
(7.3) (79.6) (9.6) (9.6) (9.6)
1.27 1.27 1.27 1.27 1.27
(0.046) (0.046) (0.046) (0.046) (0.046)
772 Design allowables substantiation

LAMINA PROPERTY REFERENCES D V23was calculated based on the following equa-

A Weight density was calculated based on the fol- tion:
lowing relationship:

Pcomposire =P f (Vf) -Pr (1- Vf)

Pf = fiber density
V f = fiber volume fraction
P r = resin density
AAAS4 and IM6 fiber properties are based on DD Hercules data for IM7/ 3501-6 from the Graphite
Hexcel supplied data which was based on tow Fiber Products Handbook, based on mechanical
test data for the fiber modulus and strength. test data.
The fiber Poisson's ratio was 'back' calculated E Assumed:
based on the composite Poisson's ratio (V I2 ), the
matrix Poisson's ratio and the fiber volume
U3 = u2 and -a3 = -a2
(see reference H). EE IM7 and IM8 fiber properties are based on
B L.L. Clements and R.L. Moore, Composite prop- Hercules supplied data (Graphite Fiber Products
erties for E-glass fibres in a room temperature Handbook) which was based on tow test data
curable epoxy matrix, Composites, 1978, 9(2), for the fiber modulus and strength. The fiber
93-99. Properties for t I3 were set equal to val- Poisson's ratio was 'back' calculated based on
ues for S2-g1ass/DER 332 since no data was the composite Poisson's ratio (V I2 ) , the matrix
reported. Poisson's ratio and the fiber volume (see refer-
BB Hercules supplied data for IM6/3501-6 from the ence H).
Graphite Fiber Products Handbook based on F The following was assumed:
mechanical test data.
C Composite is assumed to be transversely
isotropic. A transversely isotropic composite is FF Hercules data for IM8/3501-6 from the Graphite
a material which exhibits a special case of Fiber Products Handbook, based on mechanical
orthotropy, whereby the properties are identi- test data.',
cal in two orthotropic dimensions, but not the G Owens Corning fiber data and short beam
third. The properties are the same in both shear data.
transverse directions, but not in the longitudi- GG Properties based on test data listed in BASF
nal direction. The following equations apply Hexcel Technical Information handbook.
for transversely isotropic materials: H Equation for calculating the fiber Poisson's
ratio (v f ):
E3 = E2

HH Amoco data for T300 fibers. Actual test

method for fiber modulus and strength was not
defined. The fiber Poisson's ratio was 'back'
calculated based on the composite Poisson's
G23 = 2(1 + v23 ) ratio using the equation in reference H.
I Fiber and resin are assumed to be isotropic,
CC Properties were set equal to those for
therefore Gf or Gr is calculated as follows:
AS4/3501-6.
Ef
 Lamina property references 773

Technology Services, Boeing Commercial

Airplane Co.
Gr ::; 2(1 + vr )
R No data was listed in DuPont literature for
IT Properties based on test data supplied by G12, therefore values for Kevlar 29/934, Kevlar
Amoco for the T40/1962 system. Amoco sup- 49/934, and Kevlar 149/934 were set equal to
plied data for the 1962 resin system. those for Kevlar 49. Hexcel rubber-toughened
J Resin properties from publication by Texaco, epoxy system based on a paper by S.R
Huntsman Chemical Co. 'Jeffamine', resin Swanson, G.R Toombes, and S.W. Beckwith,
properties are based on 100 parts of epoxy resin In-Plane Shear Properties of Composites Using
and 45 parts of Jeffamine T-403. The DER 332 Torsion Tests of Thin-Wall Tubes, 29th National
epoxy resin was cured with Jeffamine T-403. SAMPE Symposium, April 3-5, 1984.
IT Properties were set equal to T300/5208 value. 5 DuPont supplied data for Kevlar 29, Kevlar 49,
KK Amoco data for T40 and T50 from technical and Kevlar 149 based on tow tests (ASTM
D2343). The fiber Poisson's ratio was 'back' cal-
information sheets. Test method for fiber mod-
ulus and strength was not specified. The fiber culated based on the composite Poisson's ratio
(V I2 ), the resin Poisson's ratio and the fiber vol-
Poisson's ratio was 'back' calculated based on
the composite Poisson's ratio (V I2 ) , the resin ume (see reference D).
Poisson's ratio and the fiber volume (see refer- T Fiberite data for 934 resin system from Fiberite
ence H). Material Handbook.
L H. Hahn, D. Hwaug, H. Chang, S. 10, Flywheel U Values for Kevlar 49/DER 332 set equal to
Materials Technology: Design Data Manual for those for Kevlar 49/Epoxy XD7575.03-XD7114-
Composite Materials, UCRL-15365 Volume 1, Tonox 60-40 from Reference 6 of this list.
P.O. 6641009, Lawrence Livermore Laboratory, V D.F. Adams, RS. Zimmerman and H.W.
Chang, Properties of Polymer-Matrix
July, 1981.
LL Amoco test data for T50/1962 Composites Incorporating Allied A-900
M Assumed Polyethylene Fiber, SAMPE J., September/
October, 1985, pp. 44-48. Note: The modulus of
Spectra composites is much lower than
expected from the rule-of-mixtures relation-
MMAmoco test data for P55/1962. ship. A possible explanation is that the Spectra
N Mil Handbook 17. fiber modulus is a function of strain rate. For
NN Amoco data for P55 from technical information example a single fiber tested at 100%/min
sheets. Test method for fiber modulus and strain rate exhibited a modulus of 17 msi ver-
strength was not specified. The fiber Poisson's sus 11 msi for a 8%/min strain rate.
ratio was 'back' calculated based on the com- W H.W. Chang, L.e. Lin, A. Bhatnagar, Properties
posite Poisson's ratio (V I2 ) , the resin Poisson's and applications of composites made of polyethylene
ratio and the fiber volume (see reference H). fibers, 31st Intern. SAMPE Symp., April 7-10,
o Properties were set equal to S2-g1ass/DER332. 1986.
00 Amoco data for P75/1962. X for Spectra 1000 set equal to the value for
T I3
P Properties for XP251S epoxy were set equal to Spectra 900.
DER 332 since no data was available. Y Hercules supplied test data for AS4/3501-6
PP Amoco data for P75 which was based on tow determined by independent firms; Delsen Labs
test data for fiber modulus and strength. The and McDonnell Aircraft Company. Hercules
fiber Poisson's ratio was 'back' calculated supplied data for 3501-6 resin. All data is listed
based on the composite Poisson's ratio (VI2) , in the Hercules Graphite Fiber Products
the resin Poisson's ratio and the fiber volume Handbook.
(see equation D). Z RY. Kim, F. Abrams and M. Knight,
Q DuPont supplied data for Kevlar 29/934, Mechanical characterization of a thick compos-
Kevlar 49/934, and Kevlar 149/934. Laminates ite laminate, Proc. Amer. Soc. Composites, 3rd
were fabricated and tested by Boeing Technical Conference, 1988, pp. 711-718.
774 Design allowables substantiation

Table 33.4 References for lamina properties of glass, aramid and polyethylene fibers in epoxy matrices

Properties E-glass 52-glass 52-glass Keviar

DER332 DER332 XP2515 29/934

Material system
Fiber type
Resin type
Fiber volume
Composite density A A A Q
Lamina elastic properties
E1 B L N Q
E2 B L N Q
E3 C C C C
G12 B L 0 R
G23 C C C C
G13 C C C C
V 12 B L 0 Q
V23 D D D D
V 13 C C C C
Lamina strength properties (psi)

°1 Tension B L N Q
Compression B L N Q
°2 Tension B L N Q
Compresson B L N Q
°3 Tension E M M M
Compression E M M M

t'12 B L N Q
t'13 B G N Q
t'23 F F F F
Constituent properties
Fiber
Ef G G G S
vf H H H H
Gf I I I I
Tensile strength G G G S
Density G G G S
Resin
Er J J p T
vr K K K K
Gr I I I I
Tensile strength J J p T
Density K K K T
 Tables 775

Kevlar Kevlar Kevlar Spectra Spectra

49/934 /DER332 149/934 900/EPON 826 1000/EPON
826

Q L Q A A

Q L Q V w
Q L Q V w
C C C C C
R L R V W
C C C C C
C C C C c
Q L Q V W
D D D D D
C C C C C

Q L Q V w
Q u Q V w

Q L Q V w
Q u Q K K

M M M M M
M M M M M
Q L Q V w
Q L Q V X
F F F F F

S 5 5 V W
H H H H H
I I I I I
5 S 5 V W
5 S 5 V W

T J T K K
K K K K K
I I I I I
T J T K K
T K T K K
776 Design allowables substantiation

Table 33.5 References for lamina properties of graphite fibers in epoxy matrices

Properties ASA/3501- IM6/ 1M7/3501- 1MB/

6 3501-6 6 3501-6
Material system
Fiber type
Resin type
Fiber volume
Composite density A A A A
Lamina elastic properties
E1 y SS DD FF
E2 Y SS CC CC
E3 C C C C
G 12 Y SS DD FF
G23 Z CC CC CC
G 13 Z CC CC CC
V 12 Y CC CC CC
V23 Z CC CC CC
V 13 Z C CC CC
Lamina strength properties (psi)

°1
Tension Y BB DD FF
Compression Y BB DD FF
°2
Tension Y BB CC CC
Compression Z CC CC CC
03
Tension Z CC E E
Compression Z CC E E
112 Y BB DD FF
1 13 Y BB DO FF
123 Z CC CC CC
Constituent properties
Fiber
Ef AA AA EE EE
VI H H H H
Gf I I I I
Tensile strength AA AA EE EE
Density AA AA EE EE
Resin
Er y Y Y Y
Vr y Y Y Y
Gr I I I I
Tensile strength Y Y Y Y
Density Y Y Y Y
 Tables 777

T300 T50/ P55/

5208 T40/1962 1962 1962 P75/1962

A A A A A

GG II LL MM 00
GG II LL MM 00
C C C C C
L II LL MM 00
CC C C C C
CC C C C C
L II LL MM 00
CC D D D D
C C C C C

GG II LL MM 00
GG II LL MM 00

L II LL MM 00
L IT IT IT IT
E E E E E
E E E E E
GG II LL MM 00
GG II LL MM 00
CC F F F F

HH KK KK NN pp
H H H H H
I I I I I
HH KK KK NN pp
HH KK KK NN pp

GG II II II II
K K K K K
I I I I II
GG II II II II
GG II II II II
MECHANICAL TESTS 34
Yu.M. Tarnopol'skii and v.L. Kulakov

34.1 STRUCTURAL HIERARCHY OF FIBROUS The test results are used to calculate the
COMPOSITES properties of materials with more complex
Fibrous composites are inhomogeneous mate- configurations of fiber arrangement and of
rials with multiple levels of structural scale. hybrids, i.e. materials with different reinforce-
The three levels of structural scale can be ments in the same lay-up.
arranged in a hierarchy. The characteristic For multilayered composites, the largest
dimensions for the three levels are: fiber diam- structural scale is the thickness of the lami-
nated plate that is equal to the sum of stacked
eter, lamina thickness and plate thickness. The
laminae and interleaves. The theory of lami-
most appropriate test methods and structural
analysis techniques are different for each level nated plates allows one to determine the
properties of the plate using the properties of
in the hierarchy. Test objectives and associated
problems are also different for each level. the monolayers and their stacking sequence. If
The smallest scale is the diameter of the the monolayers are part of a structural ele-
reinforcing fiber. The properties of the rein- ment, then the highest level of scale is the
forcing fiber and polymer matrix and their characteristic size of the object. The dimen-
interaction are studied in the field of microme- sions .of structural elements typically exceed
chanics. the thickness of the plate by several times. The
p~~perties of components are measured by tra-
The second level scale is the thickness of the
unidirectional lamina. Macromechanics ditional mechanical and physical test
methods1.
describes the properties of a monolayer under
loading at an angle to the fiber direction. A The results of the analyses or tests at the
monolayer is defined as a flat or curved ele- first level are used as input for the analyses at
ment of material composed of a polymeric the second level. The same relationship holds
~or the second and third levels. Upon convert-
matrix and reinforcement of the same type and
orientation throughout the layer. It is the basic mg each scale to a continuum at the next
structural element of laminated and fibrous higher scale, it is important that for each level
composites. The characterization of monolay- under consideration the number of elements
ers by mechanical test methods is given be sufficient, e.g. through the width and thick-
ness of the lamina or through the thickness of
particular emphasis in this chapter since test-
ing of anisotropic materials is a relatively the plate, so the transition from a discrete
novel and seldom studied field of mechanics. medium to a continuum is performed without
great error. This progression up the hierarchy
eventually leads to solutions of real life prob-
lems involving structural elements or
prototypes.
Handbook of Composites. Edited by S.T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7
 Problems of composite testing 779

34.2 PROBLEMS OF COMPOSITE TESTING systems of coordinates are introduced: the

axes of elastic symmetry in the material (1, 2,
34.2.1 HISTORY 3) and axes of loading (x, y, z for flat speci-
mens; (J, z and r for ring and tubular
The first reference to mechanical testing of a
specimens). It is preferable to use methods in
structural material for an engineering applica-
which the x, y, z axes (or (J, z, r) coincide with
tion is dated July 4, 1662. The objective of the
the I, 2, 3 axes.
tests was to compare the tensile strengths of
The majority of laminated and fibrous com-
cords made of Riga and Dutch yams. The
posites exhibit low interlaminar shear and
stronger in this contest was the thinner cord
transverse tension strengths. Shear strength is
manufactured in Riga. In the years that fol-
characterized by the relations between EJGxz
lowed, equipment and methods for testing
(shear stiffness) and a//rxzu (shear strength).
engineering materials, particularly metals,
Transverse tension and compression strengths
have reached a high degree of perfection and
perpendicular to the fibers are determined by
consistency. The appearance of composites
the relations Ex a tu / aztu, axtu / azCU, where Ex
/ Ez, x
and their ever-expanding use has once again
and Ez are the moduli of elasticity in the x and
made it necessary to improve mechanical test
z directions; Gxz is interlaminar shear modulus;
methods. Although significant progress has
a xu and a zu are strengths in the x and z direc-
been made, there are vast differences in their
tions; rxzu is shear strength in xz plane. The x
maturity. The methods differ primarily in the
and y axes are located in the fiber lay-up (rein-
degree to which they minimize extraneous
forcement) plane, the z-axis is perpendicular
stresses and strains. Although test methods
to this plane; the (t) and (c) designate tension
tend to become more complex as their accu-
and compression, respectively.
racy increases, economics must be considered
in their selection. Factors such as complexity
of specimen preparation, amount of material 34.2.3 UNIQUE REQUIREMENTS OF
required, and the requirements for specially COMPOSITE TESTING
designed equipment must be considered.
The anisotropy and unique structural proper-
ties of composite materials cause serious
34.2.2 DETERMINABLE VALUES difficulties. For example, a large number of
strength and elastic properties must be deter-
The purpose of mechanical tests is to deter- mined for complete characterization of the
mine the strength and elastic properties of a material. Since the number of determinable
material. However, only loads, displacements characteristics depends on the state of stress
and strains can be measured in a mechanical and the degree of anisotropy2,3, one should
test. The theory of elasticity for an anisotropic select the loading methods for which the
body is used to determine the desired proper- experimentally determinable values are most
ties of composites from these measurable simply related to the material characteristics.
quantities. It should be remembered that The selection of techniques for analyzing the
advanced fibrous composites with unidirec- data is critical as well as the determination of
tional, laminated or spatial fiber lay-ups are their range of validity. Since the composite
inhomogeneous, essentially anisotropic mate- analysis techniques are based on the theory of
rials. The customary terms, i.e. tension, elasticity for an anisotropic body, it is neces-
compression, shear and bending, are meaning- sary to consider the error in treating an
less without specification of the direction of inhomogeneous anisotropic medium as a con-
the load and its relationship to the axes of elas- tinuous anisotropic medium. For example, the
tic symmetry of the material. Therefore, two number of structural elements (fibers, lamina,
780 Mechanical tests

etc.) must be sufficiently large to support this of fiber waviness can cause the measured val-
approximation4,5.6. ues of longitudinal modulus of elasticity and
Once the general test method has been strength to be considerably lower than those of
selected, the details of the loading and the materials with ideally straight fibers. Fiber
sample geometry must be selected. For fibrous waviness also influences on the coefficient of
composites, the principal difficulties lie in the thermal expansion in the fiber direction. The
generation of a uniform stress field in a repre- modulus of elasticity perpendicular to the
sentative volume of material, i.e. the fiber direction and the in-plane shear modulus
elimination of end and edge effects. This is dif- are not significantly affected by fiber wavi-
ficult even for the most simple types of tests. ness.
The difficulties increase with increasing All of the aforementioned unique testing
degree of anisotropy, i.e. materials reinforced requirements apply to composites of a fibrous
by high-modulus or high-strength fibers and laminated structure. Additional difficul-
(boron, carbon and organic fibers). ties arise when spatially reinforced composites
End effects are primarily influenced by the are tested because the transverse strength and
method of fastening and loading of the speci- stiffness are derived from a rigid framework
men, the length of the grip section, and the rather than from a compliant matrix.
fiber orientation. The region involved in end
effects extends in the direction of the greatest
34.2.4 SUMMARY TABLES
stiffness of the material and increases with the
anisotropy of the material. The most common methods of testing com-
Edge effects are primarily influenced by the posites in tension, compression, torsion and
size and shape of the specimen, the fiber ori- bending are described in Tables 34.1-34.5. The
entation and the angle of specimen cutting. H high performance test fixtures designed
strength anisotropy is present, improper load- specifically for composite testing, their
ing and fastening can lead to changes in the description and recommended applications
failure mode and the resulting strength value. are given in Reference 7.
A most important consideration is the selec-
tion of the specimen width. The width must be
34.3 TEST SPECIMENS
large enough to avoid the effect of cut fibers at
free edges which is important for specimens of The important relationships between fabrica-
off-angle, angle-ply and cross-ply materials. tion methods, test methods and required
Edge effects are manifested as interlaminar specimen shapes are shown in Fig. 34.1.
stresses at the free edges of the specimen, the Specimens for mechanical testing are classified
direction and magnitudes of which depend on as flat specimens (bars and plates), rings (com-
the fiber lay-up. plete and segments) and tubes2.B.
Material quality also causes unique require- The specimens and test methods in Fig. 34.1
ments for testing composites. Quality cannot are used to characterize the monolayer. Flat
be ignored during testing because the material monolayers can be characterized with speci-
and structure are formed simultaneously. In mens that have a different fiber lay-ups but the
addition, composites are extremely sensitive same general, flat, long, narrow shape. To ade-
to mechanical and thermal history. Structural quately characterize wound monolayers, it is
imperfections, in particular porosity, waviness necessary to use both rings and tubes. Ring
and misalignment of fibers, require special specimens of a unidirectional fiber lay-up are
attention. The presence of porosity affects the used to assess characteristics in the fiber direc-
measurement of polymer matrix dominated tion. Tubular specimens with a 90° wind angle
properties, e.g. shear strength. Small amounts are used to measure properties perpendicular
Table 34.1 Tension and compression testing of flat specimens

Information about methods Tension (t) Compression (c) Off-axis tension

(a) (b) (c) (d) (e) (f)
fpi
x
81 P
Loading scheme

WmmD~ p
pI:
Determinable characteristics axtu(cu), Ext(C), v /)iC), v!C) Gzy
Specimen sizes L = 229mm; L = 140 mm; (") L=229mm;
L L L L L
-=22;-=70 -w = 9-15;
w t
-w = 18; - t = 100-225 (0°)
(ForCFRP)
L L L
- = 9' - = 225-300 (90°) - =225-300
w 't t
Standards or references ASTM D 3039; ISO 3268; ASTM D 3410; DIN 29971; ASTMD3039;
BSI 2782; Refs. 2, 14, 16 Refs. 2, 14, 16 Refs. 2, 14
Measurable values pt<C) ptu(cu) e t(c) e tIc) e tIc)
I Ix'y'Z P, efJ" e450' e9fJ'
Relationships for axtu(CU) = pt<c) /wt; Ext(C) = pt(c) /wte~(c); See Ref. 2, p. 174-180
calculations v tIc) = e f(c) / e tIc) v tIc) = et(c) / et(c)
yx y x'zx x x

Limitations: ~
!J>
......
Structural:
Lay-up 0°; 90°; 0° /90° 0° ]
Angle of cutting 0°; 90° 10° + 15°
Physical: For Ex; v: linear range of the curves no ~.
Geometrical: See Ref. 2, p. 54-56 L/w > 4(a); L/w > 5(b); L > Lcr (Ref. 14, p. 106) L/w = 15 (for GFRP)
Note: Pu is a load at specimen failure; I", is the critical length. ASIM D 3410 contains the specimen sizes for different types of composites ("). ~
......
Table 34.2 Tension and compression testing of ring specimens ~
IV

Pressure
Infort1Ultion about methods Half-discs Compliant ring Hydraulic system ~
Loading scheme ~
(a) "~) (d) [
• -t ...a
(e)e i(t

,. i

(c) pC (f)

Determinable characteristics 0otu(CU), Eotu(c)

Specimen sizes D j = 146 mm; w = 6.35 mm; t = 1.52 mm; 3.18 mm

Standards or references ASlM D 2290 (a). Refs. 2, 11. ASlM D 2586. Ref. 2.
Measurable values pt(c), ptu(cu), Ei 9t(c) pl(e), ptu(cu), Ei81(C)
Relationships for °0tu(cu) =ptu(cu) /2wt; ° 0 tu(cu) = ptu(cu) D;f2t;
calculations Eet(C) = pt(c) /2wt£01(e) E9tu(CU) = pt(c)D;f2tEi ot(C)
Limitations:
Structural:
Lay-up 0°; 90°; 0°/90°
Angle of cutting 0°; 90°
Physical: For E9t(c): (linear range of the curves) pt<c) vs. Eiot(c), pt(e) vs. £9t(c)
t 2
Geometrical: 0.08:S; R ~ 0.18 (for GFRP); L;::"3R (b)

Note: Pu is the load at specimen failure

Table 34.3 In-plane shear test methods

Information about methods Rnil shear test Torsion of square plate Non-symmetric bending
(a) (b) (c)

Loading scheme h
WlIOO
-I;
YLx
Determinable characteristics r u
Gxy'xy Gxy r u
Gxy'xy

Specimen sizes see Ref. 16, p. 115-116 L=50mm; 100mm L = 76.5 mm; w = 11 mm; h = 19 mm
Standards or references Refs. 2, 12, 18 Refs. 2, 17 ASTM D 5379; Refs. 7, 10, 16
Measurable values .p, PU, £4,50 P,Wp P, PU,r
Relationships for Gxy = P/2F£4,50; Gxy = 3PU
-_.
calculations r; = PUfF, Gxy = rXY/r~
where F = Lt (a); fw p '
F = 2Lt (b); rU=PU/F
xy
see Ref. 2, p. 163 see Ref. 2, p. 172
Limitations:
Structural:
Lay-up 0°; 90°; 0° /90° no
Angle of cutting 0°; 90° no ~
l!l.
Physical: no Linear range of the curve For Gxy: linear range ~
P vs. w; wp ~ O.5t of the curve r vs. r ~
L L no
Geometrical: w>lO For BFRP: 25< T < 100 I·
Note: Pu is the load at specimen failure; wp is the deflection at load P (c).
~
Table 34.4 Methods of torsion of bars, ring specimens and tubes
~
Information about methods Rods with circular cutting-out Split rings Tubes
(a) (b)
;::t
...
f[
'"'iJ'
1i
~
Loading scheme z .
/ .p Ii

Determinable characteristics TU G(IrIGez G(IrIT;;

'"
Specimen sizes dmean = 5-10 mm; t = 2 mm;
see Ref. 11 See Ref. 2, p. 149
see Ref. 2, p. 208
Standards or references Ref. 2 Refs. 2, 11 Refs. 2, 16
Measurable values MUT p,Wp M T, Mj!, IPI' IP2
2M~ MT L
Relationships for calculations T" -
!'2 -
c= 3 -J til -IP2
7td2meant Wp 1 Gez = p 7'1

TtPR3 - EJ M~
T" -
see Ref. 2, p. 209 see Ref. 2, p. 216 ez - Wp

Limitations:
Structural:
Lay-up Symmetric about z-axis 0°; 90°, 0° 190°

Angle of cutting no 0°; 90°, 0° 190° 0°, 90°

For Gez: linear range
Physical: no no
of the curve MT vs. IP
Geometrical: 1 w tlR = f(EJE e)
t == 2 mm (for GFRP) Term EJ --+ 0 (atT--+ 3) tlR < 1/10 (for GFRP);
t/R < 1/40 (for GFRP)i

Note: M~ is the torque at specimen failure; lp is the polar moment of inertia; J = tW /12; wp is the displacement at the load application points (b).
Table 34.5 Methods of bars and ring specimens bending

Information about methods Three-point bending

Wfwle-rings bending
Straight bar Curved bar

p"
p
.,.,
)(
Loading scheme (a) J;k
-4- z
J.

Determinable characteristics Ex Gx~' rx~u, u i OO Ger , ii: u

b, U: er
Specimen sizes w = 15 mm; 30 mm; t = 1-10 mm; R = 73 mm; w = 6.35 mm; R = 73 mm; w = 6.35 mm;
Lit = 15 -17 L = 16.1 mm; t = 3.18 mm t = 10-14.6 mm
Standards or references ASTM D 790 Refs. 2, 7, 14, 16 ASTM D 2344 Ref. 2. ASTM D 2290 Ref. 2.
Measurable values P, pu, wpmax P P, PU, W o
3 QU
Relationships for For E:; Gx~ : see Ref 2, p. 241 ibu = - - For Ger:
er 2 wt
calculations without shear; see Ref I, p. 237
= PV 14wP w pmax;
E: where QU is ~ vs. {~}2
ix';' = 3 PUl4wt defined in Ref. 2, p. 279 Eo R
3ptu
ibu =- -
er 4 wt
(see Ref. 2, p. 268)
Limitations:
Structural:
Lay-up 0°; 90°; 0° 190 0
Angle of cutting 0°; 90° ~
<n
Physical: For E:, G~: linear range of the curve P - wp no For GOr : linear range of the curve P vs. Wo .....
Geometrical: For E:: Lit> 40 ;For E:, G;; ~ < u:" For u:
i:r tlR > 0.08-0.18 (for GFRP); ]
2R i:r u For Eer: ~l
8:;;
L .I( 1 - k E:)
t :;; 1.095 y ~ ex';, _t_ i:r u ""~
-b-
> B. <- 0.72~( 1 -k k. eEe)
2L Uo u t T 8r
Note: Pu is the load at specimen failure; kr = wrJJwo is the fraction of deflection induced by transverse shear; Wo is the change in ring diameter. ~
U1
786 Mechanical tests

7isJ ~imel1s

~'~$-+~~hf?
~ ~
~
~
~\ fJO- lithe

1- /
Fig. 34.1 Methods of material fabrication and respective specimens for testing.

to the fiber direction. Tubular specimens with methods, loading types, and failure modes are
different balanced fiber lay-ups (fiber angles all different for the two test methods (Table
are symmetric with respect to the longitudinal 34.1).
specimen axis) are employed to assess shear A uniaxial tension or compression test
characteristics and to study complex states of specimen has several functionally different
stress. parts: two loading sections, a gage section,
The specimen shape, to a great extent, and two transition sections. The loading sec-
depends on the objective of the test: verification tions provide a means of fastening the
of scientific hypotheses, engineering specifica- specimen in the testing machine. They receive
tion of the material, or quality control of the and transmit the external loads to the gage
materials. The most rigorous requirements are section of the specimen. In the gage section,
imposed on specimens of the second group. deformations are measured and stresses are
calculated according to the geometrical
dimensions and external load. The transition
34.4 TENSION AND COMPRESSION
sections serve to attenuate stress-strain per-
turbations in the loading section to isolate
34.4.1 TENSION TESTING OF FLAT SPECIMENS
them from the gage section. The specimen
In spite of any analogy that may be drawn dimensions that are recommended in the
between loading of flat composite specimens existing standards meet these requirements.
in tension and compression, only the calcula- The specimen dimensions (length, width and
tion relationships (taking into account the sign thickness) specified in standards as a function
of the stresses and strains) are common to both of the type of fiber lay-up are shown in Table
cases. The specimen shape and size, gripping 34.1 (Methods (a) and (b».
 Tension and compression 787

forces acting on the specimen ends (Table 34.1,

The greatest technical challenge in tension
testing of composites, especially unidirec- Method (c)), it is impossible to achieve a suffi-
ciently uniform stress distribution over the
tional composites, is the reliable transmission
faces of the specimen. As a result, premature
of tensile forces from the grips to the speci-
men. This is generally performed through the local failure of the specimen occurs. When the
loading is achieved by shear forces acting on
use of friction forces. Tabs bonded to the spec-
imen improve the efficiency of load the sides of the specimen as specified by
transmission considerably. The tabs should beASTM D3410 (Method (d», the stress distrib-
made of a material that has a much lower ution in the specimen is also non-uniform,
modulus of elasticity and a higher total elon-
especially when flat wedge grips are used.
The best method is a combination of the two
gation than the respective characteristics of the
specimen material. Tabs have been made of methods in which normal forces are applied
fiberglass reinforced composites, aluminum at the ends and shear forces are applied along
and wood veneers. The thickness of tabs the faces of the specimen grip section
should be between 1.5t and 4t, where t is the(Method (e».
specimen thickness. The tabs must have a In compression testing of unidirectional
composites in the fiber direction, three basic
large enough area that the ultimate shear load
capacity of the bond between the tabs and themodes of failure are observed: buckling of the
specimen is greater than the breaking load ofreinforcing fibers, transverse cracking of the
the specimen gage section. matrix, and shearing of reinforcing fibers at a
The mode of failure in tension depends on 45° angle without local buckling of the rein-
the relationship between the external load forcement. Transverse cracking is caused by
and the reinforcing fibers and on the type ofdifferences in the Poisson's ratios of the mate-
reinforcement lay-up. When unidirectional rial components and by a non-uniform
composites are loaded in the reinforcement transverse strain distribution along the speci-
direction, they fail by breakage of the rein-men length. Materials reinforced at an angle to
forcing fibers. This is accompanied by the specimen's longitudinal axis fail in shear
transverse cracks, longitudinal shear cracks without crumpling at the end faces because all
and delamination of the polymer matrix. of the shear load is taken up by the matrix. The
Increasing the angle between the load and theaforementioned basic modes of failure can be
reinforcing fibers causes the mode of failureaccompanied by a series of other phenomena:
to change gradually from shear and splitting inelastic and non-linear deformation of the
of the polymer matrix parallel to the fiber reinforcing fibers and matrix, delamination,
direction to pure transverse tensile crackingsurface peeling, overall buckling and crushing
of the polymer matrix. The failure mode of of the end faces. Failures with different combi-
composites with a balanced angle-ply rein- nations of these phenomena can make the
forcement depends on the angle of the fiber determination of the failure mode very diffi-
lay-up. cult.
In compression testing, great care must be
taken to ensure stability of the specimen, espe-
34.4.2 COMPRESSION TESTING OF FLAT
cially in the gage section. Buckling of the
SPECIMENS
specimen side face is not always detectable
The main problem in compression testing of and will cause erroneous strain measure-
flat specimens is the selection of a loading ments. Special test fixtures are used to prevent
method that ensures compressive failure. overall buckling of the specimen.
When the loading is achieved by normal
788 Mechanical tests

34.4.3 TENSION TESTING OF RINGS The stress concentration problem is elimi-

nated in tests that use uniform internal
The most popular means of tension testing
pressure generated by the use of a compliant
rings uses a half-disk loading device (Table
ring (Method (d» or by a hydraulic system
34.2, Methods (a) and (b». This is because the
(Method (e». The disadvantages of the com-
test is easy to perform and the data is easy to
pliant ring test method are the need for
analyze. However, it has several significant
periodic calibration of the loading element
disadvantages: the strain distribution over the
and the need for very careful preparation of
specimen circumference is non-uniform, fric-
the specimen surface. The disadvantage of the
tion between the specimen and the half-disks
hydraulic test technique is the need for expen-
has a strong effect on the results, and there is a
sive hydraulic equipment.
stress concentration in the specimen at the gap
between the two half-disks because of a varia-
tion in the radius of curvature of the specimen.
34.4.4 COMPRESSION TESTING OF RINGS
In thin-walled rings, the stress concentration
takes the form of a slight increase in radial ten- Radial compression of rings is accomplished by
sile stresses that causes a minor effect on the external pressure (Table 34.2). The analog of the
test results. However, in thicker specimens, half-disk tension test is the simplest compres-
the resultant interlaminar shear stresses can be sion test method (Method (c». The primary
high enough to cause failure at lower loads difference is that in the compression test, it is
than failure due to the circumferential stresses possible to reduce the stress concentration in
alone. The error increases with increases in the the specimen at the split line. The best results
relative specimen thickness t / R, the degree of are obtained with a semi-circular housing that
anisotropy, and the ultimate strain of the mate- has a locking arrangement that prevents radial
rial. Since this test method yields erroneously growth at the split-line (Fig. 34.2). There are
low strength values, it can only be used for also compression analogs for the compliant
qualitative comparison of composites. ring (Method (f) and the hydraulic system
Corrections have been suggested but have not (Method (g». In the compliant ring method, the
been used in practice. compliant ring is the elastic foundation of the

Fig. 34.2 Typical interlocking features for ring compression fixture.

 Shear 789

specimen and to a certain extent it prevents gage section where the state of stress is the
buckling of the specimen. The external pressure most uniform. However, edge effects have
may also be applied by mechanical devices considerable influence on the shear strength.
such as multiple cams. Therefore, it is better to bond the specimen to
The primary difficulty in compression test- the rail links than to use mechanical fasteners.
ing of rings by external pressure is the The stress distribution is not affected by the
selection of a relative thickness, tlR, to reduce loading direction, i.e. along the diagonal or
secondary loading effects. Depending on the parallel to the sides of specimen gage section.
relative thickness of specimen, t I R, and the Measuring the shear properties by tensile
degree of anisotropy of the material E/GfJrI loading of an anisotropic strip is distinguished
three different failure modes have been by its apparent simplicity (Table 34.1, Method
observed. Thin-walled rings fail by buckling, (f). The strip can have one of several different
thick-walled rings fail by biaxial compression, fiber lay-ups. This method is not used to deter-
and optimum thickness rings fail by circum- mine in-plane shear strength because it yields
ferential compression. Analysis of the test data low values. A state of pure shear is not assured
must consider the radial as well as circumfer- even with a +45° lay-up.
ential stresses. A similar test method involves tensile load-
Delamination of the inner layers of the ring ing of a strip of a unidirectional material cut at
makes it difficult to correctly calculate the an angle, 0, to the reinforcing fibers. The opti-
compressive strength. This delamination is mum angle is the one for which the relative
often noisy. The occurrence of this failure shear strain r1/£ x is maximized and the shear
mechanism also depends on the relative thick- stress f12 reaches its critical value. This angle
ness of the ring. Delamination of helical depends on the anisotropy of the elastic and
windings can lead to unwinding of the speci- strength properties of the material tested. For
men. Hoop wound rings can fail by advanced composites, the optimum angle is 10
layer-by-Iayer delamination. to 15°. Because the stress ratios are very sensi-
tive to changes in the angle, rigid tolerances,
±l degree, are set on the specimen cut angle,
34.5 SHEAR
the strain gage angle, and the direction of
loading. In order to ensure that the stress state
34.5.1 IN-PLANE SHEAR
is uniform, relatively narrow strips, Llw = 14
Shear properties, especially shear strengths, to 16, are used.
are difficult to measure. The simple and eco- The in-plane shear modulus is often mea-
nomical rail shear test is often used for this sured by twisting a square plate with four
purpose (Table 34.3, Methods (a) and (b». The point loading (Method (c». The wide accep-
extent of edge effects and the uniformity of the tance of this method may be attributed to the
shear stress distribution over the specimen simplicity of its calculations. However, the
width depends on the length-to-width ratio of experiments should be performed with utmost
the specimen gage section Llw and on the care. This method is only applicable for small
relation of elastic constants G IE of the mate- deflections, wp < O.tt, on plates made of mate-
rial. Edge effects are negligible lor Llw > 10. rials which are uniform in thickness and
Edge effects cannot be eliminated for materials orthotropic along the specimen axes. Test
with vxy =vyx == -1. Thus, this method cannot be results for several different materials have
used for such materials. The elastic constants shown that the P-w relationship remains lin-
obtained by the rail shear test are less sensitive ear up to wit == 1. PHowever, in practice, the
to the relative ratio, Llw, since the measure- deflection, w ,should be limited to O.5t to pre-
ments are taken in the center of the specimen vent instability. Only the initial linear section
790 Mechanical tests

of the P-w curve should be used to determine 34.6 TORSION

the shear ri,.odulus. The optimum range of rel-
Torsional loading of thin-walled tubes is a
ative plate thickness, Lit, is determined by
standard test for measuring in-plane shear
two conditions: the contribution of transverse
modulus and strength (Table 34.4, Method (c».
shear to the deflection at small values of LI t
In this test, the stresses are distributed uni-
and the possible loss of stability at large values
formly around the circumference and alo~g
of Lit. The limits of Lit are given for BFRP in
the length of the specimen. The shear strams
Table 34.3. However, tests run on GFRP (glass
are practically constant through the thickness
fiber reinforced plastic), CFRP (carbon fiber
of the specimen wall. In torsion, the definition
reinforced plastic) and BFRP (boron fiber rein-
of 'a thin-walled tube' is a function of the
forced plastic) with different fiber lay-ups have
degree of material anisotropy EJ Eo, which can
shown that reliable data can be obtained at LI t
vary over a wide range. The disadvantages of
> 15. The specimen must be flat and of constant this method are that it requires relatively large
thickness because the calculated shear modu-
specimens, special test fixtures, inserts to pre-
lus is related to f3. The distance from the point
vent buckling of some specimens and wound
of support or load application to the comers of
specimens or specimens of special configur~­
the plate should not exceed 2t. Experimental
tions, e.g. materials in which the fiber lay-up IS
evaluation of these three shear test methods
parallel to the specimen axis.
has shown that they all yield comparable val-
The results obtained by torsional shear tests
ues of in-plane shear modulus9•
compare favorably with results obtained by
The successful application of the double V-
test methods using flat specimens.
notch or losipescu shear test (Method (d» to
Torsional loading of split rings is also used
all types of fiber lay-ups is well known10• It is
to measure shear moduli (Method (b». If the
invaluable for testing spatially reinforced com-
sample size limitations indicated in T~~le 34.4
posites since these materials do not possess
are followed, bending effects are neghgtble.
planes of low shear strength. All other shear
strength test methods use this characteris~c of
laminated composites to induce shear fallure 34.6.1 INTERLAMINAR SHEAR
first and therefore are useless for spatially
Good estimates of interlaminar shear proper-
reinforced materials. In-depth investigations
ties, especially for spatially reinforced
have shown that the losipescu method and its
materials, have been obtained by torsion test-
modifications yield good results in shear tests
ing of rods with a circumferential no~ch
of carbon--carbon composites reinforced along
(Method (a». The specimens can be tested WIth
three mutually perpendicular directions (3-D)
or without a central bore. The important geo-
and four principal diagonals of cube of 4-Dll.
metric parameters of the notch are the relative
Spatially reinforced composites are less sensi-
width Lid, diameter, d, and wall thickness, t.
tive to the dimensions of the notches and gage
It has 'b~ shown that within a range of Lid
length than laminated composites. The distrib-
ution of shear stresses is essentially uniform
=0.2-1.0, the length of the notch does not affect
the measured shear strength, f rz The gage sec-
U.
throughout the gage section of 3-D and 4-D
tion diameter can be increased from 5 mm to 15
materials cut at a 90° angle with a total notch
mm (0.6 in) without affecting f rzu • However,
depth equal to a half of the specimen thick-
increasing the diameter beyond 15 mm (0.6 in)
ness. Moreover, it is possible to prevent stress
causes a sharp drop in the measured strength.
concentrations at the notch tips. Specimens
with extra side notches yield the best shear
strength data.
 Bending 791

34.7 BENDING for measuring interlaminar shear strength.

However, refined analysis has shown that the
34.7.1 THREE-POINT BENDING state of stress in a short bar of anisotropic
The most popular type of bend test is the material is significantly different from the
state of stress predicted by isotropic theory2,12.
three-point bend test (Table 34.5, Methods (a)
and (b». The four-point and five-point bend The shear stresses through the thickness of a
tests are less popular in spite of their consider- relatively short anisotropic bar have a para-
bolic distribution only in the middle of the
able technical advantages over the three-point
bend test. span. At the loading points, the distribution of
Theoretically, the moduli of elasticity of shear stresses through the thickness of the
specimen has peaks near the surface directly
~omogeneous ~terials in tension, compres-
Slon, and bending are the same, i.e. E t = E c = beneath the loading points. In relatively short
bHowever, d anisotropic bars, there are no planar regions of
Ex' ue to'lffiperfections, thex statex of
stress in bending and differences in the fiber constant maximum shear stress. Moreover, on
r~latively short bars (L/w ~ 5), the compres-
lay-up through the thickness of the material,
the bending modulus of elasticity Exb can differ Slve transverse stresses from the load
somewhat from Ext or E c. This difference is application points can extend over the entire
emphasized by the super;cript 'h'. length of the specimen and can exceed mean
shear stress by up to a factor of 15. These com-
The formulae used to determine the elastic
pressive stresses constrain crack opening at
~ending constants, E/ and G/', from the deflec-
~ample dela~ation and result in an apparent
tion of a bar at its midspan must take into
consideration the effect of interlaminar shear. mcrease of mterlaminar shear strength. As a
result of these deviations from the ideal para-
The effect of interlaminar shear can be
bolic stress distribution, the experimentally
neglected for large values of relative span
determined interlaminar strength appears to
len~ L/ t. For highly anisotropic materials, the
decrease with increasing relative span width.
relative span length must be greater than 40.
Therefore, shear testing of relatively short bars
When determining bending strengths, fail-
ures limited by the normal strength, a bu, and can provide only a qualitative comparison of
failures limited by the shear strength different composites.
x
, r xzbu' Interlaminar shear strength can also be
must ~e separated. Unlike those of isotropic
matenals, the two strengths for composite measured by three-point bending of curved
materials can differ by an order of magnitude. segments. The shear strength is calculated
wi~h th~ same equations used for bending of
The shear stresses can have a considerable
effect on the failure. pnsmatic bars. However, the additional inter-
laminar normal stresses unique to curved
Failure due to normal stresses occurs by
fracture of the extreme outer layers in com- beams must be taken into consideration. The
pression or tension. Failure due to shear normal stresses act over the entire length of
the specimen. The sign of the stress depends
stresses occurs by delamination approximately
at the midplane of the specimen. Laminated on the orientation of the specimen. In the case
of segments loaded with their convexity
materials can fail by a violent debonding of the
upwards (center load applied to the outer
compressed outer layer. Very short bars experi-
diameter of the specimen) the stresses are ten-
ence a. third failure mode. They fail by
sile, a/. When the convexity is downwards, the
crumplmg and shearing which is accompanied
radial stresses are compressive, a,<. In the for-
by an apparent increase in shear strength.
mer case, shear and tensile radial stresses
Three-point bend testing of short bars or
combine to decrease the apparent shear
ring segments is the most widely used method
strength. In the latter case, the compressive
792 Mechanical tests

radial stresses impede the growth of the shear and II), and the Arcan Test (mixed Modes I and
delamination crack and raise the apparent II). The geometry of the specimens, prepara-
interlaminar shear strength. For accurate tion methods, and analysis procedures have
determination of the interlaminar shear been described in detaip3,14.
strength, "C(Jrbu, segment dimensions must be Advanced composites, especially those
selected so that the normal circumferential reinforced with carbon and aramid fibers,
stresses, a9, and normal radial stresses, ar, are have highly anisotropic thermophysical prop-
negligible compared to the shear stresses. erties. This is reflected in the thermomechanical
behavior of the structures fabricated with these
materials. The combination of a polymer
34.7.2 BENDING OF RINGS
matrix having a high coefficient of thermal
Bending of complete rings by diametrically expansion and fibers having a negative coeffi-
opposed loads (Method (c» is used to deter- cient of thermal expansion allows the
mine elastic and strength properties of fabrication of composites with extremely low
composites. Reliable results are obtained if the thermal expansion. This property of compos-
relative specimen thickness, tlr, is properly ites is used in the fabrication of structures
selected. The acceptable range of relative which are stable over wide temperature
thickness for determining the shear modulus, ranges. Low thermal expansion is commonly
G9rb, is based on the material anisotropy, achieved in two directions and the process can
E9b IG(Jrb. The shear modulus is calculated from be extended to materials which are spatially
the load-diametral deflection data using the reinforced in three directions (3-D) or along
same equations for three point bending of four diagonals of a cube (4-0)15. These com-
prismatic bars with a correction factor for the posites possess a thermal expansion coefficient
fraction of deflection induced by shear which is both isotropic and very low.
stresses. When the test is used to determine Measurement of the thermal expansion
the interlaminar shear strength, "C(Jrbu, the rela- coefficients of carbon and aramid composites,
tive specimen thickness must be chosen to especially in the reinforcement direction, is not
ensure failure by shear delamination at the a trivial experiment. It must employ modem
specimen mid-radius rather than failure by dilatometer and interferometer methods14.
normal circumferential stresses, alJbu.
34.9 STRUCTURAL TESTING
34.8 SPECIAL TESTS
The unique challenges of designing and test-
Laminates can fail by interlaminar delamina- ing composite parts are due to the fact that the
tion along specific planes. Toughness data is as material and its micro- and macro-structures
important for characterization and failure pre- are created at the same time as the part. The
diction of composites as strength and stiffness design of critical structures must include the
data. Cracks in composites can propagate by design of the material and must consider the
Mode I, (crack faces opening normal to the unique behavior of composite materials which
crack plane), by Mode II (crack faces sliding in is influenced by processing techniques, actual
their planes), or by Mixed-Mode (combination service loading, and environmental condi-
of Mode I and Mode II). The interlaminar frac- tions. Structural testing should start with tests
ture tests include the Double Cantilever Beam of small-scale models fabricated by the same
Test (Mode I), the Edge Delamination Test manufacturing process as the full-scale struc-
(Mode I), the End Notched Flexure (Mode II), ture, followed by tests of prototype parts,
the Notched Three-Rail Shear Test (Mode II), specimens cut from structural elements, and
the Cracked Lap Shear Test (mixed Modes I finally, full-scale tests9,19.
 References 793

REFERENCES 11. Greszczuk L.B., Shear Modulus Determination

of Isotropic and Composite Materials. ASTM
1. Lubin, G. (ed.). Handbook of Composites, New Special Technical Publication, 1969,460: 140-9.
York: Van Nostrand Reinhold, 1982. 12. Uemura M., Problems in Mechanical Testing
2. Tamopol'skii, Yu.M. and Kincis, T.Ya, Static Test Methods of Advanced Composite Materials. In
Methods for Composites. New York: Van Proc. 10th Tsukuba General Symp., 1990, pp.
Nostrand Reinhold, 1985. 43-54.
3. Witney, J.M., Daniel, I.M. and Pipes, RB., 13. Pagano, N.J. (ed.), Interlaminar Response of
Experimental Mechanics of Fiber Reinforced Composite Materials, (Composites Materials
Composite Materials, Rev. Ed.; Society for Series; 5) Amsterdam: Elsevier, 1989.
Experimental Mechanics. Englewood Cliffs, NJ: 14. Carlsson, L.A and Pipes, RB., Experimental
Prentice-Hall,1984. Characterization of Advanced Composite Materials.
4. Kelly. A (ed.). Concise Encyclopedia of Composite Englewood Cliffs, NJ: Prentice-Hall, 1987.
Materials, Oxford: Pergamon Press, 1989. 15. Tarnopol'skii, Yu.M., Zhigun, I.G., and
5. Tamopol'skii, Yu.M. and Vasiliev. Y.V. (eds.). Polyakov, V.A., Spatially Reinforced Composites.
Structural Composites. A Handbook, Moscow: Lancaster: Technomic, 1992.
Mashinostroyenie, 1990. 16. Mechanical Testing of Advanced Fibre Composites,
6. Tsai, S.w., Theory of Composites Design, Dayton, Imperial College of Science and Technology,
Paris and Tokyo: Think Composites, 1992. London: University of London, 1992.
7. Wyoming Test Fixtures. High Performance Test 17. Sims G.D., Nimmo, W., Johnson, AF. and
Fixtures. Product Catalog, Laramie: Wyoming Ferriss, D.H., Analysis of Plate-Twist Test for In-
Test Fixtures Inc, 1993. Plane Shear Modulus of Composite Materials.
8. Peters, S.T., Humphrey, W.D. and Foral, RF., Teddington: National Physical Laboratory, 1992.
Filament Winding Composite Structure Fabrication. 18. Lee, S. and Munro, M., Evaluation of in-plane
Covina: SAMPE, 1991. shear test methods for advanced composite
9. Nikolaev, Y.P., Panfilov, N.A, Popov, V.D., and materials by the decision analysis technique.
Sinitsyn, E.N., Analysis of the Failure Composites, 1986, 17(1), 13-22.
Mechanism of Large-Scale Structures. Meeh. 19. Nikolayev, V.P., Popov, Y.D. and Sborovskii,
Composite Mater., 1993,29(2):203-11. A.K., Strength and Reliability of Wound
10. Pinderra, M.J., Gurdal, Z.C, Hidde, J.S. and Fiberglass Reinforced Plastics. Leningrad:
Herakovich, CT., Mechanical and Thermal Mashinostroyeniye,1983.
Characterization of Unidirectional Aramidl
Epoxy. Report CCMS-86-08, VPI-E-86-29,
Virginia Polytechnic Institute and State
University, Blacksburg, VA, 1987.
DURABILITY AND DAMAGE TOLERANCE 35
OF FIBROUS COMPOSITE SYSTEMS
Ken Reifsnider

35.1 DEFINITIONS AND ISSUES simple, and are illustrated in Fig. 35.l.
Durability and damage tolerance are critical Damage tolerance is the remaining strength
after some period of service, and durability, in
to the design of composite structures.
Damage tolerance is the approach often general, has to do with how long the compo-
required for the certification of safety-rated nent will last, i.e. with the life of the structure.
structures such as aircraft components; dura- In this context, durability is often discussed in
terms of the resistance or susceptibility to
bility has been identified as one of the most
damage initiation. Both concepts imply that
important technical drivers for the design of
major composite structures such as the High the subject component is being exposed to
Speed Civil Transport. Recent reports from applied conditions such as mechanical loading
the National Materials Advisory Board and a and environments such as temperature and
chemical agents over long periods of time that
great volume of other literature focus on
these issues l - 3• are typical of the projected service life of the
Of course, there are many nuances in the component.
There are several technical concepts that
definitions of durability and damage toler-
ance. However, the basic concepts are quite fonn a foundation for our discussion of these
closely related topics. The first of these is the

Damage Tolerance (Remaining strength)

)
1

Normalized
stress level

Life Locus / ,,
,
Durability (Life) --.:,,
Time I Cycles

Handbook of Composites. Edited by S.T. Peters. Published Fig. 35.1 Basic definitions of 'durability' and 'dam-
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 age tolerance'.
 Definitions and issues 795

question of the relationship of material 'how strong is this composite structure', even
strength to structural strength. In general, the if the array of material strengths are known
strength of (fiber reinforced) composite mate- (shown in Fig. 35.2). Hence, there is a need to
rials is represented by an array of values that develop understandings and representations
reflect the anisotropic nature of the materials of the critical damage and failure modes that
(Fig. 35.2). For planar materials, at least the control the performance of engineering com-
tensile strength and compressive strength in ponents. This technology is currently
the fiber direction and transverse to the fibers incomplete, but discussions of those topics
and the in-plane shear strength are required will follow.
for a complete answer to the question of 'how A second fundamental concept is
strong is this material'. However, as an array, microstructural architecture. As shown in Fig.
those values do not directly show 'how strong 35.3(a), many fibrous composite components
is a composite structure'. Several possible are made in layered or laminated form, with the
answers to that question are typically given. fibers in different layers having different direc-
One may use a 'failure criterion' that compares tions; in some cases the plies are made from
all of the point stress components with all of different materials to form a 'hybrid' compos-
the material strength components (such as the ite. In addition, the fibers may not be straight,
Tsai-Hill or Tsai-Wu criteria)4 in some collec- but may be woven, braided, or arranged in
tive form based on concepts such as critical mats of various types (Fig. 35.3(b),(c». These
energy, critical shear resistance, etc. The salient details have a major influence on the durability
point to be made is that the complexity of and damage tolerance of the materials. In fact,
(inhomogeneous) composite materials and most composite material systems are 'designed'
their array of anisotropic material strengths to be 'fiber dominated', to take advantage of
give rise to the development of a correspond- light, strong and stiff (but brittle) fiber materials
ing array of damage and failure modes in that are available. Typically, the fibers, their
these materials that must be understood and geometry and their arrangement are important
correctly modeled to answer the question of parts of the question.

Five In-plane strength values for flberous composites:

Tension and compression Tension and compression

strength In the fiber
strength In the direction
direction transverse to the fibers
Xt orXc
..J Vt orVc
In-plane shear strength
tl..---S

Strength tensor:
Composite ~ X II =(X t ,X c ,V t ,V c ,S)

Fig. 35.2 Schematic illustration of 'principal strength' directions in a unidirectional continuous fiber com-
posite laminate.
796 Durability and damage tolerance offibrous composite systems

--
Mid;l~ne:)-
./' x
/
/
/
,/

(0 )

(e)
(b)

Fig. 35.3 Typical engineering composite reinforcement types: (a) fibrous, unidirectional pile; (b) fibrous
woven; (c) fibrous, braided.

A third technical issue has to do with the • Remaining strength and life models are
degradation of intrinsic strength and stiffness. developed and predictions are made for
For metals, the material stiffness and strength each independent failure mode (such as
are generally constant during the life of the fiber failure in tension or micro-buckling in
engineering component. This may not be true compression, etc.).
for composites. Stiffness changes of the order • Mechanics representations of the state of
of 10-20% may be caused by micro-cracking, stress and state6f material are constructed
for example. Since many structures are stiff- on the basis of a 'representative volume' of
ness designs, this mode of degradation must the material that is typical of the distributed
be considered. In addition, the intrinsic mate- damage state that controls the remaining
rial strengths (indicated in Fig. 35.2) may also stiffness and strength of the composite. A
be degraded, especially by such things as typical representative volume of material is
physical or chemical aging. This behavior a controlling ply in a laminate, but may be a
must also be part of the supporting predictive micro-buckling ligament, a small group of
technology developed for these materials. fibers, etc.
Nondestructive methods of tracking such • Various methods are used to characterize
degradation are under development, but this and monitor the rate of strength degrada-
remains as a challenge currently. tion in composites. A typical parameter
Methodologies for the assessment and pre- which is useful for that purpose is stiffness
diction of durability and damage tolerance of change; however, that parameter is not
composite materials typically involve the fol- appropriate in some cases.
lowing features: • Micromechanics (mechanics analysis at the
 Damage modes and failure modes 797

fiber/matrix level of representation) is made. Composites consist of mechanical'mix-

increasingly used for remaining-strength tures' of distinct phases (such as fibers or
modeling, for the calculation of stiffness particles) in a matrix material. The geometry
change (which leads to internal stress redis- and arrangement of the reinforcement phase
tribution), and for the estimation of is carefully chosen to achieve the desired com-
remaining strength for a given failure posite properties. As a result, such material
mode. systems are always inhomogeneous, often
• Statistical considerations are essential for anisotropic, and often brittle. These three
the correct representation of the long-term basic characteristics control the nature of
behavior of composites. Composites typi- damage development and failure in compos-
cally fail because of the statistical ite materials.
accumulation of defects, which eventually The most salient single feature of damage in
interact to create a critical condition. This is composites is the process of damage accumula-
in contrast to self-similar single crack prop- tion. Damage development usually involves
agation that is the typical mechanism of many damage modes which create a widely
failure for common metals. distributed damage state, and failure is usually
• Time-dependent behavior such as vis- the result of a statistical accumulation of dam-
coelastic creep, creep rupture (driven by age (rather than the statistical occurrence of
such things as internal stress redistribution damage). As discussed below, these multiple
or oxidation), and aging are typically damage accumulations on failure modes are
important in the consideration of the long- often closely related to the manner in which
term durability and damage tolerance of the composite is made, especially to the basic
polymer composites, particularly for com- nature of the inhomogeneity and anisotropy of
ponents that serve at elevated temperatures. the material. This damage development
process ultimately controls durability and
This chapter will discuss the range of physical
damage tolerance, so we will discuss some typ-
and engineering details that define and control
ical major features of that phenomenon.
this subject. Of course, a complete discussion
The most pervasive damage mode in com-
would fill several volumes, so the reader
posite materials is microcracking, most often
should regard this discussion as only a start-
in the matrix material. Figure 35.4 shows two
ing point for further study.
embodiments of this mode. Figure 35.4(a)
shows an X-ray radiograph of a cross-ply lam-
inate with cracks in both ply types, and Fig.
35.2 DAMAGE MODES AND FAILURE
35.4(b) shows matrix cracking parallel to the
MODES
fibers in the off-axis plies of a laminate, as seen
The failure of 'typical' (homogeneous from a tracing of those cracks as they appear
isotropic) engineering materials is a familiar on the edge of this [0,45,-45,90]. laminate. A
topic. The subjects of ductile rupture and brit- typical scenario for the development of such
tle fracture are widely discussed and taught in cracks is the formation of matrix cracks as a
undergraduate and graduate courses. function of increasing applied load or increas-
However, composite materials generally do ing cycles of loading. These cracks typically
not behave in a manner easily described by extend through the thickness of a ply and gen-
either plasticity (or yield) theory or by self- erally extend quickly in the fiber direction if
similar crack growth concepts. the local stress is uniform.
The reason for this different behavior is the Several other important features of matrix
fundamental difference in the micro-structure cracking are suggested by Fig. 35.4. As shown
of composite materials, i.e. in the way they are by Fig. 35.5, matrix crack formation releases
798 Durability and damage tolerance of fibrous composite systems

Stress
" " Crack initiation point

Strain
I
I I Fig. 35.5 Change in slope of the elastic stress-strain
f curve induced by microcracking.
I I I
I (
APPLIED STRESS (MPa 1
b 100 200 300 400 500 GOO i'OO

Fig. 35.4 Microcracking in the matrix, parallel to e <> FATIGUE

OOA
the fibers; a radiograph of a cross-ply laminate with !
o QUASISTATlC
(a) inter-ply delamination at crack intersections
(!)
DATA
(arrow) and (b) a tracing of matrix cracks on the
edge of a [0,+45,-45,90] laminate. ~
l<:
4.0
U
«
II: 2.0
U
stored energy in the cracked ply or material,
0
and changes the stiffness of material propor- 0 0.2 0.4 0.6 0.8 1.0
tionately, a matter of concern to engineering NO. OF CYCLES (MILLIONS)

applications, as noted earlier. However, the

density of cracks in the ply of a laminate Fig. 35.6 Data showing identity of the equilibrium
reaches a stable saturation value, as first crack spacing (,characteristic damage state' or CDS)
observed by Reifsnider et aJ.5,8, called the char- for quasi-static and cyclic loading of a laminate.
acteristic damage state (or CDS) of the ply.
That CDS is a function only of the properties the local loads if they were not bonded together
of the plies, their thickness, and their stacking in the composite. Hence, stored energy is
sequence. Figure 35.6 shows that the same released if they separate, and that energy drives
CDS is formed by static or cyclic loading. This the separation process. The most common
CDS can be readily predicted since the crack example of this damage mode is the separation
spacing is determined by the rate at which the of the plies of a laminate near a free edge, as
surrounding material can transfer stress back shown in Fig. 35.7. This process has been
into the broken ply. Moreover, the stiffness widely studied and is well described. More will
change caused by this cracking can also be be said of this driving mechanism below.
predicted as we1l7,9-11. It should be noted that delamination is usu-
A second important damage mode is delam- ally nucleated by other damage modes (such
ination, as shown in Fig. 35.7. Delamination is as matrix cracking)12-14 and although it is a
driven by the fact that local regions of the com- common damage mode, it is not usually a fail-
posite would deform differently in response to ure mode, per se. Delamination usually begins
 Damage modes and failure modes 799

matrix cracks in one ply may cause fiber frac-

ture in an adjacent ply due to the local stress
concentrations21 ,22.
Figure 35.8(b) shows a second feature of
importance. When the matrix and fibers have
comparable stiffness and strength, the fibers
may break many times along their length
before the composite fractures. In this situa-
tion, fiber fracture can cause a significant
stiffness loss as well as a strength reduction23 •
Another generic damage mode is micro-
buckling, induced by local or global
compressive loads, as shown in Fig. 35.9.
Fig. 35.7 Edge micrograph of delamination (arrow)
showing (a) relationship to matrix cracking; (b) plan
view radiograph of edge delamination in a cross-
ply laminate (shaded regions).

at an edge, such as a cutout, bolt hole, rivet

hole, etc. If it is in a region of nonuniform
stress, it may stop growing when it reaches the
boundary of that region. Even if it grows to
large dimensions, it usually does not cause
significant loss of strength in engineering
sized structures. Still, the loss of integrity can
lead to other damage and failure modes, so it
should be avoided.
A third generic damage mode is fiber frac-
ture. Many composites are 'fiber dominated',
i.e. they depend on the fibers for their stiffness
and strength. Hence, fracture of the fibers is
both an important damage mode and failure
mode. However, fiber fracture is difficult to
detect and has been studied less completely
than many other damage modes. However,
considerable data have been collected 1s- 17•
Figure 35.8 shows two examples of such data,
driven by two important mechanisms. Figure
35.8(a) shows fibers broken beside one
another, a typical situation. In many compos-
ites, the fibers are coated with a material that
decreases the tendency for the fracture of one
fiber to cause the fracture of neighboring fibers
by forming an 'interphase region' around the Fig. 35.8 (a) Adjacent fiber fracture in a polymer
fibers that tends to 'isolate' the fracture matrix lamina and (b) fracture in ceramic matrix
effectsl 8-20. It is also important to note that the laminate.
800 Durability and damage tolerance offibrous composite systems

There are several aspects of this phenomenon

that are of importance to durability and dam-
age tolerance. For example, the compression Steady-state region
strength (or remaining strength) of the com- Strain
posite may be controlled by the local stress
required to initiate the local instability, in
which case one wants a large diameter, stiff
fibers in a stiff matrix. Or, the strength may be
controlled by local resistance to shear defor-
mation after buckling begins, in which case
one would choose a tough matrix or inter-
phase region between the fibers and the Time
matrix. This is another case in which a damage
mode mayor may not be a failure mode, an Fig. 35.10 Schematic of typical creep deformation
at constant load and temperature.
important distinction.

or rheological models that represent the behav-

ior in terms of a change in the stiffness of the
material with time, as a function of tempera-
ture. Quite often, the reinforcing fibers do not
show creep behavior at low temperatures, but
at high temperatures, essentially all con-
stituents may creep. The changes of stiffness
with time can be characterized in the labora-
tory, and must be modeled carefully, based on
those data. In fact, this part of the behavior is
critically important to the correct calculation of
internal stress states, since the creep of the con-
stituents changes the internal stress distribution
greatly in some cases. For example, if the matrix
Fig. 35.9 Localized microbuckling in a polymer creeps more than the reinforcing fibers (a typi-
matrix composite. Printed with permission, I.M. cal situation), that creep 'relaxes' the stress in
Daniel and O. Ishai, Engineering Mechanics of the matrix, and increases the load carried by the
Composite Materials, Oxford University Press,
Oxford,1994.
fibers. If we wish to calculate a fiber-controlled
strength, for example, a correct representation
of this behavior must be included in our model.
Another 'damage mode' of considerable impor- Finally, another failure mode is creep-rup-
tance to polymer composites and all composites ture. This is a fairly general terminology used
used at high temperatures is the phenomenon to refer to a variety of physical phenomena
of creep, i.e. time-dependent deformation at that produce time-dependent failure. This can
constant applied stress. Figure 35.10 shows a be due to, say, oxidation of the fibers, or to
typical form of that behavior, with an initial other physical degradation processes which
transient region, a steady state region (in which eventually cause rupture. It is clear that these
most engineering design is done), and a tertiary phenomena must also be modeled correctly if
(usually unstable) region. This phenomenon is we are to discuss durability and damage toler-
usually represented by introducing viscoelastic ance of material systems.
 Damage drivers and damage 'resistance' 801

35.3 DAMAGE DRIVERS AND DAMAGE by local discohtinuities in stress state, typically
'RESISTANCE' caused by neighboring plies or ply groups
(bonded together) that would have very dif-
In the previous section, a number of damage ferent strain states if they were not bonded.
and failure modes that occur in composite Hence, the orientation of the plies in a lami-
materials, and ultimately control durability nate and the stacking sequence of those plies
and damage tolerance, were identified. Many are controlling players in the development of
of these modes are related to the manner in the interlaminar stresses that drive delamina-
which the composites were put together. This tion. This problem has been exhaustively
raises the basic question of 'can one design studied, and methods of reducing interlami-
composite materials to be durable and damage nar stresses have been widely discussed12- 14,
tolerant?' Most of the rest of this discussion but because of the inhomogeneous and often
will address this question. anisotropic nature of composites, interlaminar
Some general concepts will be followed by stresses generally cannot be eliminated in lam-
some micro-mechanics methods of quantifying inated systems, so mechanical methods are
answers. Microcracking is likely to be the most widely used to control that tendency. The most
pervasive damage mode in typical composites, successful of these is weaving, i.e. to use
especially under long-term loading, and most woven fiber architectures to reduce the
especially under cyclic loading. Even though anisotropy of a given ply, and therefore, to
most matrix materials are chosen because they reduce the 'disagreement' between the
offer some level of ductility, in most composite response of any two or more plies. Woven
systems the matrix is highly constrained so materials are now widely used, especially for
that cracks develop due to local constraint, this reason. A second approach is to 'stitch' the
local stress concentrations, and local defects composite in the region of non-uniform stress,
that grow rapidly under what is generally a typically near an edge of the laminate.
'plane strain' condition. Hence, matrix tough- Stitching simply 'clamps' the edge of the mate-
ness, in the general sense, is the key to the rial to prevent it from separating; the internal
reduction of matrix cracking. Increasing the stresses are still present. Stitching has a some-
strain to failure of the matrix material is a pri- what smaller number of proponents, but is a
mary objective, and increasing the plane-strain successful method as well. Finally, three-
fracture toughness of the matrix is a compan- dimensional reinforcement, such as mats or
ion objective. There is a richly developed braids, also serve the purpose of providing
science and technology associated with matrix constraint to the delamination drivers. These
toughening; some starting points are listed in methods are not as widely used at this time,
Wilkinson et aI.24 and Hedrick et aI.25. largely because of the difficulty associated
A second damage mode identified earlier is with manufacturing.
delamination. This problem is a strong combi- A less obvious influence on durability and
nation of structural and material concerns. The damage tolerance is the bonding between the
material concerns are essentially the same as fiber and matrix. The nature of this influence
those discussed for matrix cracking, with one has only come to light in recent years. Some of
important exception. Matrix toughness does the mechanics models needed for this discus-
not translate directly into interlaminar tough- sion will be developed in the next section; only
ness. Hence, resistance to delamination cannot a few general points will be made here. First,
be controlled entirely with material property the properties of composite materials are
increases. The structural part of the problem determined not only by the properties of the
does, however, present opportunities. It was constituents, but they are also greatly influ-
mentioned before that delamination is driven enced by the manner in which the constituents
802 Durability and damage tolerance offibrous composite systems

interact. This critical interaction is, of course, strain to failure by as much as a factor of
controlled by the bonding between the con- four2 7,28! In general, although design rules are
stituents, between the fibers and matrix in our not yet fixed, design of the interphase region is
case. Typically, this bonding is 'controlled' by a new and important opportunity for the
a fiber coating or 'sizing'. However, it is now enhancement of the durability and damage
known that such things as notched fatigue tolerance of composite systems29--32.
behavior can be improved by as much as two The final subject in this section is 'failure
orders of magnitude by carefully 'designed' criteria'; which are used to describe remaining
'interphase' regions between the fibers and the strength. In general, failure criteria are chosen
matrix26 • There are at least two basic concepts on the basis of the known failure mode. If fiber
operating in these effects. First, if one can fracture controls strength, then a suitable crite-
toughen the composite by toughening the rion may be just the stress in the fiber direction
interphase between the fibers and matrix, the divided by the strength in that direction. If
composite is likely to be more durable, as dis- matrix behavior is controlling, a shear stress or
cussed above. Second, the interphase region combined stress criterion may be appropriate.
can greatly influence the local stress state, and Figure 35.12 shows a comparison between
reduce the driving force for fiber-matrix strength 'envelopes' predicted by two popular
debonding. An illustration of that is shown in criteria. It is important to note that the inputs
Fig. 35.11. If one considers the strength of a to the failure function will, in general, change
composite under loads applied perpendicular as a function of time and loading history. The
to the fiber direction, then it is clear that the general form of any failure criterion will usu-
fiber causes a local stress concentration, in pro- ally be some function of the ratios of stress in
portion to the difference between its principal material directions to strength in
properties and those of the matrix. However, if those directions, as mentioned earlier. Under
a coating around the fiber is introduced, this long-term conditions which induce damage,
local concentration can be greatly reduced. In the local stress changes as damage causes
fact, for a 'rigid' fiber, compared to the matrix, redistribution, and the principal values of
it is not surprising that a compliant coating on material strength change, due to such things
the fiber will increase the transverse composite as constituent degradation or micro-damage.
strength by as much as a factor of two, and the Hence, to calculate damage tolerance by using

Fiber Interphase region

.. ~
Allowable
... ~ Applied
.. .. Stress (ksi)

.. ~

10
0 90
Angle of Loading (deg)
Composite
Fig. 35.12 Allowable uniaxial loading as a function
Fig. 35.11 Schematic diagram of the geometry of of angle of loading relative to the fiber direction in
the interphase region in a fibrous composite, sub- a unidirectional lamina, estimated from a maxi-
jected to loading transverse to the fibers. mum stress and a popular effective stress criterion.
 Composite micro-strength and remaining strength models 803

failure functions (or criteria) to calculate utility of such models. The example is a recent
remaining strength, one must be careful to use model of tensile strength. (Figure 35.13). The
the correct local stress state and material state stress in the broken fibers builds back up to the
in those expressions, especially when degra- undisturbed level by shear transfer from the
dation has changed those states from their surrounding matrix, composite, and inter-
initial values. phase region. That rate of buildup is directly
proportional to the stress concentration in the
next nearest fibers; if the buildup occurs over a
35.4 COMPOSITE MICRO·STRENGTH AND
short distance (a short 'ineffective length'), the
REMAINING STRENGTH MODELS
stress concentration in the neighboring fibers
The importance of material principal strengths is great, and they tend to break causing very
was noted, and the importance of composite brittle composite behavior. However, if the
microstructure in the determination of those buildup occurs over a large distance (i.e. if the
strengths has been emphasized. The proper- material around the fiber is very compliant or
ties, geometry, arrangement, and bonding of breaks easily), the strength of each fiber is lost
the constituents determine the resulting val- completely when the first local fiber break
ues of composite principal strengths. So, if occurs. A model has been developed that
those factors are understood, strong, durable, describes the physics and mechanics of this
damage tolerant composites can be designed. behavior, which estimates the fiber strength as:
That understanding is currently incomplete,
_ mlm+l( 2toL )1/m+l(_2_)lIm+l m+1
but some models are available. Such models °c - 0 0 D m+2 m +2
are very valuable since they can tell us the
preferable way to make composite materials, (1 + m)1/m
in contrast to how they can be made (the job of x (cnm + cnm-l + ... + l)l/m (35.1)
the materials science community).
In this limited space, one example will suf· where 0 0 is the Weibull characteristic strength
fice to demonstrate the general nature and of the fibers, to is the shear stress between the

Tensile
, loading Normal stress In:
Composite
• broken fiber

A
" neighboring
Fibers . . fiber

V Jf
~~
I('~

V''\ I\.. V
Fiber
breaks 1'-,1/

, t---J
Average global values
away from fiber fracture

Fig. 35.13 Schematic diagram of the local stress distribution around broken fibers in a unidirectional
composite.
804 Durability and damage tolerance offibrous composite systems

fibers and the matrix (usually taken as the continuous fiber reinforced composites is
interphase strength), m is the Weibull shape outlined. A great many details will have to be
parameter for the fiber strength distribution, omitted due to space limitations; the inter-
and en is the local stress concentration when n ested reader can find them in other
fibers are broken together in a local region. publications37-40.
Hence, the tensile strength in the fiber direc- A start is to identify a well-defined failure
tion can be estimated on the basis of the mode, as defined earlier. Since damage is dis-
properties of the constituents and the inter- tributed, this damage mode will be 'typical' of
phase region between the fibers and the any 'representative volume' of material; a
matrix. If any of those constituent characteris- mechanics boundary value problem on such a
tics change, the model can show how the representative volume (RV), as suggested in
strength of the composite changes, i.e. the Fig. 35.14 can be 'set'. This RV may be discon-
model can be used to calculate the damage tol- tinuous; i.e. it may have cracks, delamination,
erance of the composite if the failure mode is debonds, etc. But some part of it will remain
controlled by fiber strength in tension. intact until fracture of the composite, and this
Comparable models can be constructed for part of the RV that defines the fracture event is
compression failure, and for other failure a critical element'. Therefore the objective is
I

modes33-36 . the calculation of the state of stress and state of

material in the 'critical element.' One can write
all failure functions, Fa' in that critical element,
35.5 ESTIMATION OF REMAINING
and claim that when these failure functions
STRENGTH AND LIFE
(for each distinct failure mode) predict failure,
As indicated earlier, damage tolerance is the composite will fail.
defined by remaining strength, and durabil- Invoking kinetic theory we can derive an
ity is usually discussed in terms of life. In this equation that relates changes in stress state and
final section, one approach to the estimation material state with time and loading history to
of the- durability and damage tolerance of remaining strength, i.e. allow the incorporation

failure modes
i i,.l
II.ll
, ,'-r\.
representative ~
volume ~ •

~r/~eal element
suberitleal element(s)

Fig. 35.14 Diagram illustrating how experimental observation of failure modes define the representative
volume (used to set the boundary value problem) and the critical elements in which all continuum states
are defined.
 Estimation of remaining strength and life 805

of the explicit time, cycles, and environmental sile fiber failure, and it is assumed that some
dependence that leads to phenomenological fatigue behavior of unidirectional material
behavior such as creep, creep rupture, fatigue, under uniaxial stress in the fiber direction has
and aging into the calculation of remaining been measured, a 1-D SN relationship can be
strength. From thermodynamic principles, the derived, of the form:
follOwing expression can be derived:
S5 = A + B (log N)P (35.3)
Fr = 1 - tn/N}l (1 - Fa)A~ tld(~) (35.2) u

where, for our example, A = 1, B = -0.1, Su =

100 ksi (the initial ultimate strength), p =1, and
where F - F( aij(n) )
Sa is the applied stress amplitude. Equation
a - a Xin)
(35.3) provides an input, N, into (35.2) since
in the critical element, Fr is the normalized

I lj
remaining strength, n is cycles, and N is the
Xl -A
lfP
life of the critical element under the current N = lOA an (35.4)
state of stress and state of material. The B
methodology of this calculation is shown in
Fig. 35.15. Remaining composite strength, Fr' where an is now the current local ply stress in
is calculated directly; life is calculated by the the fiber direction, and Xl is the current local
coincidence of Fa and Fr' Numerous compar- principal material strength in tension, given
isons of such calculations with experimental by eqn (35.1). Substituting eqn (35.4) in
data have been made over the last 10 years or eqn (35.2) and assuming that no other
so, and there are a few examples at the end of phenomena are present (and using j = 1.2, a
this chapter. The immediate purposes are known typical value), the curve (a) in Fig.
served by using eqn (35.2) to examine the 35.16 results. Now, suppose the ply is the
effects of some hypothetical change in mater- critical element in a multiaxial laminate
ial state and stress state on remaining strength having off-axis plies that crack and 'dump'
(damage tolerance). If the failure mode is ten- stress into the critical element as a function of

Fig. 35.15 Schematic flow diagram of the manner in which the MRLife simulation scheme calculates
remaining strength and life.
806 Durability and damage tolerance offibrous composite systems

Remaining I
Strength Fiber-
0.5 Direction
Strength

80~--~--~~~~~~
oL---~--~--~~.-~
10 10 100 1000 Ix104 Ixi c?
Cycles

Fig. 35.16 Calculated remaining strength predic- Fig. 35.18 Assumed degradation of fiber strength
tions for (a) 0° lamina degradation alone; (b) added for the sample laminate.
effect of matrix cracking; (c) added effect of fiber
degradation (e.g. by oxidation).

cycles, according to the rate shown in Fig. Fig. 35.18. Then the strength model, eqn (35.1),
35.17 (from cracking rates that must be correctly integrates that micro-change into the
measured or estimated). With this internal global calculation, and eqn (35.2) shows the
stress redistribution, only, added to eqn (35.2), damage tolerance to be curve (c) in Fig. 35.16
the damage tolerance changes to curve (b) in for that situation. Hence this 'micro-kinetic'
Fig. 35.16. Of course, if creep occurs in the approach has the capability to estimate
matrix (perhaps because of increased durability and damage tolerance for very
temperature), in which case the local fiber complex situations involving combinations of
stress will increase again as a function of many time and cycle dependent phenomena
cycles to change the form shown in Fig. 35.17. in composite systems, using a mechanistic
Finally, suppose creep rupture is occurring, approach.
driven, for example, by oxidation of the fibers An example follows. Using the methods
that is reducing the diameter of the fibers, D, described above, the rate of matrix cracking
in eqn (35.1), in the manner shown in and the unidirectional SN curve of a carbon
fiber reinforced PEEK matrix composite were
determined, and used to estimate the remain-
80~----~~----~ ing strength and life of several different
laminates made from that material. Figure
Fiber- Stress increase due
35.19 showl an example of the predicted and
to matrix cracking
Direction observed life for several load levels of a quasi-
60 isotropic laminate made from such material
Stress
and Fig. 35.20 shows comparisons of the pre-
dicted and observed remaining strength of
40~------~----~4
such laminates for two load levels and cycles
1 5xl04 9.999xlO of load application. It can be seen that this
Cycles approach can produce quite useable results.
Many such predictions have been compared
Fig. 35.17 Assumed increase in stress in the 0° ply using the MRLife performance simulation
due to matrix cracking. code based on this method37-40.
 Estimation of remaining strength and life 807

AS-4/PEEK (APC-2)
Quasi-Isotropic Notched Fatigue (R=-1)
0.75

0.7

-
-s 0.65
t)
0.6
t)
........................................................................ ..... ...........................................
0.55
•
0.5

0.45 L....._ _ _ _.1-._ _ _ _....I.-_ _ _ _....L..-_ _ _ _..J

3 4 5 6 7
Log N (Cycles)
Simonds & Slinchcomb MRUfe

•
(1989) Prediction

Fig. 35.19 Predicted (line) and observed life for a quasi-isotropic A5-4/PEEK notched coupon under fully
reversed loading.

Residual Strength at 0.70 Suit Residual Strength at 0.90 SUit

'1.0& r-------=.-----~ 1.05 r-------~-----.......:~

1.00 J----------................ ,.00 1-----_=.................................

I! o.D5

1
II:

1--
0.85 ...................................................... . 0.85 ...................................................... ,.

__ __..J-__-L_ _ __
0.80
~~L..... ~ ~. ~.~-..J

L.....--L_-..JL-~_..L-_.....I-_.L....--I
1 10 100 1,000 10,000 100,000 1,000,000 1 2 5 '0 20 so 100 200
Cycles Cycles
~~t ~"~1111111

Fig. 35.20 Predicted (lines) and observed residual strength of A5-4/PEEK specimens subjected to under
fully reversed cyclic loading.
808 Durability and damage tolerance offibrous composite systems

35.6 SUMMARY 6. Damage in Composite Materials: Basic Mechanisms,

Accumulation, Tolerance, and Characterization,
This has been a short outline of the physical STP 775, American Society for Testing and
behavior associated with the durability and Materials, (ed. K.L. Reifsnider), 1982.
damage tolerance of composite material sys- 7. Highsmith, A.L., Stiffness Reduction Resulting
from Transverse Cracking in Fiber-Reinforced
tems, and a few modeling approaches to the Composite Laminates. Master of Science Thesis.
estimation and prediction of that behavior have Blacksburg, Virginia: Virginia Polytechnic
been indicated. It should be noted that there is Institute and State University, 1981.
every evidence that composite materials are 8. Reifsnider, K.L., Some fundamental aspects of
remarkably durable and damage tolerant. the fatigue and fracture of composite materials.
Fatigue allowables for carbon fiber reinforced In Proc. 14th Ann. Soc. Engng Science Mg.
Bethlehem, PA: Lehigh University, 1977.
polymer composites, for example, exceed those
9. HighSmith, A.L., Damage Induced Stress
of structural steels, and the durability and dam- Redistribution in Composite Laminates. PhD
age tolerance of ceramic composites make them Dissertation. Blacksburg, Virginia: Virginia
the only choice for ultra-high temperature Polytechnic Institute and State University, 1984.
applications in turbines, etc. Understanding of 10. Bader, M.G., Bailey, J.E., Curtis, P.T. and Parviz,
this subject, which is admittedly incomplete, A., The Mechanisms of Initiation and develop-
has reached a level that is sufficient to support ment of damage in multi-axial fibre-reinforced
engineering applications of composites to even plastic laminates. In Proc. 3rd Intern. Con!
Mechanical Behavior of Materials. Cambridge,
the most demanding situations in the most U.K.,1979.
severe environments. In fact, that is exactly the 11. Reifsnider, K.L., Damage and damage mechan-
situation in which the application of compos- ics. In Fatigue of Composite Materials, (ed. K.L.
ites is most beneficial and cost effective. Reifsnider), Amsterdam: Elsevier Science
Composite material systems can provide many Publishers, 1990.
new opportunities to design for damage toler- 12. O'Brien, T.K., Characterization of delamination
ance and durability. onset and growth in a composite laminate. In
Damage in Composite Materials, ASTM STP 775,
1982, p. 140.
REFERENCES 13. O'Brien, T.K., Analysis of local delaminations
and their influence on composite laminate
1. Life Prediction Methodologies for Composite behavior. In Delamination and Debonding of
Materials, NMAB-460. Washington, D.C: Materials, ASTM STP 876, 1985, pp. 282-297.
National Academy Press, 1990. 14. O'Brien, T.K. and Hooper, S.J., Local delamina-
2. High-Temperature Materials for Advanced tion in laminates with angle ply matrix cracks:
Technological Applications, NMAB-450. Part I, Tension tests and stress analysis. In
Washington, D.C: National Academy Press, NASA TM 104055, 1991.
1988. 15. Razvan, A. and Reifsnider, K.L., Fiber fracture
3. Horton, P.E. and Whitehead, R., Damage and strength degradation in unidirectional
Tolerance of Composites, Vol. I and II, Air Force graphite/epoxy composite materials. In
Wright Aeronautical Laboratories, AFWAL TR- Theoretical and Applied Fracture Mechanics, 1991,
87-3030, 1988. 16,81-89.
4. Daniel, I.M. and Ishai, 0., Engineering Mechanics 16. Razvan, A., Bakis, CE. and Reifsnider, K.L.,
of Composite Materials. New York: Oxford Univ. SEM Investigation of fiber fracture in composite
Press, 1994. laminates. In Materials Characterization, 1990, 24,
5. Reifsnider, K.L. and Highsmith, A.L., 179-190.
Characteristic damage states: A new approach 17. Lorenzo, L. and Hahn, H.T., Fatigue failure
to representing fatigue damage in composite mechanisms in unidirectional composites. In
laminates. In Materials: Experimentation and Composite Materials: Fatigue and Fracture, ASTM
Design in Fatigue, Guildford, U.K.: STP 907, American Society for Testing and
Butterworth/IPC, 1981, pp. 246-260. Materials, Philadelphia, PA, 1986, pp. 210-232.
 References 809

18. Ishida, H. (ed.), Controlled Interphases in 30. Jayaraman, K., Gao, Z. and Reifsnider, K.L., The
Composite Materials, New York: Elsevier, 1990. interphase in unidirectional fiber reinforced
19. Warren, R (ed.), Ceramic Matrix Composites, epoxies: effect on local stress fields. J. Compo
New York: Chapman and Hall, 1992. Tech. Res., 1994, 16(1): 21-31.
20. Reifsnider, KL. (ed.), Fatigue of Composite 31. Jayaraman, K., Reifsnider, K.L. and Swain, R.E.,
Materials, London: Elsevier Science Publishers, Elastic and thermal effects in the interphase:
1991. Part II. comments on modeling studies. J. Compo
21. Jamison, RD., Fiber fracture in composite lami- Tech. Res., 1993, 15(1):14-22.
nates. In Proc. Inti. Con! on Composite Materials 32. Jayaraman, K, Reifsnider, K.L. and Swain, RE.,
VI, 1987, no. 3, pp. 185-199. Elastic and thermal effects in the interphase:
22. Jamison, RD., Schulte, K, Reifsnider, K.L. and Part I. comments on characterization methods.
Stinchcomb, W.W., Characterization and analy- J. Compo Tech. Res., 1993, 15(1): 3-13.
sis of damage mechanisms in tension-tension 33. Gao, Z. and Reifsnider, KL., Micromechanics of
fatigue of Graphite/Epoxy Laminates. In Effects Tensile Strength in Composite Systems. Fourth
of Defects in Composite Materials, ASTM STP 836, Volume, ASTM STP 1156, (eds W. W. Stinchcomb
American Society For Testing and Materials, and N. E. Ashbaugh), Philadelphia, PA:
Philadelphia, PA, 1984, pp. 21-55. American Society for Testing and Materials,
23. TIwari, A, The Development of an Interpretive 1993, pp. 453-470.
Methodology for the Application of Real-Time 34. Xu, Y. and Reifsnider, K.L., Micromechanical
Acousto-Ultrasonic NDE Techniques for modeling of composite compressive strength. J.
Monitoring Damage in Ceramic Composites Under Compo Mater., 1993, 27(6): 572-587.
Dynamic Loads. PhD Dissertation. Blacksburg, 35. Reifsnider, K. L. and Gao, Z., Micromechanical
Virginia: Virginia Polytechnic Institute and concepts for the estimation of property evolu-
State University, 1993. tion and remaining life. In Proc. Intern. Con!
24. Wilkinson, S.P., Liptak, S.c., Lesko, J.J., Dillard, Spacecraft Structures and Mechanical Testing,
D.A, Morton, J., McGrath, J.E. and Ward, T.C., Noordwijk, the Netherlands, 1991, pp. 653-657.
Toughened bismaleimides and their carbon 36. Curtin, W.A, Theory of mechanical properties
fiber composites for fiber-matrix interphase of ceramic-matrix composites. J. Amer. Ceram.
Studies. In Proc. 6th Japan-U.S. Con! Composite Soc., 1991, 74(11), 2837-2845.
Materials, 1992. 37. Reifsnider, K.L., Performance simulation of
25. Hedrick, J., Patel, N.M. and McGrath, J.E., polymer-based composite systems. In Durability
Toughening of epoxy resin networks with func- , of Polymer-Based Composite Systems for Structural
tionalized engineering thermoplastics. In ACS Applications, (eds AH. Cardon and G.
Advances in Chemistry Series, no. 233, Verchery), New York: Elsevier Applied Science,
Toughened Plastics I: Science and Engineering, (eds. 1991, pp. 3-26.
c.K. Riew and A.J. Kinloch), 1993, pp. 293-304. 38. Reifsnider, K.L. and Stinchcomb, W.W., A criti-
26. Swain, RE., Reifsnider, K.L., Jayaraman, K. and cal element model of the residual strength and
El-Zein, M., Interface/interphase concepts in life of fatigue-loaded composites coupons. In
composite material systems. J. Thermoplastic Composite Materials: Fatigue and Fracture, ASTM
Compo Moler., 1990, 3, 13-23. STP 907, (ed. H.T. Hahn), Philadelphia, PA:
27. Case, S.W., Micromechanics of Strength-Related American Society for Testing and Materials,
Phenomena in Composite Materials. MS Thesis. 1986, pp. 298-313.
Blacksburg, Virginia: Virginia Polytechnic 39. Reifsnider, KL., Use of mechanistic life predic-
Institute and State University, 1993. tion methods for the design of damage tolerant
28. Carman, G.P., Eskandari, S. and Case, S.W., composite material systems. In ASTM STP 1157,
Analytical investigation of fiber coating effects (eds M.R Mitchell and O. Buck), Philadelphia,
on shear and compression strength, symposium PA: American Society for Testing and Materials,
on durability and damage tolerance, ASME 1992, pp. 205-223.
WAM, (in press), 1994. 40. Reifsnider, KL., Evolution concepts for
29. Jayaraman, K. and Reifsnider, K.L., The inter- microstructure property interactions in compos-
phase in unidirectional fiber-reinforced epoxies: ite systems. In Proc. IUTAM Con!
effect of residual thermal stresses. Compo Sci. Microstructure-Property Interactions in Composite
Tech., 1993,47,119-129. Materials. Aalborg, Denmark, 1994.
ENVIRONMENTAL EFFECTS ON COMPOSITES 36
Ann F. Whitaker, Miria M. Finckenor, Harry W. Dursch,
R.C. Tennyson and Philip R. Young

36.1 INTRODUCTION stabilizers, vertical fins and fairings were

flight-tested with annual inspections.
Composite usage has increased dramatically
over the last three decades due to the advan- Satisfactory performance of the composite
materials was noted over fourteen years, with
tages of light weight, specific strength and
stiffness, dimensional stability, tailorability of some parts experiencing more than 39 000
properties such as coefficient of thermal hours of flight loads. Also evaluated were
composite parts from military aircraft such as
exp~nsion and high thermal conductivity.
EnvIronmental effects on these properties may the C-130 center wing box, S-76 tail rotors and
compromise a structure and must be consid- horizontal stabilizer, 206L fairing, doors, and
vertical fin and the CH-53 cargo ramp skin.
ered during the design process.
Because of the variety of uses, the compos- Boron/ epoxy, graphite/epoxy, Kevlar / epoxy
and Nomex honeycomb were used in these
ite environment cannot be exactly defined.
aircraft and helicopter components.
This chapter details the major environmental
concerns for the composite designer, problems
encountered with these environments in the 36.3 ENVIRONMENTS AND EFFECTS
past, and some materials or protective systems
effectively used. The use of trade names, how- 36.3.1 BIOLOCICAL ATTACK
ever, does not constitute endorsement, either
expressed or implied, by the authors. Biological attack on composites may consist of
fungal growth or marine fouling. As reported
in the literature, fungal growth does not
36.2 HISTORICAL PERSPECTIVES appear to be as damaging as the wet condi-
Use of composites in commercial aircraft tions that promote growth. Fungicide has been
increased under two NASA programs, the mixed in with resins to retard this growth.
Flight Service Evaluation Program and the Marine boring organisms do not appear to
attack glass-reinforced composites. Even
Aircraft Energy Efficiency Program, begun in
the 1970s. These programs included evalua- though marine organisms will grow on com-
posite surfaces, mechanical properties do not
tion of environmental effects on the composite
parts of the Boeing 727, McDonnell Douglas appear to be affected, and the fouling can be
removed by scraping (Fried, 1969).
DC-10 and Lockheed L-1011 commercial air-
craft. Elevators, rudders, ailerons, horizontal Composites with graphite fibers have been
used in medical applications for both internal
and external purposes. Internal composite
Handbook of Composites. Edited by S.T. Peters. Published structures, such as artificial joints or plates for
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 bone fracture support, must be biocompatible
 Environments and effects 811

or the material may degrade over time. further if heat is present or if the composite is
External composite designs, such as artificial undercured or has a large amount of voids.
limbs or orthotic braces, may experience Moisture is absorbed into the composite
impact damage, flexural and torsional loading until a saturation point is reached. This has
during use. been described as a non-Fickian process,
meaning the rate of relaxation in the material
due to water absorption is comparable to the
36.3.2 FATIGUE diffusion rate of water: As the material proper-
Fatigue, either through mechanical loads or ties change, such as a decrease in glass
acoustic vibrations, can cause crack growth or transition temperature, the diffusion process
local defect formation. Fatigue design depends changes. Swelling stresses due to non-uniform
not only on the load, but also on the use tem- water absorption have been investigated
perature range and amount of moisture (Ashbee, 1989). Volume expansion due to
present. Very cold temperatures (below -50°C water absorption can be a few percent at satu-
(-58°F» may increase the stiffness of some ration. Moisture absorption is usually
composite materials, thereby increasing the dependent on the matrix, but aramid fibers
susceptibility to fatigue damage (Staunton, will also absorb water. The mechanical proper-
1982). ties degrade in relation to the amount of
Destructive effects of fatigue vary with the moisture absorbed, with no further deteriora-
composite system tested. One example of tion after saturation is reached. Strength
fatigue resistance is the B-1 horizontal stabi- reductions in polyester laminates have been
lizer torque box, an all-mechanically fastened found to be 10-15%, while epoxy resins are
hybrid composite structure (Staunton, 1982). less vulnerable.
Acoustic fatigue testing produced no degrada- In a few cases, drying of the composite
tion, nor did the service environment of restored the original mechanical properties.
moisture, mechanical fatigue, and tempera- Testing of a glass/polyester laminate allowed
ture cycling from -12 to +167°C (10 to 260°F). to dry after ocean exposure at 1700 m (5700 ft)
below sea level for three years showed little
change in compressive strength and modulus,
36.3.3 FLUIDS flexural strength and modulus, or interlaminar
shear strength (Fried, 1969).
Moisture
Fiberglass composites with either polyester
Moisture effects on composites have been stud- or epoxy resins have been used extensively in
ied for decades. Water acts as a plasticizer marine structural applications because of their
when absorbed by the matrix, softening the strength-to-weight characteristics and resis-
material and reducing some properties of the tance to the marine environment. Glass
laminate. Moisture may also migrate along the reinforcement is preferred over carbon fibers
fiber-matrix interface, affecting the adhesion. due to carbon's electrical conductivity, which
Moisture in composites reduces matrix-domi- may result in severe dissimilar metals galvanic
nated properties, such as transverse strength, corrosion with sea water acting as an elec-
fracture toughness and impact resistance. trolyte.
Lowering of the glass transition temperature MIL-HDBK-l7B, besides providing guide-
may also occur in epoxy and polyimide resins lines for characterizing materials and
with an increase in absorbed moisture. designing a composite system, contains a
Debonding can occur due to formation of dis- wealth of mechanical property and environ-
continuous bubbles and cracking in the matrix. mental effects data. The effect of moisture
Mechanical properties can be reduced even absorption or water immersion on weight,
812 Environmental effects on composites

operational temperatures, and mechanical (150°C (302°F» environment, microcracking

properties such as stress rupture characteris- occurred. The amount of moisture absorbed, as
tics is discussed. measured by weight gain, is directly related to
the change in mechanical properties. Salt water,
antifreeze and gasoline had the most pro-
Aircraft fluids
nounced effect. Drying did restore some, but
The aircraft fluid environment consists of fuel, not all, of the strength and modulus.
hydraulic fluid, lubricants, deicing com- Volkswagen of Germany tested composite
pounds, and water. Polysulfone has been systems for compatibility with gasoline, oil
found to be sensitive to phosphate ester based and coolant for engine use (Beckmann and
hydraulic fluids. Some polymer resins, such Oetting, 1985). These materials were tested
as PEEK, may have lower glass transition both as pure resin and in fiber reinforced lam-
temperatures after exposure to fluids with a inate form. Among the materials tested were
high aromatic content. A study of stressed and glass, carbon and aramid fibers, high-temper-
unstressed composite materials (Dexter, 1987) ature epoxy, polyimide and polyester resin
evaluated short-beam shear strength and ten- systems. Mechanical properties of the samples
sile strength after immersion in JP-4 fuel, were measured after each 100 h of immersion
hydraulic fluid, a fuel-water mixture and up to 1000 h total. Adverse reactions of the car-
fuel! air cycling for 5 years. The composites bon reinforced materials with metal parts, oil
tested were T300/5208 graphite/epoxy, and combustion residues were noted.
T300/5209 graphite/epoxy and Kevlar
49/5209 in (±45)s configuration. The
Other fluids
fuel-water immersion appeared to be the
most damaging, reducing the tensile strength Liquids accidentally spilled on composite sur-
of the T300 / 5209 and the Kevlar / 5209 by 11 % faces may also affect the mechanical
and 25%, respectively. Fuel-water exposure properties. Methylene chloride, found in paint
also degraded short-beam shear strength by strippers, may cause severe damage to epoxy
as much as 40%. resins and a number of other polymers.
Graphite/polyimide composites samples were
immersed in various fluids for 10 min then
Automotive fluids
tested for flexure strength and modulus
The automotive fluid environment consists of (Lisagor, 1979). Slight increases in these prop-
gasoline, oil, battery acid, brake fluid, transmis- erties were noted for samples immersed in
sion fluid and coolant. A study by University of hydraulic fluid, nitrogen tetroxide liquid,
Michigan and General Motors (Springer, monomethyl hydrazine liquid and unsymmet-
Sanders and Tung, 1981) details the effects of rical dimethyl hydrazine. Samples exposed to
prolonged exposure of E-glass/polyester and hot hydrazine vapors were degraded beyond
E-glass/vinylester to automotive fluids. the ability for mechanical testing. Solvents,
Composite samples were immersed in water, bases and weak acids at room temperature do
salt water, No. 2 diesel fuel, lubricating oil, not appear to affect graphite/epoxies and
antifreeze or gasoline. Property changes mea- Kevlar/epoxies. Molten metal, such as alu-
sured were weight, ultimate tensile strength, minum or titanium, may react with carbon
tensile modulus, short-beam shear strength and fibers.
shear (flexural) modulus. Specific materials
exposed were OCF-E-920-1 polyester /E-glass,
OCF-E-980 polyester /E-glass, and vinyl
ester /E-glass. In a moist high-temperature
 Environments and effects 813

36.3.4 WEATHERING effects on composite materials. Over 35 differ-

ent types of organic matrix composites were
Warm, moist climates may affect the perfor-
flown on LDEF during its 69-month mission in
mance of composites. Decreases of 10-20% in
LEO. The post-flight testing and analysis of
tensile strength have been noted for fiber-
composites flown on LDEF has become the
glass/polyester and fiberglass/epoxy (Graner,
basis of understanding the long-term effects of
1982) where the surface resin has been eroded
the LEO environment on composites. Data
away due to extended weathering. Resins
from the Solar Array Materials Passive LDEF
which are more weather-resistant have been
developed since this study. In a 10-year study Exp~riment (SAMPLE), the Space
EnVIronment Effects on Spacecraft Materials
of real-time weathering of graphite/epoxy,
Experiment, the University of Toronto
graphite/polysulfone, and Kevlar/epoxy
Institute for Aerospace Studies (UTIAS)
(Dexter, 1987), the materials that absorbed the
Experiment, and the Space Exposure of
most moisture were most affected by UV radi-
Composite Materials for Large Space
ation.
Structures Experiment are presented here and
Erosion due to rain, snow or ice impact
in the section on design considerations.
may be a problem for some aircraft parts,
Further information may be found in the
such as radomes or leading edge parts.
LDEF Post-Retrieval Symposium Proceedings
Coatings, such as polyurethane, may be used
(Levine 1991, 1992).
to make composite parts more resistant to
this erosion. A real-time weathering study
was performed by Grumman on fiberglass Atomic oxygen
parts from E-2A and A-6A aircraft (Staunton,
Atomic oxygen is formed through the dissoci-
1982). Length of service varied from 12 to 19
ation of O2 by UV radiation. It is the
years. Effects of weathering were dependent
on the material used and whether a protec- predominant molecular species at 100-1000
km orbital altitudes. Its destructiveness is
tive coating was intact. When a hygroscopic
caused by its strong chemical reactivity com-
BF3-400 curing agent was used, the fiber-
bined with its translational energy of 5 eV
glass/Epon 828 epoxy lost nearly half of its
(8 x 10-19 J) from the high velocity of the space-
fl~xural strength. Fiberglass/Epon 828 epoxy
craft. Studies on the effects of atomic oxygen
Wl~ methylene dianiline/benzyl dimethyl
amme (MNA/BDMA) curing agent per- have been performed both on samples
formed well, retaining tensile and flexural exposed to the low earth orbit environment
and in ground-based simulators.
strength. Tensile and flexural test samples
taken from a fiberglass rotodome demon- The Mass Spectrometer Incoherent Scatter
(MSIS) neutral atmosphere model is generally
strated the value of good coatings. Where the
paint was intact, the material retained more used for predicting the atomic oxygen fluence
during a mission. Orbital altitude, inclination,
than 90% of its original strength and 82-94%
other orbit parameters and solar activity are
of modulus. Where the paint had been eroded
away, the composite retained only 68% of its used as inputs. The amount of atomic oxygen
original strength. received by a surface also depends on its ori-
entation to the RAM or velocity vector.
RAM facing composites (facing into the
36.3.5 SPACE
velocity vector) flown on LDEF were sub-
jected to an atomic oxygen fluence of
The Long Duration Exposure Facility (LDEF) approximately 9 x 1021 atoms/cm2 which
has provided a wealth of information on the resulted in a thickness loss up to 0.013 cm
low-earth-orbit (LEO) space environment (0.005 in) of material, the equivalent of
814 Environmental effects on composites

Meteoroid/debris impacts
A spacecraft in any orbit is susceptible to
micrometeoroid impact. These small particles,
fragments of asteroids or comets, may impact
at velocities up to 60 km/s (37 mils) but aver-
age 17 km/s (10.5 mils). Those spacecraft in
near-earth orbit are also susceptible to impact
from pieces of space junk or debris, also trav-
elling at high speeds. Damage from impact
may reduce the strength of composite struc-
tures or cause rupture in filament-wound
tanks. Damage may consist of cratering, pene-
tration, including penetration of thermal or
protective coatings, and spallation Fig. 36.2 is
Fig. 36.1 Cross-sectional photomicrograph of of a classic Whipple-type meteoroid/debris
atomic oxygen exposed LDEF graphite-reinforced
shield before and after particle impact.
OMC showing approx. one ply (0.013 em/0.D05 in)
of erosion.

approximately one ply of laminate (Fig. 36.1). • Impact particle

For unidirectional reinforced specimens, the
reduction in mechanical properties was pro-
portional to the reduction in cross-sectional
area. The SAMPLE experiment flown on the ?
LDEF contained tensile specimens of One or more protective bumpers
'-...
graphite/epoxy systems and S-glass/epoxy.
The composite systems flown were HMF Pressure wall
322/P1700 polysulfone in a (±4S). weave,
HMS/934 in both 0° and 90° unidirectional
configurations, and P75S/934, also in both Before impact
OOand 90° unidirectional lay-ups. The S-
glass/epoxy was flown with and without Debris cloud
aluminized thermal control tape as a protec-
tive coating. These samples received direct
atomic oxygen, as well as UV radiation, ther-
/ Bumper breaks
mal cycling and micrometeoroid/ debris up impact into
debris cloud
impacts. Atomic oxygen reaction efficiency Result - less
was calculated to be approximately 1 x 10-24 damage to
pressure wall
cm3 /atom.
The thickness loss due to atomic oxygen
erosion of the S-glass/epoxy samples in this = = Penetration or
spallation of
, ; ==:::J pressure wall may
experiment is estimated to be 9.14 ~ (0.36 Spallation or may not occur
mil). The glass fibers are not susceptible to ero-
sion, and thus protect the underlying matrix. After impact
The atomic oxygen reaction efficiency for
these samples was calculated to be 0.13 x 10-24 Fig. 36.2 Hypervelocity impact of Whipple bumper
cm3 /atom. design.
 Environments and effects 815

The potential hazard of a meteoroid or Facility have been performed on a number of

debris particle is dependent on its size, veloc- composite materials. These materials were
ity, density and angle of impact. NASA Kevlar 49/epox~ non-impregnated Kevlar 49,
SP-8042 may be used to predict the meteoroid Spectra 900/epoxy and IM6/3501-6
environment encountered by a spacecraft in graphite/epoxy. The test configuration con-
near-earth orbit, earth-to-moon space and sisted of a classic Whipple design with the
near-lunar orbit. The space debris environ- composite material used as a bumper protect-
ment in earth orbit is continually changing as ing a 2219-T87 aluminum pressure wall
more debris is being added with every launch. sample.
At the time of publication, NASA TM-100471 The results of these tests (Schonberg, 1990)
is being used as the definition of the space indicated that some composite systems, such
debris environment. as the Kevlar 49 cloth and Spectra 900/epox~
Prediction of damage caused by mete- performed better than equivalent weight alu-
oroid/ space debris impact can be made either minum systems as measured by damage to the
by using a variety of models, study of impact pressure wall. However, the Kevlar 49/ epoxy
sites in flight hardware, or simulation in the and IM6/3501-6 graphite/epoxy panels did
laboratory. Hydrocodes such as CTH and not perform as well as an equivalent weight
HULL use finite elements or finite differences aluminum system at protecting the pressure
to predict penetration or spallation of a com- wall plate from damage, but damage by spal-
posite laminate. Inspection of composite lation was far less for composite than metal
samples from LDEF (Fig. 36.3) revealed 2-5 bumper plates. The composite bumpers did
impacts of <0.1 em diameter. «40 mil) per experience large areas of delamination,
sample. These impacts did not affect tensile though this was restricted to the outer layers
strength as much as the atomic oxygen erosion with peeling in the direction of the surface
(see previous discussion), but the potential laminate fibers. Some charring and melting
effect of many impacts on a composite struc- did occur with the Kevlar 49 cloth (Fig. 36.4)
ture is obvious. and Spectra/epoxy test articles.
Ground simulations in the NASA/Marshall
Space Flight Center Space Debris Impact

Fig. 36.3 Cross-sectional photomicrograph of

coated graphite-reinforced OMC showing impact Fig. 36.4 Simulated meteoroid/ debris impact in
damage and post-impact atomic oxygen erosion. Kevlar 49 cloth panel.
816 Environmental effects on composites

Particulate radiation Thermal extremes

Particulate or ionizing radiation in the space As a spacecraft orbits, it moves from sunlight
environment consists of trapped radiation in to shadow. Solar radiation, both direct and
the Van Allen radiation belts, solar and galac- reflected from the earth (albedo), and thermal
tic cosmic radiation and bremsstrahlung radiation from the earth cause temperature
radiation. Protons, electrons, neutrons, alpha extremes of -185 to +lS0°C (-300 to +300 0 P).
particles, X-rays, gamma rays and other Spacecraft heat sources, such as engines, elec-
charged particles bombard spacecraft in orbit. tronic equipment and batteries, also contribute
Bremsstrahlung radiation is ionizing radiation to the thermal environment. Active thermal
resulting from the deceleration of electrons. control systems or protective coatings with
Various computer models exist for predicting optimized solar absorptance and infrared
the radiation dose during a mission at a par- emittance can minimize thermal extremes.
ticular orbit, but the high variability of solar Solar absorptance is defined as the ratio of
activity and the resulting magnetic storms absorbed light to incident light, and infrared
may give radiation calculations an uncertainty emittance is defined as the ratio of emitted
of an order of magnitude or more. heat to input heat. Low absorptance values
Protons, electrons and cosmic rays may reduce the maximum temperatures when
affect composite materials by surface chemical exposed to direct or albedo radiation and low
changes. The radiation threshold for observing values of emittance reduce minimum temper-
effects is between 108 and 109 rads for atures when exposed to deep space. Stability
graphite/epoxy and graphite/polysulfone. In of these properties is important for long-term
a cooperative effort between Boeing and exposures. Thermal cycling effects on compos-
Langley Research Center (Fogdall, Russell and ites will be discussed later in this chapter.
Cannaday, 1980), these materials were irradi- Spacecraft leaving and returning to earth
ated with high-energy protons and electrons must endure heating during ascent and re-
(approximately 1 MeV) then tested for flexural entry through the atmosphere. Re-entry
durability. The composite systems tested were temperatures can reach 1482°C (2700°F).
T300/934, T300/5208, and C6000/P1700. Reinforced carbon-carbon composites are cur-
Radiation doses of 5 x 109 rads or more rently utilized in the Space Shuttle nose-cap
resulted in surface blistering of the T300/934, cover and wing leading edges. Another appli-
although this had a limited effect on mechani- cation under investigation is a composite nose
cal properties. Flexural testing of irradiated cone for the Space Shuttle External Tank (Sigur
samples showed no significant changes for and Gray, 1990). This design not only weighs
low radiation doses. less than an equivalent metallic design but
A follow-on study performed at Boeing also can maintain the internal environment
(Fogdall et al., 1983) involved high radiation within desired temperature limits without
doses to T300/934, C3000/934, C3000/PMR- thermal insulation. Temperatures on the Space
15 and C3000/P1700 composites. In this test, Shuttle External Tank nose cap may range
compressive breaking strength at a 45° fiber from -lS0°C (-297°F) around the liquid oxy-
orientation and dynamic mechanical analysis gen vent louvers to SOO°C (930 0 P) during
(DMA) were used to measure the effect of 1010 ascent. Mechanical property tests, including
rads of electron radiation. As much as 15% tensile, compression, in-plane shear and inter-
degradation in breaking strength was noted. laminar shear tests, have been performed on
The DMA tests also showed a decrease in laminates, moisture saturated per ASTM-D618
glass-transition temperature for irradiated to reproduce the high humidity at the launch
samples. site. These composite materials were also
 Environments and effects 817

tested for 'ply lift' or moisture-induced delam- Because moisture is removed from the com-
ination at high temperatures. Several posite in a vacuum, mechanical properties,
composite systems were studied, including such as compressive and interlaminar shear
graphite/bismaleimide and graphite/PMR- strengths, may be improved through expo-
15, but Celion G~30/500~3K graphite/BASF sure. However, the designer must consider
506 phenolic was chosen because of its dura- possible dimensional changes due to moisture
bility during the ply lift testing. desorption. Dimensional stability is critical for
space hardware such as optical benches and
truss structures. Another design concern is the
Ultraviolet radiation
effect of outgassed moisture on sensitive
Ultraviolet radiation is that band of light from optics or electronic equipment present on a
300 to about 4000 A. Ultraviolet radiation may spacecraft. Vapor barriers of metallic foils have
cause degradation through molecular weight been used to prevent line-of-sight deposition
change and cross-linking in the resin system. of moisture and other outgassing products.
However, this damage is generally limited to The UTIAS experiment flown on LDEF
darkening of the resin in the surface layer. (Tennyson, 1991) contained a variety of flat
Figure 36.5 is a photomicrograph of a LDEF and tubular composite samples consisting of
composite laminate exposed only to UV. T300/5208, T300/SP-288 and T300/934
Coatings, such as thermal control tape, have graphite/epoxy samples, boron/SP-290 epoxy
been used to protect composite materials from and Kevlar /SP~328 epoxy. A data acquisition
degradation. system recorded outputs from 16 strain ther-
mal gages attached to the composite samples
for 371 days. Outgassing was measured, rang-
ing from 40 days for the T300/934 to 120 days
for the Kevlar/SP-328. Coefficient of thermal
expansion increased slightly for the 0° config-
uration (Table 36.1). This change should be
considered when designing zero CTE lami-
nates for space applications.

36.3.6 TEMPERATURE

Temperature effects on composite materials

discussed in this section include cryogenic
temperatures, elevated temperatures and ther-
mal cycling between these extremes.
Cryogenic temperatures do not appear to
Fig. 36.5 Cross-sectional photomicrograph of uv- affect the mechanical properties of
exposed LDEF graphite-reinforced OMC showing
minimal degradation. graphite/epoxies or graphite/polyimides sig-
nificantly.
Elevated temperatures for a prolonged
Vacuum and outgassing
period of time can seriously affect the proper~
Orbital atmospheric pressure varies according ties of a composite, with even greater effect if
to altitude and solar activity. Average pressure moisture is present. Susceptibility to matrix
is generally 133 x 10-7 N /m2 (1 x 10-7 torr) in softening is not only dependent on the resin
low earth orbit, decreasing to 133 x 10-14 N/m2 but also the lay-up. A study of graphite/poly-
(1 x 10-14 torr) at geosynchronous orbit. imide properties used two different lay-ups, a
818 Environmental effects on composites

Table 36.1 Summary of LDEF /UTlAS composite material thermal data (Tennyson, 1991)

Material type Configuration Strain Thermal Ambient" Space b Thermal

gage gage CTE CTE vacuum
number number (1(J6/0F) (1 (J6fF) facilityc
CTE (1(J6fF)
Stainless Steel
(calibration) Tube 1 1 9.84 10.0 10.0
Graphite/Epoxy Flat 2 (0°) 2 1.32 3.33 2.50
T300/934 (0°)4 3 (90°) 14.7 13.9-15 16.1
Kevlar /Epoxy Tube 4 (90°) 3 33.9 30-35 35.3
SP-328 (90°)4 5 (0°) 0.10 0.71 0.63
Graphite/Epoxy Tube 6 (90°) 4 14.6 13.6-14.3 15.4
T300/SP388 (0°)4 7 (0°) 0.97-1.57 -1.14-3.33 3.75
Graphite/Epoxy Tube 8 (90°) 5 15.6 12.5-15.3 16.1
T300/5208 (90°)4
Boron/Epoxy Tube 9 (±300) 6 1.57 1.67-2.22 0.44-2.0
SP-290 (:t300)28 10 (±60) 11.7 7.5-11.1 12.7
• At atmospheric pressure prior to launch.
b Measured in space environment on LDEF during first 371 days in orbit.
'Measured in laboratory thennal vacuum test facility (22 h at 133 x 10-5 N/m2 (1 x 10-5 torr), -4O"F to 150°F) after 2114
days in orbit and 184 days at ambient conditions.

0° unidirectional lay-up and a (0, ±45, 90)2 lay- composites, and advanced carbon/carbon
up, tested at temperatures ranging from -157 composites.
to +315°C (-250 to +600°F) (Lisagor, 1979). Test Temperature effects are not limited to the
results showed little change in interlaminar matrix material. Extended operation at 350°C
shear strength for the quasi-isotropic lay-up (660°F) and 450°C (840°F) can cause oxidation
while the 0° unidirectional samples dropped of low-modulus PAN-based fibers and high
to approximately 40% of original strength at modulus PAN- or pitch-based fibers, respec-
elevated temperatures. Reduction of normal tively. Oxidation resistance can be improved
moisture content by vacuum drying reduced with higher purity fibers.
the loss to only 70% of room temperature Thermal cycling conditions are common for
strength. a number of applications, including aircraft
High-temperature resins under develop- and spacecraft. Thermal cycling may induce
ment, such as AFR700B developed by the Air microcracking in some composites. A study of
Force (Brown, 1991), have reached glass transi- this microcracking behavior in graphite/PMR-
tion temperatures of 416°C (780°F). AFR700B 15 composite materials was performed at Rohr
retains 50% of its mechanical properties up to Industries (Sullivan and Ghaffarian, 1988).
370°C (700°F). PMR-15 is another high-tem- Woven laminates of C3000/PMR-15 and unidi-
perature resin with excellent properties, but rectional tape lay-up of C60oojPMR-15 were
large or thick structures require debulking a thermally cycled between -18 and +232°C (0
few layers at a time due to the high volatile and 450°F) up to 2000 cycles. The C3000/PMR-
content. Other materials being developed for 15 developed microcracks, with the number of
high-temperature applications include tita- microcracks dependent on the number of ther-
nium matrix composites, ceramic matrix mal cycles. Decreases in compressive and shear
 Protective coatings 819

strengths were noted, although tensile may be applicable to other environments.

strength, a fiber-dominated property, was not On the LDEF, 5-901 glass/epoxy compos-
significantly affected. The unidirectional tape ites were exposed to the space environment.
lay-up, C6000/PMR-15 did not crack during Three of these samples were flown with ther-
testing. mal control tape, consisting of 50.8 IJ.IIl (2 mil)
aluminum with 50.8 IJ.IIl (2 mil) pressure-sensi-
tive SR574 silicone adhesive. The glass/epoxy
36.4 PROTECfIVE COATINGS
samples without tape suffered more mass loss
When an environmentally resistant composite than the protected samples, and solar absorp-
material cannot be utilized, protection of the tance of the unprotected samples increased
material through the use of coatings is neces- 5.4%, probably due to ultraviolet radiation.
sary. A variety of coatings have been developed Post-flight peel tests of the thermal control
for protecting composites from various envi- tape showed an increase in strength with some
ronments. Standard marine paint, pigmented embrittlement of the adhesive. Figure 36.6, a
gel coatings and polyurethanes have been used SEM photograph, also shows the degree of
to prevent ultraviolet damage and weathering micrometeoroid/ debris protection provided
erosion of marine composites. The following by the tape.
coating examples have been used in space but

Fig. 36.6 Micrometeoroid/debris impact in thermal control tape. Substrate is S-glass/epoxy.

820 Environmental effects on composites

A 60 nm (600 A) SiO/100 nm (1000 A) good corrosion resistance and good unifor-

nickel vapor-deposited coating appears to mity. Irregular shapes, such as end fittings,
have prevented resin loss in graphite/epoxy may also be coated with this process. The com-
samples flown on LDEF (Young et al., 1991). posite surfaces were sanded prior to plating to
Inspection of selected composites improve adhesion. However, atomic oxygen
(C6000/P1700, C3000/P1700, T300/934 (with attack resulted in loss of adhesion in the nickel
two different fiber areal weights) and coating without SiOx overcoat.
T300/520S) revealed dramatic visual effects, Numerous other coatings have been stud-
material loss and a deterioration in mechanical ied for space environment resistance. Braided
performance for unprotected composites. and double-braided aluminum coverings,
Chemical characterization suggests that there indium-tin eutectic coatings, various silicone-
is no significant change in molecular structure based paints and zinc oxide pigment in a
of the surviving polymeric matrix resin in potassium silicate and silicone elastomer
these composites. No change in glass transi- binder have been tested (Piellisch, 1991). A
tion temperature was apparent. The majority proposed design for solar array backing on the
of the unprotected flight samples showed a NASA space station employs woven E-glass
decrease in tensile strength and modulus. sandwiched between two layers of silicon
Coating/ composite systems for large space dioxide-coated Kapton.
truss structure applications have been tested
in ground simulations of atomic oxygen, ultra-
violet radiation, and thermal cycling (Dursch REFERENCES
and Hendricks, 1987; Dursch and George,
Anon. 1970. Meteoroid Damage Assessment. NASA
1993). Coatings studied were chromic and SP-8042
phosphoric acid anodized aluminum foil, Ashbee, Ken. 1989. Fundamental Principles of Fiber
sputtered SiO/ sputtered All Al foil, bare AI Reinforced Composites. Lancaster, PA: Technornic
foil, electroplated nickel with and without PublislUng.
SiOx coatings and inorganic sol gel solutions. Beckmann, Hans-Dieter and Oetting, Hermann.
P75S/934 composite was used to form the 1985. Fiber reinforced plastics move inside
engine. Automotive Engineering 93(5):34-41.
truss tubes. Brown, Alan S. 1991. The Air Force finds an ultra-
The aluminum foil used was 1145-H19 alu- high temperature resin. Aerospace America 29
minum alloy with a thickness of 50.S pm (2 (10):56-57.
mil). Thicker foils can be used, depending on Dexter, H. Benson. 1987. Long Term Environmental
weight constraints and flexibility of any coat- Effects and Flight Service Evaluation of
ings applied to the foil. The foil was bonded to Composite Materials. NASA TM-89067.
the graphite/epoxy truss tube with an epoxy Dobyns, Alan. 1991. Structures. Aerospace America
29(12):38-39.
film adhesive and was co-cured. The chromic Dursch, Harry W. and Hendricks, Carl L. 1987.
acid anodized foil provided good protection as Protective coatings for composite tubes in space
well as optical properties. The phosphoric acid applications. SAMPE Quarterly 19(1):14-18.
anodized foil also provided good protection Dursch, Harry W. and George, P. 1993. Composite
while maintaining adhesion. Desired optical Protective Coatings for Space Applications.
properties were obtained by varying thickness Paper read at Third LDEF Post-Retrieval
of silicon dioxide and sputtered aluminum on Symposium, November 8-12, 1993, at
Williamsburg, VA.
top of the aluminum foil. The optimized thick- Fogdall, L.B., Russell, D.A. and Cannaday, 5.5. 1980.
nesses were 1 pm (0.04 mil) Si02 and 300 nm Radiation Exposure of Composite and Polymer
(3000 A) AI. Materials. Report for Contract NAS1-1S606.
The electroplating process was chosen for Fogdall, L.B., Lindenmeyer, P.H., Sheppard, C.H.,
study because of its low application costs, Russell, D.A. and Cannaday, 5.5.. 1983.
 References 821

Development of Facilities, Quality Control Springer, George S., Sanders, Barbara A and Tung,
Procedures, and Testing Techniques for Randy W. 1981. Environmental Effects on Glass
Irradiation of Spacecraft Composite Materials. Fiber Reinforced Polyester and Vinylester
Report for Contract NASI-16854. Composites. In Environmental Effects on
Fried, N. 1969. Marine Applications. In Handbook of Composites Materials, pp. 126-144. Westport, CT:
Fiberglass and Advanced Plastics Composites, (ed. Technomic Publishing.
George Lubin) pp. 747-783. New York: Van Spry, William J. 1987. Sports and Recreational
Nostrand Reinhold. Equipment. In Engineered Materials Handbook,
Graner, William R 1982. Marine Applications. In Vol. 1, (ed. Theodore J. Reinhart et al.) pp.
Handbook of Composites, (ed. George Lubin) pp. 845-847. Metals Park, Ohio: ASM International.
514-529. New York: Van Nostrand Reinhold. Staunton, R 1982. Environmental Effects on
Kamenetzky, RR and Whitaker, AF. 1992. Composites. In Handbook of Composites, (ed.
Performance of Thermal Control Tape in the George Lubin) pp. 514-529. New York: Van
Protection of Composite Materials to Space Nostrand Reinhold.
Environmental Exposure. NASA TM-I03582. Sullivan, Lawrence J. and Ghaffarian, Reza. 1988.
Kessler, D.J., Reynolds, RC. and Anz-Meador, p.o. Microcracking Behavior of Thermally Cycled
1989. Orbital Debris Environment for Spacecraft High Temperature Laminates. Paper read at
Designed to Operate in Low Earth Orbit. NASA 33th Intern. SAMPE Symp. Exhib., March 7-10,
TM-I00471. 1988, Anaheim, CA
Levine, Arlene, (ed.). 1991. LDEF-69 Months in Tennyson, RC. 1991. Composite materials in space
Space: First Post-Retrieval Symposium. NASA - results from the LDEF satellite. Canadian
CP-3134. Aeronautics and Space J. 37(3):120-133.
Levine, Arlene, (ed.). 1992. LDEF-69 Months in Vette, J.I., Lucero, A.B., Wright, JA, King, J.H. and
Space: Second Post-Retrieval Symposium. Lavine, J.P. 1967. Models of the Trapped
NASA CP-3194. Radiation Environment. NASA SP-3024.
Lisagor, W. Barry. 1979. Mechanical Property Whitaker, A.F. 1991. Coatings Could Protect
Degradation of Graphite/Polyimide Composites Composites from Hostile Space Environment.
After Exposure to Moisture or Shuttle Orbiter Advanced Materials and Processes 139(4).
Fluids. In Graphite/polyimide Composites, (ed. H. Whitaker, A.F. and Young, L.E. 1991. An Overview
Benson Dexter and John G. Davis, Jr.) of the First Results on the Solar Array Materials
pp.273-287. NASA CP-2079. Passive LDEF Experiment (SAMPLE), A0171.
Piellisch, Richard. 1991. New solar arrays mean Paper read at First LDEF Post-Retrieval
new materials. Aerospace America 29(5):20-23. Symposium, June 2-8, 1991, Orlando, FL.
Pilpel, Edward D. 1982. Expanded Design Analysis Whitaker, A.F. 1991. Preliminary Assessment of
of the Use of Composites in Determining Snow LEO Effects on LDEF Experiment AOI71
Ski Characteristics. In Materials Overview for CompOSite Material Surfaces. Paper read at
1982, Soc. Advanc. Mater. Proc. Engng, pp. LDEF Materials Workshop, Nov. 19-22, 1991, at
616-627. Langley, VA.
Schonberg, W. 1990. Hypervelocity Impact Testing Young, Philip R, Slemp, Wayne S., Witte, Jr.,
of Non-metallic Materials. 17th Congo Intern. William G. and Shen, James Y. 1991.
Council of the Aeronaut. Sci., Sept. 9-14, 1990, Characterization of Selected LDEF Polymer
at Stockholm, Sweden, paper ICA5-90-1.7.3. Matrix Resin Composite Materials. Paper read
Sigur, W.A and Gray, C.R 1990. Composite at 36th International SAMPE Symposium and
Applications on the External Tank. Final Report Exhibition, April 15-18, 1991, at San Diego, CA
from Martin Marietta Manned Space Systems,
NASA Contract NAS8-33708.
SAFETY AND HEALTH ISSUES 37
Jennifer A. Heth

37.1 INTRODUCTION Safety and Health Subc9mmittee of the

Suppliers of Advanced Composite Materials
On a weekly basis, there are conflicting media
Association (SACMA) researched and pub-
reports indicating that something is good or
lished Safe Handling of Composite Materials (3rd
bad for us, depending on the study done, the
Edn, Aprll1996). This chapter is an edited ver-
topic and the amount of press given to it.
sion of the booklet, with new information
However, by reviewing the information and
added as appropriate. The information is not
adding common sense and moderation to our
meant to be inclusive. Rather, the reader
lifestyles, we are able to discern good from bad,
should be aware of the issues addressed and
thus allowing us to lead healthy and safe lives.
make further investigation as needed. Note
Working with composite materials, or any
that data cited was current in 1997.
chemicals for that matter, should be viewed in
the same manner. Know what material is used,
how it is handled, what is known about it and 37.2 HEALTH INFORMATION
how to reduce the risk of injury from any haz- TERMINOLOGY
ard associated with it.
From a safety perspective, proven engineer- It is important to understand the terminology
ing and administrative techniques and in order to read and assess toxicological data.
controls exist that can make the workplace This section will concentrate on some of the
safe if implemented correctly. There are fairly basic terms and definitions that are applicable
universal industrial hygiene protocols that to the composite user, especially in reading
merely need to be implemented to effectively Material Safety Data Sheets, the most common
minimize potential exposure to any hazard, source of information on materials.
whether it be chemical or physical.
On health issues, there are numerous toxi-
cological papers published on chemicals used 37.2.1 TOXIC, HAZARD AND RISK
in the composites industry. However, without These three words are used frequently and
a tOxicological background or proper analysis sometimes interchangeably. To avoid confu-
and interpretation of the data, it is difficult to sion, they are defined as follows.
know what the studies' conclusions are and,
more importantly, if they are valid. Toxic refers to a poison or poisonous sub-
To assist users in understanding safety and stance that may cause a harmful effect in the
health issues for composite materials, the body. A substance's toxicity characteristic is a
property of the chemical, similar to its color or
odor. This is as true for chemicals like water,
Handbook of Composites. Edited by S.T. Peters. Published salt or sugar as it is for cyanide, snake venom
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 or botulism toxin (Fig. 37.1).
 Health information terminology 823

Risk describes the probability or likelihood

'All chemicals are toxic, there
are none which is not ...
that a hazard will result in a harmful effect.
Regardless of the toxicity, hazard or risk
The right dose differentiates a
poison from a remedy.'
associated with a substance, care should
always be taken to minimize exposure. This is
Paracelsus
16th Century because what is known today may be inade-
quate compared to what tomorrow's science
may discover (e.g. asbestos usage and asbesto-
sis or tobacco and lung cancer).

37.2.2 ACUTE TOXICITY VERSUS CHRONIC

TOXICITY

When discussing toxicity, reference will typi-

cally be made to two types:· acute or chronic.
Fig. 37.1 Toxicology principle. One can be instantaneous and the other mani-
fests itself over a long period of time.
Hazard takes into account the toxicity of a sub- Acute toxicity occurs when a harmful effect is
stance along with exposure to it. Something experienced after a single or short-term expo-
can be extremely toxic, but if there is no expo- sure to a substance. It usually occurs instantly
sure, there is no hazard (Fig. 37.2). Conversely, or within a short time period. Typically, expo-
it is important not to be negligent in handling sure will be by skin or eye contact or
low toxicity substances, as extreme exposure inhalation. In industrial settings, it would be
(high concentration or long exposure time) unlikely for exposure to be by ingestion,
could result in high hazard. though this is another route.

o° 0
o
o
o.

° 0
0 0
x Minimal hazard
Low exposure
o
o •

4A1
High toxicity

·m 0°.·
0°.
Extreme exposure

Rlr
•• 0.
00
x Elevated hazard
.
• 0
o•
.
Low toxicity

Fig. 37.2 Risk assessment equations.

824 Safety and health issues

The most common measure of acute toxicity (possibly years or decades). Because chronic
through the oral or dermal routes is called the effects develop slowly and are measured later,
median lethal dose or LDso' The LDso (mg of and it is not possible to keep the workplace
chemical per kg of body weight) is the amount toxic-free, it is important to eliminate or mini-
of material that kills 50% of a group of experi- mize exposure.
mental animals, usually mice or rats (Fig.
37.3). The measure by inhalation is LC5Q1 or
37.2.3 SENSITIZATION
median lethal concentration. It is expressed as
an airborne concentration in milligrams of In some cases, an allergic reaction to a sub-
chemical per cubic meter (mg/m3) of air, or stance will develop with one exposure, or over
parts per million (ppm) in the air (Fig. 37.3). time with repeated exposures. This is called
sensitization. Once a person is sensitized to a
• Oral and dermal routes: substance, the extent of the reaction does not
necessarily relate to the degree of exposure.
Median lethal dose - LD50 (mglkg) Also, people who are sensitized to one chemi-
A large LD50 (e.g., 5000 mg/kg) equals a low degree cal may react to other similar materials. This is
of acute toxicity, typically relating to a low health known as cross sensitization (Fig. 37.4).
hazard. Certain individuals may be sensitized to some
chemicals used in the composites industry. For
• Inhalation route: example, there have been reported cases of
Median lethal concentration - LC 50 (mglm3) epoxy sensitization. In these instances, it is
For rats, a four hour exposure period is commonly important that an employee not be exposed
used. Because other time durations are used, data further, or allowed the opportunity to be
will usually be reported with time specified. exposed. To do this, engineering or adminis-
trative controls should be implemented. If
these are not possible or practical, then the
Fig. 37.3 Acute toxicity measures.
employee should be removed from chemicals
causing reaction.
The acute inhalation hazard is dependent on
the material's toxicity, together with its physi-
cal properties, such as vapor pressure for gases 37.2.4 EXPOSURE LIMITS
or particle size for aerosols, particulates and
There are limits or values (Fig. 37.5) estab-
dusts. Knowledge of the physical properties
lished by health groups to assist the user in
will assist in determining whether a material is
controlling exposure to certain chemicals;
likely to become airborne, thus inhalable.
some are regulated (OSHA's Permissible
Chronic toxicity occurs when adverse health Exposure Limit, PEL), others are recom-
effects are manifested after exposure to a sub- mended (ACGIH's Threshold Limit Value,
stance over a long period of time (e.g. TLV). These limits are based on actual indus-
throughout a working lifetime) or by long-term trial experience, experimental animal and
effects resulting from one or a few doses. These human studies and, when possible, a combi-
effects can occur following repeated exposures nation of all three.
to chemical substances through dermal or ocu- Exposure limits are developed for protec-
lar contact, inhalation or ingestion routes. tion against serious health effects or irritation,
Chronic toxicity testing includes systemic narcosis, or nuisance. They are intended for
toxicity, mutagenicity, carcinogenicity, repro- the control of potential health hazards in the
ductive toxicity and sometimes epidemiological workplace. Because each individual is differ-
studies, all of which are very time consuming ent, there will be incidents of people affected
 Industrial hygiene 825

Once sensitized (have an allergic reaction), the following may occur:

x Allergic reaction
Low exposure
Low toxicity

4A1
nlr
Extreme exposure
x Allergic reaction

cu·
Low toxicity

At
'. .
. +..
•• +.

,
. .
Chemicals similar to x Allergic reaction
substance causing Low exposure
sensitization
Fig, 37.4 Sensitization equations.

by some substances at concentrations below 37.3 INDUSTRIAL HYGIENE

the PELs or TLVs (Fig. 37.5). This is because the
Potential hazards can be controlled through the
limits/values are based on the exposed popu-
implementation of sound industrial ~ygi~e
lation being 'normal', and does not address
practices. Whether employees are working ill a
aggravation from pre-exis~g conditio~, ~l­ high or low hazard area, the recognition, evalu-
nesses or lifestyle choices. It IS best to mamtam
ation and control of conditions are necessary to
concentrations of atmospheric contaminants at
avoid occupational illness or discomfort. This is
the lowest level reasonably possible through
industrial hygiene.
process or engineering controls. , ., It is important to apply good engineering,
When referencing PELs and TLVs, a skin
process and administrative controls in the work-
notation may be assigned. This indicates that
place along with effective industrial hygiene
dermal exposure should also be considered
practices (e.g. monitoring, surveys, proper per-
when evaluating exposure to that substance.
sonal protective equipment usage). Without
Always check the current OSHA standards for
addressing the fundamentals, it is difficult to
PELs and the latest edition of the ACGlli's
maintain employee comfort and, hence, protec-
TLVs: Threshold Limit Values and Biological
tion against exposures that may be present.
Exposure Indices, along with the supplier's
Industrial hygiene practices for composites
Material Safety Data Sheet, for information on
are common to those for other industries.
limits and protection against exposure.
826 Safety and health issues

Exposure limits
PEL (Permissible Exposure limits) Airborne exposure limits issued by the Occupational Safety and Health
Administration (OSHA), 29 CFR 1910.1000, which are legally binding.
TLV (Threshold Limit Values) Recommended exposure limits for airborne concentrations of substances.
They are issued by the American Conference of Governmental Industrial Hygienists (ACGIH) and pub-
lished annually.

Exposure limit descriptions

Both PEL and TLV are expressed in similar terms describing exposures:
TWA (Time Weighted Average) This represents a normal 8 h day and a 40 h work week, to which nearly ali
workers may be exposed without adverse health effects. In literature, data will appear as PEL-TWA or
TLV-TWA.
STEL (Short-Term Exposure Limit) Concentrations to which people can be exposed for a short period of
time without suffering from
• irritation;
• chronic or irreversible tissue damage; or
• narcosis of sufficient degree to increase likelihood of accidental injury, impairment of self-rescue or
reduction of work efficiency
provided the daily PEL or TLV-TWA is not exceeded.
Data will appear as PEL-STEL orTLV-STEL in written works.
C (Ceiling) The concentration not to be exceeded during any part of the working exposure. This will
appear as PEL-C or TLV-C in documentation.

Fig. 37.5 Exposure limit terminology.

Periodic exposure assessments should be con- or a new process is to be used. After any new
ducted taking into consideration routes of process or modification of an old one, good
exposure associated with composite use (Table work practice is to ensure that existing con-
37.1). Monitoring should be routine, and on trols to prevent exposure are implemented
demand whenever unusual odors are noticed, and effective.
visible contamination is heavier than normal, Two potential exposures posed by composite

Table 37.1 Routes of exposure for composite users

Skin and eyes Typically hands, lower arms and face are exposed. However, if personal hygiene is not
good, other areas of the body may be susceptible to exposure. Avoid exposure,
especially in cases where dermatitis or sensitization has been confirmed. Avoid contact
with chemicals that can be absorbed through the skin, as systemic and! or local
effects may occur.
Inhalation Good ventilation will minimize possible exposure from the release of solvents or dusts
generated.
Ingestion Thorough washing of the hands prior to eating or smoking provides significant
protection from the effects of accidental ingestion.
Injection Needles and sharps are not normally a concern. However, shards from cured
composites or brittle fibers, or needles from weaving or sewing processes can
puncture the skin and chemicals could enter the body.
 Industrial hygiene 827

usage are skin contact with materials that controls are defined as 'methods of controlling
could result in irritation leading to dermatitis employee exposures by job rotation, work
or sensitization, and inhalation of particulates assignment, or time periods away from the haz-
from operations such as cutting, grinding and ard'. These controls may be implemented when
finishing. Both concerns can be eliminated or engineering controls cannot reduce the exposure
minimized with implementation of proper to permissible levels (Fig. 37.6). Administrative
gloves/clothing, good ventilation and process controls also encompass 'other work rules', such
conditions, and effective training. as company policies (Table 37.2).
One example of a process hazard in the
composites industry, controllable through
37.3.1 ADMINISTRATIVE CONTROLS
engineering and administrative rules, is an
In the National Safety Council's Fundamentals out-of-control exothermic reaction. This is an
of Industrial Hygiene (3rd Edn), administrative unintentional runaway chemical reaction of a
Table 37.2 Workplace tools to prevent exposures

1. Safety controls Confined space entry; hot work permits; lockout/tagout procedures; pipeline
breaking permits; safe work permits.
2. Personal controls Knowledge and use of Material Safety Data Sheets (MSDS); company policies
and procedures; regulation of employee's exposure time in a work area.
~----------------

3. Education and Training in the use and handling of materials and any associated hazards;
training understanding of the training.
4. Materials handling Information from MSDS on proper handling of materials
5. Process controls Isolation of process controlled work areas; eating, drinking, smoking should be
separate from any work area as should food storage; control of process-related
hazards, such as an out-of-control exothermic reaction potential. There should
be procedures, equipment, ventilation systems and process hazard analysis
and training in place; wet processes to reduce dust generation.
6. Personal hygiene Thorough washing of hands before each job break and eating/ drinking/ smoking
and toilet use along with use of good hand creams after each wash.
~~~~~~~~~---

7. Warnings Proper labels on chemicals; follow OSHA's Hazard Communication

and labels Standard.
8. Housekeeping Keep work areas clean and free of clutter; remove dust by vacuuming instead of
blowing it off work areas into the air.
9. Emergency plans Plans and procedures for dealing with any chemical emergency; personnel
should have the knowledge, skills and equipment to respond appropriately.
10. Ventilation General, diluted (supplied air) and local exhaust ventilation.
11. Medical controls Good occupational medical program and protocol.
12. Engineering Design, maintenance and hazard studies; shielding, monitoring devices, and
practices alarm systems.
13. Personal protective Respirators, gloves, body suits, boots, jackets, goggles, eyewear, face shields,
equipment safety shoes/boots, headgear.
14. Administrative Job rotation, scheduling machine times to reduce number of workers exposed,
controls scheduling work times to avoid exposures (e.g. cooler times of day to avoid heat
stress), reduction of work periods.
828 Safety and health issues

Exothermic reactions can be avoided and min-

imized by clearly defining and following the
proper procedures, thoroughly training
employees who work with resins and prepreg
materials, making certain that equipment is in
good working order, and that safety devices or
systems (e.g. quench systems such as alkyl
Personal protective equipment (PPE) benzyl phthalate, ABP) are in place and func-
tioning properly.
Fig. 37.6 Industrial hygiene order of priorities.

resin system, alone or in prepreg form, typi- 37.3.2 PERSONAL PROTECTIVE EQUIPMENT
cally called an exotherm. It may occur under (PPE)
any of these conditions: Personal protective equipment is used to con-
• heating or mixing a resin too long; trol exposures when they cannot be reduced to
• heating a resin too fast; an acceptable or practical level using engineer-
• allowing a resin to get too hot; ing and administrative controls. In the
• contamination or mislabeled chemicals. composite industry, there are three main expo-
sure areas that may require PPE to protect the
These factors have been known to start or skin, eyes and lungs.
extend an exotherm:
• resin mass scale up and equipment cannot 37.3.3 SKIN PROTECTION
dissipate heat;
• deviating from procedures; Hand contact exposure is most common when
• disabling safety equipment; using composites. The use of proper protective
• casting hot melt resins too deep; gloves is important to control exposure. Along
• process equipment malfunction; with gloves, other equipment such as jackets,
• variability in raw materials; arm protectors, or body suits may be neces-
• mixing incompatible chemicals and curing sary depending on the process or potential for
agents; exposure.
• contaminating chemicals, e.g. poor house-
keeping;
Gloves
• uneven dispersing or mixing of chemicals;
• trying to mass cure resin or prepreg in an Glove selection should take into account the
oven or autoclave. parameters in Table 37.3 to ensure the correct

Table 37.3 Proper glove selection criteria

Chemical resistance Glove must be impermeable to the chemical being used. An incorrect glove
choice may allow the material to come in contact with the skin.
Dexterity and If the user cannot work comfortably with the glove, it will not be used.
comfort
Glove lining Lining material may cause irritation or excess sweating. Consider liners or other
alternatives.
Glove surface The outer surface may contain residual coatings that may contaminate and
possibly delaminate a composite part.
 Industrial hygiene 829

glove choice and use by the employee. aramid protect well against heat, cuts or
Additionally, Table 37.4 gives a brief listing of scrapes. Glove choice must depend on the
glove types used in the composite industry job and its duration.
and what they protect against. SACMA's com-
panion video to the Save Your Skin! booklet
demonstrates correct glove selection and Skin creams
usage for composite users.
There are creams that are applied directly to
It is important to know that there is not an
the hands and forearms for protection. Two
'all purpose' glove for protection from all
types of cream discussed in Table 37.5 indicate
exposure. Proper protection may require
the difference between moisturizers (no pro-
wearing one type of glove over another.
tection against chemicals) and barrier creams
Basically, there are two glove types:
(limited protection). Barrier creams may be
1. Chemically resistant: used alone or in combination with gloves.
A selection of gloves used for protection When using barrier creams, there is a possi-
against exposure to chemicals. Depending bility of contamination to the composite
on the permeability of the glove material to material. If there is a heavy reliance on the use
the chemical used, the glove chosen will of barrier creams, process engineers need to
protect the wearer for only a limited time. investigate alternatives in engineering, process
Even the proper glove does not last forever. changes or alternate PPE. Since barrier creams
2. Mechanically or thermally resistant: are not moisturizing creams, a good moisturiz-
These gloves, such as leather, cotton or ing cream should also be applied after work to

Table 37.4 Examples of glove types

Glove type Hazard ...

Aramid Cutting, intense heat
Cotton Abrasions
Disposable plastic (Latex) Microorganisms, mild irritants, fibers
Natural rubber (Latex) Acetone « I h), epoxies, methyl ethyl ketone (MEK), light work
Leather Abrasions, punctures, fibers
Metal mesh Cuts, scrapes
Neoprene Acetone « I h), solid epoxies, DMSO, isopropyl alcohol
Nitrile Solid epoxies, TDI, isopropyl alcohol
Polyvinyl alcohol Methylene chloride, bisphenol A epoxies, toluene, MIBK, TDI,
styrene, THF « 4 h), l,l,l-trichloroethane
Polyvinyl chloride TDI, isopropyl alcohol, « 4 h)
Rubber, butyl Isopropyl alcohol, Dimethylformamide (DMF), DMSO. MEK, MIBK,
acetone, N-methyl-2-pyrrolidone (NMP)
Rubber, insulated Electrical shocks and burns
VITON@ Toluene, xylene, styrene, isopropyl alcohol, l,l,l-trichloroethane
Sources: Save Your Skin!, SACMA 1990; Safe Handling of Advanced Composite Materials, SACMA 1996.
830 Safety and health issues

Table 37.5 Skin creams

Moisturizing creams Replenish moisture in the skin which is lost after was~g up. Use regularly to
avoid dry skin conditions which could lead to dermatitis. Gloves must always be
worn when moisturizing creams are used.
Barrier creams 1. Water repellent creams: Leave a thin film of lanolin, b~eswax, petroleum or
silicone on skin. Used mainly in machine shop operations. .
2. Solvent repellent creams: Leave a thin film on the skin, visible or not, which
will repel oils, paints, and solvents. Barrier creams do wear off and must be
reapplied for constant protection.

keep skin healthy. Additionally, barrier creams 37.4 POTENTIAL HEALTH AND SAFETY
may actually increase abrasion from fiber dust, HAZARDS IN COMPOSITE PROCESSES
so their effectiveness needs to be monitored.
When designing equipment, processes and
modifications, complete containment of
37.3.4 EYE PROTECTION vapors and dusts should be a goal. General
ventilation should be provided to all work
Eye protection should be selected based. on areas, with local exhaust equipment designed
impact (flying particles) and/or chemIcal
to pull contaminants away from the
splash possibilities. Selections should include
employee's breathing zone. In almost every
appropriate safety glasses, goggles, face composite process, the engineering emphasis
shields or a combination of these. is on good ventilation to control solvent and
dust exposures, along with other contami-
37.3.5 RESPIRATORY PROTECTION nants.
Figure 37.7 gives a summary of potential
During operations such as resin mixin~, health and safety exposures that could occur
prepreg lay-up, machining or clean-up, respI- in composite processes. However, with good
ratory protection may be necessary to reduce engineering, administrative and industrial
exposures to vapors or dusts. hygiene practices, personal exposure can be
There are two main types of respirators: air minimized.
purifying and air supplying. Before using any Process improvements are driven by tech-
respirator, familiarity with OSHA's Respiratory nology and regulation (e.g. CNs South Coast
Protection Standard (29CFR 1910.134) is essen- Air Quality Management District Rules 1171
tial. Not everyone can or should use a
and 1128 are for emission reduction of volatile
respirator. An employee must be medically organic compounds in coating and cleaning
approved, fit tested, and trained to assure that operations). In the composites industry, some
the respirator used is both appropriate and examples of minimized solvent vapors and
protective. exposures include closed loop systems, cabin
If using a cartridge (filter) respirator, the surroundings for processes, improved ventila-
cartridges are specific to hazards and must be
tion designs and capture efficiencies, and
maintained and changed periodically to pre- elimination of solvents or substitution of low
vent exposure. If they are the wrong type, are vapor pressure solvents for high vapor pres-
used too long, or become dirty, respirators are sure solvents.
ineffective.
 Toxicological properties of composite components 831

COMPOSITE PROCESSES
1..0
0';:;
~0
,,"6
~ 0~ ,>q ~ .".&
o~
, (;:'
. ~~ ~"
~C$. ~lS «\'li ,,"6
~
~C}
iF 'b-
~.;:;
.#'. ~~
~ ~\'li ~'" d- ~0 I.. i:P~ ....~~ «.,(;:'
if','§ «\0 0~ ,>q 0'""6 ~q.'li C} ~~C$ 'b- ~
«.0 ~'b- .....'Ir~ ,,'-\ r'8-'li "'..."6 "..o~ . ~~ ~<Q . ~~~ Q;-0q
.~ .~«'''' 0 ~ .~ ~~ (to ,:) ~ 0
POTENTIAL HEALTH HAZARDS Q;-0~ «~~ ~e- . ~.f ~.s-0 Q;-0~ «~ ,i::-~ ~'lrv';:; ~'lrv «'~

1. Skin Contact, Dermatitis/Sensitization X X X X X X X X X X

2. Dust and/or Solvent Inhalation X X X X X X X X X X X
3. Systemic Effects X X X X X X X X
4. Eye Contact X X X X X
5. Smoke Inhalation (Exothermic) X X X X X X X
6. Oxygen-Deficient Atmosphere X
7. Noise X X
POTENTIAL SAFETY HAZARDS
8. Mechanical X X X X X X X X X
9. Electrical, Static Electricity X X X X X X X X
1O. Housekeeping X X X X X X X X
11 . Exotherms X X X X X X X
12. Flammabilitv/Exolosion. X X X X X X X X X X
13. Cuts X X X X X X
14. Burns X X X X X
15. Laser Cuts/Burns X
16. High Pressure Water X

Fig. 37.7 Potential health and safety exposures in composite processes.

37.5 TOXICOLOGICAL PROPERTIES OF effects, the resin system is of primary concern.

COMPOSITE COMPONENTS The reinforcement is secondary due to its lower
hazard potential compared to resin systems.
The supplier's Material Safety Data Sheet
(MSDS) and other sources of toxicological Solvents are addressed as a separate category as
information should be used for practical haz- they are major sources for potential skin and
ard assessment on any material. Rarely will inhalation exposures. Some are part of the resin
systems, while others are only used for clean-
data be available on the composite system
itself because the methodology has not been up purposes.
fully developed to adequately test these kinds
of materials. Therefore, the hazards of com-
37.5.1 RESINS
posites are generally expressed in terms of the
components' hazards. This is a conservative Composites are by definition based on a resin
approach that provides the user with the most system matrix applied to some reinforcing
protective information because it is based on material. The resin system is often quite com-
the most hazardous component which may plex, consisting of a basic resin (e.g. epoxy),
only be a percent or less of the entire mixture. which is formulated with other materials,
Typically, when discussing toxicological such as curing agents, diluents, accelerators,
832 Safety and health issues

pigments or solvents. Components of the any product containing additives with known
resin system may be supplied individually or potential health effects. Reference to the
and formulated by the user, supplied as 'Part MSDS will advise you of hazardous materials.
A and Part B' and blended before use, or sup- Some commonly used additives are listed in
plied mixed as a 'one-pack' system. Table 37.7.

Epoxy resins Other resin types

Epoxy resins, also known as glycidyl com- Over the years, epoxies have been the back-
pounds, are commonly used in composite bone of the composite industry. However,
matrix resin systems. In addition to the stan- other resin systems play an important role and
dard bisphenol A based epoxies, other are used for particular applications and perfor-
glycidyl ethers, glycidyl esters, glycidyl mance properties. Table 37.8 discusses health
amines, and epoxy novolacs are used in com- effects of some of these other resins. The asso-
posite matrices. ciated health hazards may appear severe, but
The greatest concern with epoxies is their exposures are avoidable through engineering
potential to cause skin irritation, dermatitis or and industrial hygiene practices, especially
skin sensitization, depending on the base good ventilation for control of vapors.
epoxy used and the sensitivity of the individ-
ual exposed. Table 37.6 discusses epoxy based 37.5.2 REINFORCING MATERIALS
resins and associated hazards.
Most reinforcement materials (Table 37.9) are
fibers, such as graphite or carbon, glass,
Hardeners/curing agents/catalysts
aramid or ceramic. Others may be used, but
One or more of these additives are used to the application is typically specialized, result-
enhance properties of composite materials. ing in small specialty runs. Overall, data
Their percentage of the resin system may be indicates that most fibers have a low hazard
small, but precautions ought to be taken for potential in initial form. However, they may

Table 37.6 Health effects of various epoxies

Epoxy type CAS number(s) Known health effects Key notes

Bisphenol A based 1675-54-3, Possible skin sensitizer; Insufficient evidence to classify as
25036-25-3, low order of acute a carcinogen according to lARC.
25068-38-6, toxicity; slightly to Considering the many studies as
25098-99-8 moderately irritating. a whole, the evidence does not
show the resins to be carcinogenic.
Glycidyl amines 28768-32-3 Possible skin sensitizer; Mutagenicity tests gave both
low order of acute negative and positive results.
toxiCity.
Cycloaliphatics 2336-87-0, Irritant to skin and mucous Not considered mutagenic or
30583-72-3 membranes. carcinogenic.
Glycidyl ethers 2210-79-9, Possible skin sensitizer. Neopentylglycol diglycidyl ether
2426-08-6, Moderate to severe skin has caused skin tumors when
3101-60-8, and mucous membrane applied repeatedly to skin of
17557-23-2, irritant. shaved mice.
26447-14-3
Table 37.7 Health effects of hardeners/curing agents

Hardeners/ CAS Known health effects Key notes

curing agents number(s)
Aromatic amines Slightly irritating to skin and mucous Several amines are suspected carcinogens.
membranes; can cause damage to internal
organs (liver) and may decrease ability of
blood to transport oxygen to tissues.
1. Methylene dianiline; 101-77-9 Liver toxin. Potential for thyrOid, spleen, red Possible carcinogen (NTP Group 2, IARC 2B). No
4,4'-methylene blood cell, kidney and eye (photoreceptors) confirmed reports of cancer in humans. TLV-TWA
dianiline, MDA damage after oral, dermal or respiratory routes. is 0.81 mg/m3 or 0.1 ppm, skin; PEL-TWA is 0.08
Potential skin sensitizer. mg/m3 or 10 ppb, STEL is 0.8 mg/m3 or 100 ppb.
2. 4,4'-sulfonyl dianiline, 80-08-0 Blood disorders after chronic oral exposure; Not classified as a carcinogenic; used over 50
DDS; Dapsone: 4,4'- nausea, vomiting and a host of isolated years as drug treatment for leprosy and certain
diaminodiphenylsulfone complaints after oral dosage. types of chronic dermal inflammations.
3. M-phenylenediamine 108-45-2 Skin and mucous membrane irritant, potential Manufacturers recommend TWA of 0.1 mg/m3,
skin and respiratory sensitizer. Liver toxin. skin.
4. Toluenediamine 25376-45-8 Skin and mucous membrane irritant; potential ~
Some isomers may be carcinogenic. l-!
isomers skin and respiratory sensitizer; liver and §.
kidney effects. Decreases ability of blood to ~
transport oxygen to tissue. [
Aliphatic and 111-40-0, Strong bases and severe skin and eye irritants; Diethylene triamine PEL and TLV-TWA is 1 ppm,
cyc10aliphatic amines 112-24-3, can cause skin and/ or respiratory sensitization; skin. 1
140-31-8, certain types can cause temporary visual ~
~.
694-83-7 disturbances with airborne material affecting CIj

the cornea. -Q..

Polyaminoamides 68410-23-1 Possible skin and eye irritation. 8
Polyamides 68082-29-1 Possible skin and eye irritation.
~
1il
Anhydride curing agents 85-43-8, Severe eye and skin irritants; irritating effect on ~
85-42-7 skin, mucous membranes, and respiratory tract
26590-20-5, (1st 3 CAS); potential skin sensitizer; some
552-30-7, respiratory sensitization (last 2 CAS).
89-32-7 (ri
i
ao
w
VJ
Table 37.S Health effects of other resin types
~
Other resin types CAS Known health effects Key notes en
number(s) ~
<"I>

Polyurethanes ~
;:,
1.Isocyanates Strong/severe irritants to skin and mucous membranes of eye and Possible human carci- s.
respiratory tract, skin and respiratory sensitization; systemic toxicity. nogen (IARC Group ~

2B; NTP Group 2).

Toluene diisocyanate, TOI 584-84-9 ~
1,6-Hexamethylene 822-06-0 Coughing; chest discomfort; shortness of breath; reduced lung function; <n'
<I>'
diisocyanate, HDI sensitization to skin. :::::
til
2. Polyols Polyether and polyester polyols are used in crosslinking components for
isocyanates. No particular health hazards in industrial processing.
However, some products contain unreacted ethylene oxide.
Phenolic and amino resins
1. Phenol-formaldehyde 9003-35-4 Hazards are considered low. Due to toxicity and absorption of phenol
resins 37337-65-8 through skin, contact should be avoided.
Formaldehyde 50-00-0 Skin, eye and respiratory irritant; potential skin sensitizer. Overexposure Probable human
may cause liver and kidney effects. carcinogen. (IARC 2A,
NTP Group 2, ACGIH
A2,EPAB1).
2. Urea- and melamine- Same effects as others. Guard against off-gassing for inhalation route of
formaldehyde resins exposure.
Imides
Bismaleimide, BMI; Skin irritation and sensitization; dust or vapor may cause eye, nose, or
polyimides and others throat irritations.
Thermoplastic resins Fairly inert; watch for unreacted monomers. Vapors generated could cause
eye and sinus irritation. Thermal bums are major effect.
1. Styrene (monomer) 100-42-5 Vapors cause eye irritation; liquid contact will cause skin, eye and mucous Possible human carci-
membrane irritation; possible systemic effects on central nervous system, nogen (IARC Group
liver and kidney. 2B). Has an odor
threshold at 0.3 ppm
high concentrations.
TLV-TWAis20ppm,
STEL is 40 ppm
2. Polyphenylene- 26125-40-6 Off-gas materials may cause headaches, dizziness and elevated blood
sulfide, PPS levels of carboxy-hemoglobin; irritation of eyes, nose, and throat; H 2S off
gases may paralyze the respiratory system.
 Toxicological properties of composite components 835

Table 37.9 Health effects of fibers

Reinforcement type Known health effects Key notes

Carbon or graphite Mechanical abrasion and irritation PEL-TWA is 15 mg/ m 3 total dust, and PEL-
fibers of the skin; possible dennatitis; TWA of 5 mg/m3 for synthetic graphite
physico-mechanical properties respirable dust. ACGlli has a TLV-TWA of
of the fibers rather than a 2 mg/m3 respirable dust for all fonns of
toxico-chemical reaction. Possible graphite except fibers. There are no limits for
reaction from the fibersizing. carbon fiber, though the U.S. Navy has set 3
See resin health hazards. carbon fibers / cc. EPA did not classify the
potential carcinogenic properties of
carbon fibers due to insufficient data.
Glass fibers Mechanical irritation Continuous fiber (used in composites, >6 J.!m
of eyes, nose, throat; in diameter) is probably not carcinogenic
pOSSible skin sensitization, (!ARC, 1988). Wool fiber is classified as a
either from sizing or fiber. possible human carcinogen (!ARC Group 2B).
Para-aramid fibers Minimal evidence for skin Commercial fibers are too large to inhale, but
irritation, none for skin RFP can be created when the fiber is abraded
sensitization. Prolonged or cut. Airborne RFP concentrations from
overexposure to respirable composite machining have been acceptably
fibrous particles (RFP) low. Inhalation studies in rats demonstrated
has potential for lasting fibrous particle breakdown in the lungs and
lung damage. no carcinogenicity. !ARC classifies para-
aramid RFP in Group III (not classified as a
carcinogen).
Ceramic fibers Skin, eye, or upper respiratory EPA proposed ceramic fibers, such as refrac-
irritation is possible. tory aluminum oxide and zirconium oxide, as
not classifiable to human carcinogenicity.
Also, EPA proposed that aluminum silicate
fiber be classified as a human carcinogen.
Refractory ceramic fibers are classified
as 2B by !ARc.

be chemically coated or stiff enough to cause believed to be 3.5 11m or less. Anything larger
irritation by penetrating the skin or tissues of than that will be removed from the body via
the nose, throat or bronchi. Little has been nose and throat functions. If the fibers are res-
studied and is known about cured materials pirable, the toxic effects may vary
being ground, drilled, milled, cut or sanded. significantly. There are low risk fibers (irritants
Should fragments of fibers be small enough to such as fiberglass) and there are significant
be respirable, there is concern that a general or risk fibers that can result in asbestosis or can-
fibrous dust hazard to the lungs can occur. cer (such as asbestos fibers). Therefore, each
Therefore, precautions should be used to min- fiber should be assessed on its own toxicolog-
imize exposure. ical properties.
Knowledge of the fiber used should include:
length, diameter, aspect ratio and fragmenta-
37.5.3 SOLVENTS
tion propensity. Knowledge of fiber
parameters, along with how the composite will Solvents are used in many aspects of compos-
be handled, is critical to protect from exposure. ites manufacturing, from resin formulation to
Fiber diameter size for respirability is clean-up activities. There are several groups of
836 Safety and health issues

Table 37.10 Health effects of solvents

Solvents CAS Known health effects Key notes

number(s)

Ketones Mild to moderate skin irritant; moderate to

severe eye irritant; if overexposure by
inhalation, possible central nervous system
(CNS) depression.
1. Acetone 67-64-1 Irritation of mucous membranes; headache, PEL-TWA of 1000 ppm and TLV-
2-propanone, DMK nausea. Skin contact can cause defatting, TWA of 500 ppm. ACGIH
dermatitis. Systemic effects only after repeated indicates a TLV-STEL of 750 ppm.
overexposure. If ingested, vomiting may cause
acute chemical pneumonitis (lung damage).
2. Methyl ethyl 78-93-3 Few ill effects have been reported. PEL and TLV-TWA of 200 ppm
ketone,MEK Objectionable odor is reported. MEK has an and TLV-STEL of 300 ppm. At the
2-butanone odor threshold of 0.25-25 ppm. Eye, nose TLY, workers complain of odor,
and throat irritation at greater than 200 ppm. but few ill effects have been
reported.
3. Methyl isobutyl 108-10-1 Moderate to severe eye irritation; slight to PEL-TWA of 100 ppm and
ketone, MIBK moderate skin irritation; toxic by ingestion and TLV-TWAof 50 ppm and TLV-
dermal exposure; may cause CNS depression. STEL of 75 ppm.
Overexposure may cause kidney and liver effects.
Chlorinated solvents Exposure to high concentrations can cause cardi-
ovascular effects (sensitization of the cardiac
muscle); CNS depression; chronic animal
exposures have caused liver and kidney changes.
1. Methylene chloride 75-09-2 Vapors below TLV levels; no adverse health TLV-TWA of 50 ppm, A2. PEL-
Dichloromethane responses expected; overexposures may cause TWA at 25 ppm, PEL-STEL of
possible respiratory irritation due to vapors 125 ppm. Possible human
which could lead to delayed pulmonary edema carcinogen. IARC Group 2B, NTP
and CNS depression. Liquid may cause skin Group 2.
and eye irritation.
2. l,l,l-trichloroethane 71-55-6 Drowsiness; overexposures can cause CNS PEL and TLV-TWA of 350 ppm
Methyl chloroform depression which could lead to respiratory and TLV-STEL of 450 ppm.
arrest; animal studies indicate possible liver NIOSH has a recommended
and kidney damage. exposure limit of 350 ppm as a
15 minute ceiling.
Other solvents
1. Dimethylformamide, 68-12-2 Irritating to the skin, eyes and mucous Possible human carcinogen (IARC,
DMF membranes; stomach pain and cramps; nausea Group 2B) PEL and TLV-TWAof
and vomiting; DMF is readily absorbed through 10 ppm (skin). A lethal single oral
the skin and may aid in absorption of other dose of DMF for humans is
materials; possible link to testicular cancer. estimated to be 10 g.
2.- N-Methylpyrroli- 872-50-4 Severe dermatitis possible; irritating to skin, Maximum airborne levels not
done,NMP severe eye irritant; inhalation of high concen- established. Manufacturers
trations can cause headaches, giddiness, mental suggest a TWA limit of
confusion, nausea, gastric upset, and vomiting; 10-100 ppm (vapor), TWA of
eye irritation. 5 mg/m3 (mist).
 References 837

solvents, ketones and chlorinated, which have enforcement' safety philosophy to control haz-
dominated the industry. Known health effects ards. This encompasses all the workplace tools
of major solvents are outlined in Table 37.10. and techniques discussed.
With increased concern for safety, health Safety, encompassing health, should be
and environmental impacts of solvents in the treated as a value. It is not a priority to be
workplace, the use of solvents is decreasing raised or lowered with business cycles or man-
and exposures are being minimized. When agement changes. It should be a constant
choosing a solvent, make certain that regula- philosophy exhibited each day (Fig. 37.8). If it
tory investigation is done to avoid is a value in the composites industry, then the
unnecessary rework. It is possible that a sol- technology of composite materials can
vent of choice may become obsolete due to develop naturally, without unnecessary hin-
emission regulations or toxicity concerns. drances impeding growth.

37.6 CONCLUSION REFERENCES

Every industry or business has hazards. The American Conference of Governmental Industrial
maturity of the industry, its products, and his- Hygienists. 1997. 1997 TLVs® and BEIs®,
tory are all factors in what the hazards or risks Threshold Limit Values for Chemical Substances and
Physical Agents, Biological Exposure Indices.
are and how they are controlled. ACGIH.
There are known potential health and safety International Agency for Research on Cancer, World
hazards in the composites industry, even Health Organization. 1997. [ARC Monographs on
though it is a young technology compared to the Evaluation of Carcinogenic Risks to Humans.
mature ones such as steel. As the composites Vol. 68. IARC.
industry develops, processes may change and Laing, Schmidt. 1992. Accident Prevention Manual for
more specific health and safety issues may be Business & Industry, Administration & Programs,
10th Edition. National Safety Council.
identified. Until then, each composite user has Laing, Schmidt. 1992. Accident Prevention Manual for
the opportunity to implement the National Business & Industry, Engineering & Technology,
Safety Council's 'engineering, education and 10th Edition. National Safety Council.
OSHA Instruction CPL 2-2.20B CH-2. 1993. Polymer
Matrix Materials: Advanced Composites, Chapter
16. OSHA.
Although safe and healthful working conditions can be
Plog, Benjamin, Kerwin, Schonfeld. 1988.
justified on a cold dollars-and-cents basis, I prefer to
Fundamentals of Industrial Hygiene, Third
justify them on the basic principle that it is the right Edition. National Safety Council.
thing to do. In discussing safety in industrial Suppliers of Advanced Composite Materials
operations, I have often heard it stated that the cost of Association. 1996. Safe Handling of Advanced
adequate health and safety measures would be Composite Materials, Third Edition. SACMA.
prohibitive and that 'we can't afford it: Suppliers of Advanced Composite Materials
Association. 1990. Save Your Skin! A Guide to the
My answer to that is quite simple and quite direct. It is Prevention of Dermatitis. SACMA.
this: 'If we can't afford safety, we can't afford to be in
business:
Admiral Ben MoreeU ACKNOWLEDGEMENT
President
Jones and Laughlin Steel Corp., 1948 Thanks and appreciation to the chemists,
industrial hygienists, toxicologists, and safety
Fig. 37.8 Safety philosophy (Laing and Schmidt, professionals in the composite industry, par-
1992). Used by permission of the National Safety ticularly those in SACMA's Environmental,
Council, Itasca, Illinois. Safety, and Health Committee.
NONDESTRUCTIVE EVALUATION METHODS 38
FOR COMPOSITES
Thomas S. Jones

38.1 INTRODUCTION of constituents (fiber-resin ratio, resin-starva-

tion, etc.), orientation and distribution of
The development and selection of nondestruc-
tive evaluation (NDE) techniques for reinforcement, voids, matrix-reinforcement
bonding and similar characteristics. The pur-
application to composite materials and struc-
tures presents several challenges and pose of the NDE is to detect these
inhomogeneities and others, including foreign
considerations that are quite distinct from the
material, fiber breakage, degradation due to
considerations given to the similar processes
moisture, ultraviolet (UV) or other reasons,
for metallic materials and structures. A princi-
cracks, abrasion, impact damage, fire or exces-
pal consideration is the nature of advanced
sive heat, etc. This chapter describes many
composites as typically layered, anisotropic
NDE methods that can be used to detect anom-
materials. The materials of interest include
alies in composite materials during
fiber-reinforced plastics such as 'fiberglass' and
manufacture and in-service.
carbon epoxy, as well as some of the more
Although composite materials have been in
exotic materials such as metal-matrix or
ceramic matrix composites. In some cases, nat- use on military aircraft for many years, the last
few years have seen a rapid escalation in both
urally occurring composites, such as wood,
the number and structural criticality of com-
with its mix of differing density summer and
posite applications entering service. The
winter growth rings and fibrous structure, may
be treated with approaches similar to those falling price of composite raw materials has
used for the man-made composites. The advan- vastly increased the number of applications in
non-aerospace industries, including automo-
tages offered by composites are focused on the
tive, sports, boating and construction. While
high strength, low weight properties of typical
the service record for composites has been
constituent materials. Yet, if the materials are to
excellent, they are subject to damage from
exhibit high strength, they must be manufac-
sources such as overload, hail, lightning, low
tured as the designer envisioned and they must
velocity impact, ballistic rounds and moisture
maintain their integrity in service. NDE pre-
intrusion. Low strain designs, coupled with
sents a technology to help assure the reliability
less critical applications, have made for rela-
of the materials.
tively large acceptable flaw sizes and lax
Inhomogeneities that may affect the perfor-
mance of a composite include the concentration inspection criteria. As composite structures
representing a more aggressive design criteria
and less tolerant materials (such as some of the
ceramic matrix materials) begin to accumulate
Handbook of Composites. Edited by S.T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 service hours, we will see a new level of
 Visual inspection 839

demands placed on nondestructive inspection as nonpigmented glass reinforced plastic

techniques to support the in-service mainte- (GRP). Stress whitening in GRP, which results
nance and repair requirements for composites. in a loss in optical transparency, may result
Several authorsl-9 have reviewed a variety from fiber-resin debonding or resin cracking.
of nondestructive methods applicable to the Another visual inspection approach, which
evaluation of composites. Several of these has been considered for the detection of impact
methods have, so far, been suitable only for damage, is to paint the composite surfaces with
laboratory applications. Some of these inspec- a paint containing micro-encapsulated dye.
tion methods are adaptations of inspection When crushed by the impact, the dye is
techniques that have been used for the evalua- released and reveals the location of the impact.
tion of metallic components for years 10 • New Dyes can be formulated either for visual obser-
methods are being developed to address the vation of a color change or using UV excited
special needs of composite materials. Many fluorescence in the dye. This method provides
inspection approaches have been automated good indications to locate suspicious areas for
for production applications; but in-service further inspection so that the extent of the
inspection of composites is frequently per- impact damage can be determined.
formed using conventional manual inspection Visual aids, such as magnifiers, borescopes,
equipment. Inspection times with manual television cameras, etc., can be used to
approaches can be long, but more significantly, enhance the detectability of surface-related
interpretation of the inspection results may damage. Small diameter, flexible borescopes
require special training or familiarization for permit the inspection of inaccessible areas
the inspectors. The long inspection times may without teardown of components.
restrict inspections to local areas based on crit- The most common visual inspection is
icality, loading, or suspicious areas detected by exemplified by the typical walk-around visual
visual inspections. inspection of aircraft. These inspections cer-
In recent years, there have been quite a few tainly apply to much of the composite
government sponsored research programs structures currently in use. Many of the more
directed at the development of new nonde- severe conditions associated with composites
structive inspection systems providing high are visually detectable. Punctures, surface ply
throughput for the increasing volume of com- delaminations, scratches, gouges and heat
posite structures in service and/or providing damage can frequently be detected by visual
some level of advanced signal analysis which inspection. Disbonds between a composite
will aid in the recognition and mapping of skin and some substructure may also be
flaws. As an example, a recent research pro- detectable in some cases as a blister in the skin,
gram resulted in the development of a an edge separation, or a distortion in the
computer-controlled ultrasonic scanning sys- geometry of the component. This inspection is
tem for field applicationsll. In this article the clearly valuable, but one cannot expect to
capabilities and limitations of some of these detect all of the forms of damage which may
typical inspection approaches are considered. be present in a composite structure. Further, if
damage is detected, it is important that other
tools are used to assess the extent of the dam-
38.2 VISUAL INSPECTION
age, since the subsurface damage may
Some translucent composites can be drastically exceed the surface detectable dam-
inspected by transmitted light. age. One common example is low velocity
Inhomogeneities such as voids, delamina- impact damage. Studies12,13 on this form of
tions or inclusions can be detected by this damage, which may result from hail, runway
straightforward approach in materials such debris, or ground handling abuse, have shown
840 Nondestructive evaluation methods for composites

that substantial subsurface damage can result disadvantages as manual tapping with the
with little or no surface detectable condition. additional disadvantage of increased cost.

38.3 TAP TEST 38.4 ULTRASONIC METHODS

The tap test method, using either a coin or a
38.4.1 ULTRASONIC THROUGH-
special tap hammer continues to be a common
TRANSMISSION TESTING
in-service inspection tool despite the avail-
ability of less subjective inspection tools. The Ultrasonic inspection makes use of high fre-
use of this method has persisted for several quency (above 20 kHz) mechanical vibrations.
reasons. The most obvious is that it does not Typical ultrasonic frequencies for inspection of
require sophisticated or expensive equipment. composite materials are in the megahertz
Another is that many of the composites in use range. Ultrasonic signal strength in a material
consist of thin laminates in low strain designs. is reduced both by attenuation and by reflec-
This combination yields critical flaw geome- tions from interfaces. Large reflection signals
tries which are fairly large and close to the are obtained when ultrasound is directed from
surface, two conditions necessary to the suc- one medium to another with the two media
cessful use of the tap method. Under these having very different acoustic impedances.
conditions, the tap method can be a useful For example, more that 99.9% of an ultrasonic
tool for the detection of problems in relatively wave is reflected from a metal-air interface as
large areas of laminate, particularly where the might be encountered with a crack in a metal.
substructure of the tested skin is relatively The ultrasonic signal is typically introduced in
consistent. The tap test method is sensitive a pulsed mode. The inspection geometry can
only to laminar type flaws, such as delamina- be through-transmission, with the receiver 'lis-
tions or unbonds, and relies on the different tening' for the ultrasonic pulse on the opposite
acoustic resonance of the loose upper layer side of the component from the transmitter, or
compared to the surrounding material. The a one-sided pulse-echo test, in which a single
tap method suffers from subjective interpreta- transducer first transmits a pulse and then 'lis-
tion (particularly with complex geometries), tens' for the reflections of that pulse from the
variable application, declining sensitivity interfaces of the inspection object.
with flaw depth and an inability to calibrate Ultrasonic through-transmission test sys-
effectively for either flaw size or depth. The tems measure the signal strength of a pulse of
more recent applications of composites in ultrasonic energy transmitted through the
thicker laminates and more highly loaded structure or material under test. Locations
designs make this approach inadequate in where there may be a delamination or a for-
many cases. eign material, for example, will show a
Several attempts have been made to instru- reduced exit ultrasonic intensity. This test
ment the tap test by providing a machine-type approach is probably the most commonly
tapper and instrumentation to interpret the used production inspection method for com-
signals. The mechanical tappers offer the posite structures. It is relatively easily
advantage of improved repeatability in terms automated, provides approximately constant
of tap impact and location. The instrumenta- sensitivity to flaws, regardless of their depth
tion developed centered on spectral analysis within the structure and is fairly easy to inter-
of the detected audio signals. These instru- pret. Sophisticated computer controlled
mented tappers have not made a significant inspection systems, such as the McDonnell
impact on the composite NDE instrument Douglas Automated Ultrasonic Scanning
market. They suffer from many of the same System (AUSS)14, shown in Fig. 38.1, are
 Ultrasonic methods 841

Fig. 38.1 McDonnell Douglas AUSS ultrasonic system.

common in aircraft production facilities. These mated systems is usually not available for field
systems scan the composite component while inspection. This is, however, changing. Many
collecting and recording the ultrasonic pulse of the military depot inspection facilities now
transmission amplitude. The method is sensi-
tive to most flaws which are planar in nature
and lie roughly parallel to the surface. Since
this describes most of the common flaw types
for layered composite structures, the method
is very appropriate. Where automated testing
systems are used, the test can be performed
very efficiently and can yield a recording, or
map of the inspection results, called a C-scan.
An example of a through-transmission C-scan
showing a delamination caused by a foreign
material inclusion is shown in Fig. 38.2.
Generally, the method requires access to
both sides of the part and alignment of ultra-
sonic search units on opposite sides of the
part. These factors drastically restrict the use-
fulness of this approach for in-service
inspection, since in many cases, access to both
sides is not available or not practical. Further, Fig. 38.2 Computerized through-transmission
the speed and recording advantages of auto- C-scan showing foreign material.
842 Nondestructive evaluation methods for composites

have automated ultrasonic inspection systems

for composite inspections. Further, inspec~on
systems such as the Automated Real-time
Inspection System (ARlS)15 provide a semiau-
tomated through-transmission inspection and
data recording capability for some on-aircraft
inspections. A 9 kg (20 Ib) yoke provides the
mechanism to place an ultrasonic transducer
on either side of the field inspection object.
One-sided pulse-echo testing can also be
accomplished.

38.4.2 ULTRASONIC PULSE-ECHO TESTING

The ultrasonic pulse-echo inspection approach Fig. 38.3 McDonnell Douglas LACIS portable ultra-
sonic scanner.
typically uses a single search unit as both the
transmitter and receiver. The approach
C-scan shown in Fig. 38.4 shows the multiple
requires access to only one side of a material or
delaminations associated with impact dam-
structure to be tested. Flaws are detected by
age. In this scan, the delaminations show as
monitoring the time of arrival and/ or the sig-
gray to black areas in which the darkness of
nal strength of returning echoes.
the indication reveals its depth. The darker
Delaminations cause the returning echo to
indications are closer to the surface. The foot-
arrive prior to its expected travel time from the
ball shaped delamination of the far surface ply
round trip from the front surface to far surface
is seen as a gray shape only slightly darker
and back to front surface. This method is com-
than the normal background thickness.
monly used for in-situ inspection since the
The pulse-echo method is also advanta-
single transducer approach usually makes it geous in that it offers increased sensitivity to
simpler to apply in a manual test. On the other
foreign material inclusions associated with the
hand, the more complex signal patterns and
manufacturing process for laminated compos-
more critical sound entry angle, make this test
ites. Many expendable paper and plastic
more difficult to automate than the through-
materials are used in handling and transporting
transmISSIOn test. Even so, computer
automated pulse-echo inspection systems
have gained popularity in recent years. .
An example is the Large Area ComposIte
Inspection System (LACIS). This semiauto-
mated pulse-echo inspection system is useful
for in-service inspection of composites. The
LACISll (Fig. 38.3) uses a reciprocating arm
holding two to four transducers to perform
pulse-echo inspection of large com~osite com-
ponents quickly. Inspection speeds m excess of
9.3 m 2 /h (100 ff/h) have been reported with
this hand-held scanning device.
An advantage to the pulse-echo systems is
that flaws at multiple depths can be distin- Fig. 38.4 pulse--echo time-of-flight C-scan showing
guished from one another. The pulse-echo impact damage.
 Ultrasonic methods 843

uncured composite materials. These materials surface loading on the ultrasonic probe18 • This
have at times found their way into a compos- can be seen as a phase, amplitude, or resonant
ite laminate and are cured in place. Some of frequency shift in the ultrasonic element. This
the plastic materials are particularly problem- method is particularly useful in complex
atic in that they may bond into the laminate bonded structures where access limitations
and go undetected by the through-transmis- restrict the use of ultrasonic through-transmis-
sion inspection. Fortunately, however, most of sion testing and the complex internal
these materials offer a reflected signal strength reflections make pulse-echo signals difficult
sufficient to be detected by pulse-echo ultra- or impossible to interpret.
sonics.
38.4.5 ULTRASONIC CORRELATION
38.4.3 ULTRASONIC POLAR BACKSCATIER
Ultrasonic correlation offers a novel approach
This inspection approach makes use of slightly to the ultrasonic inspection of highly attenua-
angled ultrasonic beams to detect flaws in tive materials19 • It has proven effective in the
composite materials. The angle orientation fre- evaluation of some of these materials where
quently allows one to associate flaws with a conventional pulsed ultrasonic systems have
particular ply orientation. The method is par- experienced difficulties in penetrating the
ticularly useful for characterizing matrix material. The correlation approach achieves
cracking within the composite plies. It is also this increased sensitivity using a continuous
sensitive to linear voids or porosity. It has been wave, cross-correlation technique which
successfully used16 to characterize the various enhances the sensitivity of the test but sacrifices
levels of damage in impacted laminates. The inspection speed. A block diagram of a continu-
approach has also been used to measure direc- ous wave correlator is shown in Fig. 38.5.
tional velocity variations and elastic properties The principal advantage of the correlator is
in composites 17• From these measurements, the that it provides substantial enhancements in
anisotropic elastic properties can be calcu- the signal-to-noise (SIN) ratio of the ultra-
lated. At the current time, this inspection sonic signals. This SIN enhancement is
method remains primarily a laboratory tool. achieved through a modification of the time
The angulating mechanism would be some- scale reqUired to produce an A-scan trace.
what difficult to manipulate and control in the Since the correlator uses continuous genera-
field and the inspection times can get long, tion and accumulation of the ultrasonic
since multiple scans at a variety of orientations signals, the maximum possible efficiency of
are required. Further, the composite designs data accumulation can occur. The SIN
currently in use have not required the detailed enhancement allows the correlator to produce
level of inspection provided by this method. usable A-scan traces where conventional
pulsed ultrasonic systems are unable to pro-
duce a recognizable pattern.
38.4.4 ULTRASONIC RESONANCE
For example, the correlator has been used to
This one-sided ultrasonic inspection method characterize polytetrafluoroethylene (PTFE,
detects laminar discontinuities within com- Teflon®), a material very difficult to penetrate
posites or bonded structures by setting up a with conventional ultrasonic techniques. The
continuous ultrasonic wave within the mater- results have been obtained by measuring
ial and sensing the mechanical stiffness, or ultrasonic velocity as the temperature of the
impedance of the material. A delamination or PTFE sample changed. Figure 38.6 shows a
disbond reduces the surface normal stiffness velocity versus. temperature profile with an
of the material and this, in turn, reduces the excellent indication of a change in the slope of
844 Nondestructive evaluation methods for composites

Transmit
Transducer
Noise RF
__- - . t Generator I---~ Power
No.1 Amplifier

Clock

Blanker Noise
(Delay) Generator ~--i-I~
No.2
. Integrator:
..............................

Strip
Chart

IBM PC

Fig. 38.5 Block diagram of ultrasonic correlator.

the curve at the phase transition temperature 38.5 X-RADIOLOGY

of 30°e. The correlator has been shown effec- X-ray imaging relies on the differential absorp-
tive for the inspection of thick and highly
tion or scattering of the X-ray photons as they
attenuative composites. Effective signals have
pass through a material. Flaws which either
also been obtained through over 30 em
allow more X-ray photons to pass or which
(11.8 in) of cured wood and through
absorb or scatter the photons can be imaged if
wood/plastic laminates, such as countertops.
the effect is significant. X-ray inspection sys-
tems are usually sensitive to changes which
result in an apparent change of at least 1-2% of
1340
the material thickness or density. Composite
1320 flaws tend to lie between plies and present a
1300 very small apparent thickness change, particu-
VelocIty 1280
larly for thick laminates. However, the low
(mls) density of most composite materials permits
1260 the use of low energy X-rays which help to
1240 enhance the sensitivity20. X-rays can be used to
detect porosity and matrix cracks as well as
1~0~--~-----+-----r----4----
20 25 30 35 40
some foreign materials. More highly loaded
Temperature (DC) applications and/or the use of more brittle
matrix materials, both of which seem to be on
Fig. 38.6 Ultrasonic velocity transition in PTFE the horizon, will make the detection of some of
measured with correlator. these conditions much more critical than they
 X-radiology 845

are today. While carbon fibers are not gener- typically possess the resolution of X-ray films,
ally imaged by X-rays, boron and silicon the use of very small X-ray sources and geo-
carbide fibers are typically deposited on a metric magnification techniques have
tungsten filament and this filament can be permitted the recovery of much of this sensi-
imaged in X-radiographs. This allows one to tivity. Figure 38.7 shows a magnified image of
detect fiber fractures and determine fiber ori- a honeycomb core area. This 'bee's eye view'
entations and placements, so long as the of the honeycomb core reveals the onset of
number of plies does not get too large. damage in the cell wall. In practice, one must
X-radiography is particularly useful for the find a suitable compromise between spatial
detection of honeycomb core defects in resolution and area of coverage. The cost
bonded sandwich assemblies. The low density advantages of this approach have made it very
and thin composite skins usually provide min- attractive. Many military aircraft maintenance
imal interference for the X-rays to image the facilities either have or are acquiring real-time
honeycomb core materials. Core defects such X-ray inspection equipment.
as blown core, crushed core, condensed core,
fatigued, corroded or cut core and foaming
38.5.1 X-RAY BACKSCATTER IMAGING
adhesive voids can be detected by radi-
ographic methods. It is also frequently Conventional radiographic inspection tech-
possible to detect water intrusion into the hon- niques rely on the attenuation of a beam of
eycomb core using X-ray methods. penetrating radiation to form an image of a
Significant progress has been made in part. At low X-ray photon energies typically
recent years in the use of non-film electronic used with composites, a large portion of that
imaging systems for X-rays. These systems attenuation is due to Compton scattering. This
allow the real time viewing of X-ray images suggests the use of scattered radiation to ana-
and can drastically reduce the demand for X- lyze the inspection object. One approach
ray film. While video imaging systems do not utilizing a novel, X-ray backscatter camera

Fig. 38.7 Microfocus X-radiograph of honeycomb cell showing onset of damage.

846 Nondestructive evaluation methods for composites

insulator and a rubber liner, as shown in Fig.

Lamina
38.9(a). Figure 38.9(b) is the X-ray backscatter
inspection result for a single location. The
illustration shows the relative X-ray backscat-
ter intensity versus thickness of the inspection
sample; depth through the sample can be cor-
Collimator related with detector number. Region A
represents the near-surface graphite-epoxy
X-ray
region. The sharp dip in response at B repre-
tube sents the decreased X-ray backscatter intensity
typical of an air gap or delamination. The insu-
Slot
camera lator response is given in region C; note the
differing response slope for this lower density
Detector array
....====ii:iii:=aaI!~ material, as compared to the response for the
second graphite-epoxy layer shown in region
D. The response from the rubber deeper in the
Fig. 38.8 X-ray backscatter system geometry.
assembly is shown in region E.
The X-ray backscatter signal contains quan-
employing a slot instead of the more common titative information about variations in
pin-hole approach has been developed21 • This density as caused by changes in material or
technique is illustrated in Fig. 38.8. The delaminations and the location of such varia-
backscattered X-rays are detected by an array tions within the depth of the material. These
of scintillation detectors. The backscatter characteristics, coupled with the one-sided
intensity information is obtained as a function inspection feature, represent advantages for
of detector position; this in tum, relates to dif- the inspection of composite assemblies.
ferent depths in the inspection object. One
interrogates the entire object thickness at one
38.5.2 COMPUTED TOMOGRAPHY
time (composite structures as thick as 7 cm
(2.7 in) have been inspected). The inspection Tomographic inspection systems, developed
results provide indications of material density originally for medical diagnostics, have been
changes by a change in the slope of the inten- further developed for applications in the
sity Iposition graph and of voids or foreign industrial inspection arena22- 24 • A tomographic
material by an abrupt change in backscatter image looks like a slice taken across the
intensity. The inspection method is particu- inspection object. A collimated single pencil
larly useful for the inspection of laminated beam or a flat fan beam of X-rays is directed at
structures such as pressure vessels and rocket the sample. The sample is shifted and rotated
motor cases. In some of these designs, ultra- relative to the X-ray beam while the X-ray
sonic inspection approaches are ineffective or beam intensity is measured by individual
impractical and the backscatter X-ray imaging detectors (typically arrays of detectors) at each
approach offers a potential solution. Tight position and rotation. The measurements per-
delaminations, with gaps less than 50 pm in mit a computer reconstruction of a density
width, can be detected. map of the inspection object. One advantage of
An example of an X-ray backscatter inspec- tomographic inspection is the extremely good
tion is shown in Fig. 38.9. The sample was a contrast sensitivity; variations of less than
graphite-epoxy composite structure as used in 0.1 % are detectable. The major advantage of
a rocket motor case. The assembly included computed tomography is that the resultant
several layers of graphite epoxy composite, an image shows all the variations across the
 X-radiology 847

BACKSCATTER
X-RATS

GRAPHITE
EPOXT

(a)

A~ep.y

10Ip

~ C~

I
D~ep.y

E~

I 2 3 4 5 I 7 I • 10 11 12 13 14 15 11 17 " " 20
(b)
DETECTOR NUM8ER

Fig. 38.9 X-ray backscatter examination results (a) composite sample configuration and (b) system
response.

image slice in the true three-dimensional per- For composites, tomography offers similar
spective. A typical radiograph is a advantages. The tomographic image shown in
two-dimensional projection of a three-dimen- Fig. 38.10 shows variations in density across
sional object. The tomographic result retains the width of the composite material.
the three-dimensional nature of the image. A
computed tomography image of a tree, for
38.5.3 NEUTRON RADIOGRAPHY
example, would appear similar to the visual
image if one sawed through the tree at one X-rays are attenuated as a function of the den-
location. The tomographic image would show sity of the material through which they pass. If
the differing density of summer and winter X-ray attenuation is plotted as a function of
growth and the presence of knots, voids, etc. increasing atomic weight of the attenuating
848 Nondestructive evaluation methods for composites

radiography. Recent advances in moderating

and collimating accelerator fast neutron beams
have led to significant improvements in
portable system thermal neutron beams26 • A
typical thermal neutron image of a composite
structure, as obtained with the transportable
inspection system shown earlier, is shown in
Fig. 38.12.

Fig. 38.10 Computed tomogram of graphite silica

composite tube.

material, one finds a monotonically increasing

function. The attenuation of thermal neu-
trons22, plotted similarly, shows a generally
random pattern in which there is high attenu-
ation for several light materials, hydrogen,
boron and lithium in particular and relatively
low attenuation for most metals. For bonded Fig. 38.11 Photograph of transportable thermal
structures, the high sensitivity of neutrons to neutron radiographic system.
hydrogen means that thermal neutron radiog-
raphy can display images of adhesives, water,
or corrosion in a metal bonded assembly more
readily than X-ray techniques. For composites,
neutron techniques offer promise for the
detection of variations in the organic matrix
materials and moisture take-up.
Advances in transportable neutron sources
have been made including source systems for
depot and field inspections. A photograph of
one system25 is shown in Fig. 38.11. The neu-
tron source is a Cockroft-Walton accelerator
that produces neutrons by accelerating
deuterons on a tritium target, a D-T reaction
that produces neutrons of about 14 Mev in
energy. These neutrons are slowed, or moder-
ated, in a low atomic number assembly and Fig. 38.12 Thermal neutron radiograph of a bonded
collimated into a beam for thermal neutron fiberglass-epoxy structure.
 Eddy current testing 849

A major neutron radiographic facility for One area where acoustic emission testing has
the inspection of military aircraft has been found acceptance is in the detection of mois-
installed at McClellan Air Force Base. A ture and corrosion in honeycomb assemblies.
maneuverable, crane-based system permits If an aluminum honeycomb assembly contain-
near real-time neutron inspection of wing and ing water is locally heated in the vicinity of the
empennage assemblies on the aircraft. Film water, the increased vapor pressure will force
techniques are used to inspect lower doors. the water to move through the leakage path
These on-aircraft inspection systems employ along which it entered the honeycomb cell.
the radioisotope 252Cf as the neutron source. The motion of the water through the leakage
The system has emphasized the inspection of path can be detected using acoustic emission
metal bonded assemblies to detect water intru- equipment.
sion and corrosion damage. A nuclear reactor
facility is available at this facility to inspect
38.7 ACOUSTO-ULTRASONICS
components removed from the aircraft.
Consideration is being given to replacement of The acousto-ultrasonic, or stress wave factor,
the californium source with an accelerator test method uses an ultrasonic transducer to
source of neutrons. inject a simulated acoustic emission pulse into
the material under test. The wave form of the
pulse is then monitored a fixed distance away
38.6 ACOUSTIC EMISSION
with an acoustic emission transducer. Damage
Acoustic emission testing involves the detec- to the material will affect the manner in which
tion of elastic energy that is spontaneously the wave is transmitted through the materiaFB.
released by materials when they undergo The stress wave factor will be affected most by
deformation27• For example, when a piece of discontinuities which impede the motion of
wood is stressed, audible cracking noises can the stress wave through the material. The
be detected just before the wood breaks. The method can be applied in a through transmis-
same phenomena occurs in other materials, sion mode, but is generally used with both
including composites. The detected signals are transducers on the same side of the part; one
frequently in the ultrasonic rather than audi- monitors the transmission of signals parallel to
ble region. the surface. In this mode the method is most
Acoustic emission testing monitors the sensitive to matrix cracking and laminate
material under test for spontaneous noise gen- porosity. It is also sensitive to fiber breakage
erated under load. It has been used extensively and, particularly in the through transmission
in the proof testing of fiberglass pressure ves- mode, delaminations. The method has suf-
sels and beams. It has also been used to fered from problems in obtaining consistent
monitor and characterize damage growth results. Some of these problems have been
mechanisms in composites under cyclic load- solved or reduced and the method shows
ing. Acoustic emission testing is capable of potential for evaluating the severity of matrix
detecting and characterizing matrix cracking, degradation in composites. One approach has
delamination and fiber breakage27• It is used been to use completely non-contact generation
extensively in the testing of composite pressure and detection of the signals29 •
vessels and in the evaluation of some struc-
tures, such as 'cherry picker' booms; however,
38.8 EDDY CURRENT TESTING
the method has not found particular applica-
tion in the aircraft inspection arena for these Eddy current methods rely on the principles of
flaws. This may be partly due to the need to magnetic induction to interrogate a material
have the component under load during testing. under tesf30. Current loops, or eddy currents,
850 Nondestructive evaluation methods for composites

are induced in a conducting material by a vary- a = Stefan-Boltzmann

constant = 5.7 x 1Q-8
ing magnetic field. While most resin matrix (W/m2K4)
materials are very poor electrical conductors, T = absolute temperature, degrees Kelvin
some of the carbon fibers are relatively good
conductors. In order to form current loops, The emittance in a particular differential
multiple fibers must make electrical contact wavelength band is governed by the Planck
with one another at various places along the distribution criterion. Important criteria for
length. Fortunately, the use of carbon fiber tows practical infrared imaging are the spectrum of
of 3000 to 10 000 fibers or more clearly pro- the emitted energy and the wavelength of
motes this. Eddy currents can then be used to maximum emittance. These are given by the
monitor fiber orientation since the current single temperature evaluation of Planck's Law
loops are not circular, but greatly elongated and Wien's Displacement Law, respectively.
along the fiber direction. Eddy currents can also Figure 38.13 shows a family of such energy
be used to detect fiber breakage. Eddy current distribution curves. For the practical case of a
techniques have been demonstrated to provide body at a temperature of 25°C (77°F, approxi-
effective results for many of the damage mech- mately room temperature), the wavelength
anisms, including impact damage and fatigue range of peak emission is in the range
damage, in carbon fiber composites31 • Even so, 9-10 pm. This turns out to be a very useful
eddy current techniques have not received a wavelength range. Various components of our
great deal of acceptance to date for production atmosphere, notably water, absorb a great deal
or in service inspection of composites. of the emitted infrared energy. However, there
are two fairly transparent windows through
the atmosphere: one between 3-5 pm and the
38.9 INFRARED THERMAL TESTING other at about 8-14 J.llll. Specific infrared
Infrared thermography has been gammg detector elements have been developed which
rapidly increasing acceptance as a nondestruc- respond well to each of these wavelength
tive evaluation tool for composite (as well as bands. As seen in Fig. 38.13, the intensity of
many other) structures. All bodies above the infrared emittance within such a band can be
temperature of absolute zero emit electromag- used to indicate the temperature of the object
netic radiation by virtue of the motion of the surface.
constituent atoms. The spectrum and intensity
of the radiation depend on the temperature 0.0005 r--------------~

and nature of the surface. When a surface is

heated, there is an increase in energy of the 0.0004 t-------IL------.:~___..j
atomic particles leading to a corresponding
increase in temperature and emitted ~
I! 0.0003 +------Jr.-.-----:;,----....,,------.:~-____I
~

I ~0002r_---+-~/~-·-··-··-··-~~~~
energy32.33. The wavelength-independent rate ///

of emission of radiant energy per unit area is

governed by the Stefan-Boltzmann law:
0.0001 +----+....;-,;.'-------------1
W = eaT'
where:
2 4 10 12 14 18 18
W = rate of emission, radiant energy per unit W......ngth(l1m)

area
e = emissivity (ratio of emittance of the sur- Fig. 38.13 Spectral radiant emittance distribution at
face relative to a black body) three surface temperatures.
 Laser shearography/holography 851

Typical thermographic applications involve

the introduction of a controlled thermal load
on the object of interest. Variations in the ther-
modynamic properties of the object then
produce surface temperature patterns which
can be detected with an infrared imaging sys-
tem. Composite materials typically offer an
excellent combination of thermal properties
for the useful application of infrared thermog-
raphy.
Thermographic methods have been used to
display object temperature differences for a
variety of inspection applications. Many of the
inspection systems provide a television-type Fig. 38.14 Thermogram of impact damage in
image display. Typical sensitivity of the instru- graphite epoxy composite.
ments under ideal conditions is O.l°C (0.2°P).
Thermal patterns can be produced by heating
the inspection surface and monitoring the sur- developed for composite materials and struc-
face of the part for relatively hot spots caused tures. This increased acceptance has been
by the blockage of the thermal flow away from assisted by improvements in the infrared
the surface by a laminar flaw. This method is imaging technology, development of
particularly sensitive to flaws near the inspec- improved heat applications techniques and
tion surface; the sensitivity reduces rapidly significant efforts to model and understand
with flaw depth34 • The thermal images can also the heat flow characteristics in anisotropic
be produced by heating the far side of the media. Effective applications have been identi-
object and monitoring the surface for cold fied for a wide range of materials and
spots where the heat flow toward the inspec- applications. Thermography has been used for
tion surface in impeded by the flaw. This aerospace laminates of graphite epoxy as well
'through-transmission' approach somewhat as for marine applications of fiberglass up to
improves the depth sensitivity, but the method around 8 cm (3 in) thick. In many instances for
is still more sensitive to near side flaws. materials which are most difficult for more tra-
Impact damage often results in matrix dam- ditional nondestructive tests, such as
age near the surface of a composite material. ultrasonics, infrared thermography has pro-
Figure 38.14 shows an infrared image of an duced rapid, effective results. An example is
impact damaged graphite epoxy composite foam core fiberglass boat hull structures.
sample. The object was heated and imaged on While this structure poses substantial difficul-
the same side. The diamond shaped indication ties for ultrasonic techniques, infrared
is typical of the delamination pattern pro- thermography has produced rapid, effective
duced in a cross-plied laminate. Infrared results35 •
thermography has been demonstrated to be an
effective tool for evaluating many glass fiber
38.10 LASER SHEAROGRAPHY I
composite structures used in marine applica-
HOLOGRAPHY
tions35,36.
Infrared thermography has been gaining A hologram is an interference pattern that can
increased acceptance in recent years as a viable be used to reconstruct. the optical wavefront
nondestructive testing tool. A significant num- that originally emanated (due to reflection or
ber of effective applications has been transmission) from an object. The hologram is
852 Nondestructive evaluation methods for composites

formed by the superposition of two wave- benches and elaborate vibration isolation sys-
fronts, an object beam and a reference beam, tems, however, the vibration problems of
on a suitable recording material, such as pho- working in a production environment (much
tographic film. When properly illuminated by less an in-service environment) were not satis-
the reference beam, an observer looking factorily solved. The holographic inspection
through the developed hologram sees a virtual approach is sensitive to very small amounts of
image of the original object. When the real relative motion in the part surface. Typically a
object undergoes a small displacement over part is tested by taking an image of the part at
part of the surface, due to stressing by thermal rest, then stressing the part surface, either with
or mechanical means, a variation in the rela- a small amount of heat or a mild vacuum. This
tive phase of the wavefronts will be produced light stressing causes the area of the surface
and a fringe pattern can be observed. Laser over the flaw to deform slightly more than the
illumination provides the spatial and temporal surrounding material. Displacements as small
coherence in the light beam to permit the as one-quarter wavelength of the laser light
observation of the fringe pattern, a technique being used cause a fringe to appear in the
called holographic interferometry37. An exam- interference pattern. Unfortunately, unwanted
ple of a holographic interferometry inspection test object motion of an equally small ampli-
of a composite tube to reveal impact damage is tude will also cause fringes. The development
shown in Fig. 38.15. of phase locked loop holographic systems has
Laser interferometric holography tech- gone a long way to eliminate the problems
niques were evaluated for composites testing associated with low frequency vibration.
in the 1960s. In fact, several composites and Holographic inspection systems which use
bonded assembly testing systems were con- video imaging systems and 'develop-in-place'
structed. In spite of the use of massive optical reference holograms are currently being used.

Fig. 38.15 Holographic interferogram of impact damage in a composite tube.

 Conclusions 853

Another development in the area of video

holographic systems, electronic shearography,
has recently provided even greater immunity
to test object vibration and motion problems,
making possible the rapid scanning of large
areas of composites on aircraft. Electronic
shearography uses no film and compares a
live video image with a stored video image to
produce interference fringes.
Conventional interferometric holography
interferes two holograms of a component to
produce the interference pattern. Each holo-
gram is produced by recording the speckle
interference between a coherent reference
beam and the coherent object beam which illu-
minates the surface of the test part. Fig. 38.16 Electronic shearography image of
Deformations between the two exposures pro- graphite-epoxy / foam core panel with edge delam-
ination.
duce the interference fringes. Phase-locked
holography uses the same basic technique, but
38.11 MICROWAVE TESTING
uses the diffuse reflection of an unexpanded
coherent beam shone on a small portion of the The use of microwave energy to interrogate
test object as the reference beam. This and characterize composite materials has been
approach minimizes the influence of full sur- investigated for some years. Microwaves are
face motions, typical of environmental very sensitive to small changes in the dielectric
vibration and is sensitive to differential sur- properties of low conductivity composites,
face displacements typical of defect such as glass and aramid fiber composites.
indications. Microwave techniques have also been investi-
Shearographf8, on the other hand, uses no gated for applications to composites with
separate reference beam. Rather, the returning higher conductivity fibers such as graphite
object beam is doubly imaged, with one of the fibers. Successful results have been reported
images slightly shifted or 'sheared' relative to for the measurement of fiber content and ori-
the unshifted image. Thus, the interference entation40, material thickness 41 and porosity
pattern does not reveal bulk surface motion, content42 • To date microwave techniques
but only the degree of differential motion of remain largely laboratory tools and have not
the surface along the direction of the shearing. gained great acceptance as production evalua-
This makes shearography particularly well tion techniques. However, these techniques
suited to many production and depot environ- offer excellent sensitivity to conditions, such
ments because of its relative immunity to as matrix porosity and cure state of the matrix,
vibration problems39• An example of a shearo- that are difficult to establish by more conven-
graphic inspection is shown in Fig. 38.16. In tional nondestructive testing techniques.
this case the inspection object is a sandwich Increased applications for these techniques
panel consisting of graphite-epoxy laminate can be anticipated in the future.
skins and a foam core. The indication arises
from a delamination created by a pull-out in
38.12 CONCLUSIONS
the skin-to-core bondline. The side-by-side
pattern of concentric rings is characteristic of A wide variety of nondestructive testing tech-
the shearography technique. niques are applicable to the evaluation of
854 Nondestructive evaluation methods for composites

various composite structures. The selection of too cumbersome to apply. Delaminations in

the most appropriate test technique for a par- the order of 1 em (0.4 in) in size can usually be
ticular application can be a formidable task. detected by shearography and even smaller,
The nondestructive inspection methods most near-surface discontinuities can be detected by
widely used in industry for composite inspec- infrared imaging. Both methods offer a televi-
tion, ultrasonics and radiography, have much sion image display which offers straight
to recommend them in that many different forward interpretation. Both methods also
types of discontinuities can be detected and offer good sensitivity in low density material
characterized. Shearographic and infrared systems which are frequently difficult to evalu-
inspection methods present the attractions of ate using the more traditional techniques.
noncontact, large area coverage and good sen- The overall summary of inspection methods
sitivity to discontinuities close to the inspected as related to composite inhomogeneities
surface. These techniques are finding a rapidly includes many inspection methods. Table 38.1
growing list of applications where the more provides a sorting approach that can be used
traditional tools encounter difficulties or are for the selection of inspection methods for par-

Table 38.1 Summary of applicability of NDE methods

Flaw TYPe
Porosity 1 1 2 2 1 1 1 2 2
Foreign material 3 2 1 2 2 2 2 2
2 2 2 3
Shallow delamination 2 1 1 2 1 1 11 3 2 1 1
Deep delamination 1 1 3 3 1 11 3 2 2
Matrix cracks 3 1 1 2 2 1 3 2
Fiber breaks 2 2 2 2 1 1
Impact damage 3 2 1 1 2 2 1 22132 1 1 1 1
Skin/ skin disbond 321 211 2 1 2 2 2 3 1 1
Skin/ core disbond 3 1 221 3 3 2 2 1 1
Crushed core 1 3 1 1 2 2 2
Condensed core 1 1
Blown core 1 1 1 1
Core node disbonds 1 1
Water intrusion 3 2 3 2 2 1 2 2
Corroded core 2 2 2 3 1 3
Fatigued core 2 2 1 2
Foam adhesive voids 3 2 1 2
Bondline adhesive voids 1 211 2 2 2 1 2 1
Key:
1. Good sensitivity and reliability. Good candidate for primary method.
2. Less reliability or limited applicability. May be good supplementary method.
3. Limited applicability. May provide some useful information.
 References 855

ticular composite discontinuities. While this 1985.

table cannot be considered definitive, it does 11. Large Area Composites Inspection System-I,
give general guidance to the applicability of USAF Contract F3361S-91-C-5664, Wright
Research and Development Center, Wright-
the various technologies to various material Patterson Air Force Base, Contract with
conditions. In any specific application, a large McDonnell Douglas Corporation, 1991.
number of considerations will influence the 12. Smith, B.T., Heyman, J.S., Moore, J.G., Cucura,
selection of nondestructive testing technolo- S.J. and Freeman, S.M., Correlation of the deply
gies. It is hoped that the discussions provided technique with the ultrasonic imaging of impact
in this chapter will offer some assistance in the damage in graphite/epoxy composites, Review
selection and application of nondestructive of Progress in Quantitative Nondestructive
Evaluation, Vol. 5B, New York: Plenum Press,
testing methods. However, final selection 1986,1238-1244.
should be based on the evaluation of a greater 13. Gros, X.E., Low energy impact detection on car-
number of parameters than can be considered bon fiber reinforced materials, Materials
here. Experience, training, experiment and Evaluation, 1995, 53, 373-381.
clear understanding of the inspection objec- 14. Jones, T.S., Inspection of composites using the
tives are required to develop effective automated ultrasonic scanning system (AUSS),
nondestructive testing applications. Materials Evaluation, 1985,43, 746-753.
15. Jacobs, B., Hamlin, D., Peterson, R. and Spinks,
R, Composites in-service inspection system
REFERENCES producibility, Air Force Report AFWAL-TR 88
4218, Wright Research and Development
1. Prakash, R, Non-destructive testing of compos- Center, Wright-Patterson Air Force Base, OH,
ites, Composites, 1980, 11(4), 217-224. 1988.
2. Scott, I.G. and Scala, CM., A review of non- 16. Blodgett, E.D., Miller, J.G. and Freeman, S.M.,
destructive testing of composite materials, NDT Correlation of ultrasonic polar backscatter with
International, 1982,15(4), 75-86. the deply technique for assessment of impact
3. Reynolds, W.N., Nondestructive testing (NDT) damage in composite laminates, Review of
of fiber-reinforced composite materials, SAMPE Progress in Quantitative Nondestructive
Quarterly, 1985, 16, 1-16. Evaluation, Vol. 5B, New York: Plenum Press,
4. Shuford, RJ., Hinton, Y.L., Murray, T.J. and 1986,1227-1238.
Brockelman, RH., Advanced NDE techniques 17. Bar-Cohen, Y., Mal, A.K. and Lih, S.S., NDE of
for composites, Technical Paper EM85-113, composite materials using ultrasonic oblique
Society for Manufacturing Engineers, 1985. insonification, Materials Evaluation, 1993, 51,
5. Bar-Cohen, Y., NDE of fiber-reinforced compos- 1285-1296.
ite materials - A review, Materials Evaluation, 18. Botsco, RJ., New methods for nondestructively
1986, 44, 446--454. evaluating airframes and jet engines, Air
6. Brahney, J,H., Inspection and repair of compos- Transport Association Forum, Long Beach, CA,
ite aerospace structures, Materials Evaluation, September 8-11,1980.
1986,44,1513, 153t 1539. 19. Kishoni, D., Rosen, M., Berger, H. and Cheng,
7. Jones, T.S. and Berger, H., Nondestructive eval- Y.T., Signal to noise enhancement by an ultra-
uation methods for composites, International sonic cross-correlation system, IEEE Ultrasonics
Encyclopedia of Composites, (S.M. Lee, ed.), 1991, Symp., Chicago, Oct., 1988.
Vol. 4, pp. 37-49. 20. Jones, T.S., Polansky, D. and Berger, H.,
8. Seidl, A.L., Inspection of composite structures - Radiation inspection methods for composites,
Part I, SAMPE Journal, 1994, 30, July/Aug NDT International, 1988, 21, 277-282.
38-44,1994. 21. Berger, H., Cheng, Y.T., Jones, T.S. and Polansky,
9. Seidl, A.L., Inspection of composite structures- D., An electronic imaging technique for one-
Part II, SAMPE Journal, 1995, Jan/Feb, 31, 42-48. sided X-ray inspection, Proc. 1989 ASME
10. Boyer, H.E. and Gall, T.L., (eds), Metals Pressure Vessel and Piping Conference, NDE, Vol.
Handbook, Part IV, Section 33, Nondestructive 5, (RD. Streit, ed), pp. 89-93, Amer. Soc. Mech.
Testing, Amer. Soc. Metals, Metals Park, OH, Eng., New York, July 1989.
856 Nondestructive evaluation methods for composites

22. Bryant, L.E. and McIntire, P. (eds.), 33. Cohen, J., Elements of thermography for nonde-
Radiography and Radiation Testing, structive testing, NBS Technical Note 1177,
Nondestructive Testing Handbook, Vol. 3, National Institute of Standards and Technology,
Radiography and Radiation Testing, Arner. Soc. Washington, D.c., (1983).
Nondestr. Testing, Columbus, OH, 1985. 34. Jones, T. and Berger, H., Thermographic detec-
23. ASTM E-1441, Standard Guide for Computed tion of impact damage in graphite-epoxy
Tomography (CT) Imaging, Amer. Soc. Testing composites, Materials Evaluation, 1992, 50,
Materials, Philadelphia, PA, 1993. 1446-1453.
24. ASTM E-1570, Standard Practice for Computed 35. Jones, T.S., Berger, H. and Weaver, E., Large area
Tomographic (CT) Examination, Amer. Soc. thermographic inspection of GRP composite
Testing Materials, Philadelphia, PA, 1993. marine vessel hulls, Thermosense XV: Intern.
25. Antal, J.J., Dance, W.E., Moravec, J.D. and Con! Thermal Sensing and Imaging Diagnostic
Carollo, S.F., Experience with an on-off mobile Applications, (Lee R Allen, ed.), SPIE Vol. 1933,
neutron radiography system, Proc. Second World pp. 197-206.
Con! Neutron Radiog., (eds. J.P. Barton, G. Farny, 36. Jones, T.S. and Lindgren, E.A., Thermographic
J.L. Person and H. Rottger), Dordrecht: DReidel inspection of marine composite structures,
Publishing Co. 1987,407-414. Thermosense XVI: Intern. Con! Thermal Sensing
26. Cluzeau, S. and Le Tourneur, P., Stationary and Imaging Diagnostic Applications, (John R
DIANE equipment: description and perfor- Snell, ed.), SPIE Vol. 2245, pp. 173-175, (1994).
mance of the thermal neutron source, Nuclear 37. Vest, C.M., Holographic Interferometry, New York:
Instruments and Methods in Physics Research B, John Wiley, 1979.
1994,89,428-431. 38. Hung, Y.Y., Shearography: a new optical
27. Miller, RK. and McIntire, P., (eds), method for strain measurement and nonde-
Nondestructive Testing Handbook, Vol. 5, Acoustic structive testing, Optical Engineering, 1982, 21(3),
Emission Testing, Amer. Soc. Nondestructive 391-395.
Testing, Columbus, OH, 1987. 39. Newman, J,w., Shearographic inspection of air-
28. Sundaresan, M.J., Henneke, E.G. and Brosey, craft structure, Materials Evaluation, 1991, 49(9),
W.o., Acousto-ultrasonic investigation of fila- 1106-1109.
ment-wound spherical pressure vessels, 40. Urabe, K. and Yomoda, S., Non-destructive test-
Materials Evaluation, 1991,49,601-606,612. ing method of fiber orientation and fiber
29. Huber, RD. and Green, RE., Noncontact content in FRP using microwave, Prog. Sci.
acousto-ultrasonics using laser generation and Engng of Composites, Fourth International
laser interferometric detection, Materials Conference on Composite Materials, Tokyo,
Evaluation, 1991, 49, 613-618. 1982.
30. McMaster, R.c., McIntire, P. and Meister, M.L., 41. Bakhtiari, S., Ganchev, S. and Zoughi, R,
(eds), Nondestructive Testing Handbook, Vol. 4, Microwave swept-frequency optimization for
Electromagnetic Testing, Arner. Soc. Nondestr. accurate thickness or dielectric property moni-
Testing, Columbus, OH, 1986. toring of conductor-backed composites,
31. Mahoon, A., Automated eddy current testing of Materials Evaluation, 1993,51(6),740-743,748.
composites, Proc. 20th Intern. SAMPE Tech. 42. Gray, S., Ganchev, S., Qaddoumi, N.,
Conf., Vol. 20, pp. 517-523, SAMPE, Covina, CA, Beauregard, G., Radford, D. and Zoughi, R,
1988. Porosity level estimation in polymer composites
32. Wolfe, W.L. and Zissis, G.J., The Infrared using microwaves, Materials Evaluation, 1995,
Handbook, Office of Naval Research, Dept. of the 53(3), 404-408.
Navy, Washington, D.C., 1985.
REPAIR ASPECTS OF COMPOSITE AND 39
ADHESIVELY BONDED AIRCRAFT STRUCTURES
Anton 1. Seidl

39.1 INTRODUCTION the composites industry, is an unfamiliar lan-

guage to the uninitiated. Inspectors are often
39.1.1 MANUFACTURABILITY AND at a loss when attempting to describe a condi-
MAINTAINABILITY OF COMPOSITE AIRCRAFT tion they perceive as a defect; the words
STRUCTURE simply do not exist in their standard lexicon. It
is intuitively clear to even the casual observer
To the manufacturer, weight reductions, struc-
that repairs using mechanically fastened con-
tural requirements, manufacturability and
ventional materials can be effected quickly,
production costs have long been obvious pri-
under almost any atmospheric conditions, and
orities. Only recently, however, and only as a
with minimal investments in tooling, raw
consequence of persistent user demands, have
materials, and training. In contrast, repairing
maintainability and repairability been added
even relatively minor damage on composite
to this list. From the operator's perspective,
structure requires an array of non-conven-
nevertheless, composite structures continue to
tional materials, highly skilled and
be a mixed blessing. Clearly, and despite state-
experienced technicians, special tooling and
ments being heard to the contrary, the
equipment, access to production drawings (to
industry would be loath to give up the many
locate and interpret the many hidden features
obvious advantages gained through the use of
characteristic of composite structures), a con-
composites and revert to all-metal airplanes.
trolled environment in terms of temperature
However, the maintenance problems associ-
and humidity, time-consuming preparatory
ated with composites cannot be
work, cold storage of shelf-life limited and
underestimated and may well be regarded as
occasionally hazardous materials, lengthy
the weak link in the new technology chain.
resin cure cycles, post-repair NDT, and legally
mandated record-keeping and follow-up
39.1.2 METAL REPAIRS COMPARED WITH activities.
COMPOSITE REPAIRS
Compared to the relative simplicity of conven- 39.1.3 COMPOSITE REPAIRS: AN AIRLINE
tional metallic structures, composites are PERSPECTIVE
replete with complexities that continue to baf-
The aim and purpose of this presentation is to
fle and confuse maintenance workers trained
highlight the principal aspects of composite
only in the traditional, i.e. metalworking,
structure repairs from an airline perspective.
skills. The glossary of terms alone, as used by
An attempt will be made to:
1. describe some of the more common defects
Handbook of Composites. Edited by S.T. Peters. Published and conditions encountered in service;
in 1998 by Chapman & Hall, London. ISBN 0412 54020 7
858 Repair aspects of composite and adhesively bonded aircraft structures

2. give a brief summary of common mainte- subsequent repairs more difficult to perform.
nance practices; When surface defects are detected, it is imper-
3. describe a limited number of typical repairs. ative to determine the extent of the damage
that may already have occurred and if possi-
ble, evacuate and decontaminate the panel. As
39.2 DAMAGE ASSESSMENT
a first line of defense against any further struc-
tural deterioration, foil tape should be applied
39.2.1 IMPACT DAMAGE - NON-METALLIC
at the earliest opportunity2.
STRUCTURE
Severe degradation is generally quite obvi-
ous, having resulted in visible disbonds and
Foreign object impact without skin
delaminations. If the damage is due to a recent
penetration
event, and the process of deterioration has
Prior to any repair action, it is important to only begun, the amount of ingested fluid may
determine the extent of the damage sustained still be relatively small and if the precise loca-
by the structure. One must always assume that tion of the contaminant can be determined,
the actual damage is more extensive than the complete evacuation and purging may be suc-
visible damage l . This is especially true for car- cessfully accomplished.
bon fiber-reinforced composites with
non-toughened 177°C (350 0 P) cured matrix
Limitations of moisture detectors3
resins. After a foreign object impact, there is
generally, but not invariably, some visual indi- Commercial moisture detectors are extremely
cation in the form of damaged paint. However, useful devices requiring no special training.
because of the elasticity of high modulus Where appropriate, they may be used to deter-
fibers, the laminate often 'springs back', leav- mine how far any ingested water has spread
ing residual subsurface damage in the form of into the core cells adjacent to the point of
broken fibers, ply separations and in the case impact. Moisture detectors, however, are effec-
of sandwich panels, crushed core and dis- tive only on non-metallic (typically glass or
bonded face sheets. Tap testing is generally aramid-reinforced structures); they cannot be
sufficient to delineate the extent of the damage used on panels containing carbon fibers, or in
and should be conducted before removing any zones reinforced with metals. Nor are mois-
damaged materials. Defects may also propa- ture detectors effective through surfaces
gate during the removal process, often as the coated with carbon-filled conductive paints,
result of stress relief. on panels having metallic coatings, metal-
coated fibers, or similar lightning protective
and EMI shielding features.
Skin penetrations: holes, cracks, tears,
gouges, cuts and abrasions
Efficacy of radiography
On the 'wetted' outer surface of the aircraft,
even minor penetrations of the face sheet must Radiography (X-ray) is presently the only
be regarded as serious because once an open- available practical technique for determining
ing exists, the part has been rendered moisture contamination in panels containing
permeable to atmospheric moisture and air- electrically conductive elements. Given the
craft system fluids. Ingested water can and limitations of moisture detectors and the lim-
will degrade the affected part, leading to pre- ited availability of X-ray equipment in the
mature failure. Fluids such as hydraulic oil, field, water detection by X-ray and subsequent
when allowed to enter, contaminate both lam- evacuation are generally carried out only dur-
inates and honeycomb core materials, making ing depot level maintenance opportunities.
 Damage assessment 859

Interim repair actions - 'speed tape' repairs Effects of skin penetration: corrosion, resin
plasticization and core dissolution
When a composite panel is found to have been
penetrated, it is important to prevent further Any impact damage resulting in skin penetra-
deterioration of the panel. When fluid detri- tion must be regarded as serious damage.
mental to adhesion (hydraulic oils, deicing However, unlike non-metallic core materials,
fluid, engine oils, etc.) is present, the affected which absorb and diffuse water, non-perfo-
area must be thoroughly decontaminated rated aluminum honeycomb cores tend to keep
before attempting a repair, or the contami- any ingested water concentrated about the area
nated material removed entirely. of the penetration. Left unattended, prolonged
When a permanent repair is to be deferred, exposure will cause the ingested water to
fractured material should be trimmed away migrate to other areas of the panel by gradual,
and the opening covered with foil tape before progressive diffusion through the adhesive
the aircraft is dispatched to a location where bondlines and, preferentially, through the core
the appropriate repair facilities exist. Foil tapes splice adhesives. As the bonding adhesives
must be applied with care to prevent their absorb moisture they become plasticized and
coming loose in flight. Loose foils have been their bond strength weakens.
known to create static noises that can interfere At the same time, unprotected areas of the
with radio communications. face sheet, doublers, substructural compo-
nents, cut edges and fastener holes, i.e. where
39.2.2 IMPACT DAMAGE ON METAL-SKINNED the anodic and primer protections have been
SANDWICH PANELS removed during the manufacturing process,
and machined edges of the honeycomb core,
Unlike laminated face sheets, which may are rendered vulnerable to corrosion attack.
show little evidence of an impact having Ingested water, if left unevacuated for long
taken place, thin metal face sheets (common periods, has been known to initiate chemical
on many honeycomb sandwich panels) reactions that lead to complete dissolution of
invariably become dented or gouged by for- the aluminum honeycomb core.
eign object impact. The resulting surface
irregularities are readily seen.
39.2.3 DAMAGED PROTECTIVE COATINGS
Minor damage - no skin penetration AND SEALANTS: LEAK PATHS
Shallow dents may be present that do not nec- Water ingestion and fluid contamination must
essarily result in disbanding of the skin, but be presumed to exist whenever the protective
there will always occur some crushing of the coatings or sealants of a panel have been dis-
core cells. A tap test will usually, but not turbed. The cause may be erosion of the
always, determine whether the skin is dis- protective finish, substrate corrosion, hail
bonded. Dents that have not resulted in skin damage, minor collisions, or similar foreign
disbonds are generally considered negligible object damage episodes. Leak paths, no matter
damage and may be filled with an appropriate how small, are detrimental to the long-term
compound to restore aerodynamic cleanness, structural integrity of the panel because they
provided the added weight does not affect the allow atmospheric moisture, aircraft system
balance of a critical control surface. Flight con- fluids, or a combination of contaminants, to
trol surfaces damaged by hailstones frequently enter the structure.
exhibit multiple dents that cannot be repaired Subsequent 'ground-air-ground' and
by dent fillers without creating an out-of-bal- 'freeze-thaw' cycling are capable of introduc-
ance condition. ing considerable quantities of water and other
860 Repair aspects of composite and adhesively bonded aircraft structures

fluids into the core of a panel. Sandwich pan- Moisture

els with thin face sheets of aramid/ epoxy are
While demonstrably corrosive to metals,
especially vulnerable to moisture contamina-
moisture is far more pernicious in its effect on
tion via small cracks in the resin gelcoat and at
composites because it plasticizes resins,
the resin-fiber interface. This phenomenon
degrades their mechanical properties and
has been explained as the result of the thermal
lowers their glass transition temperature4 •
expansion behavior of the aramid fiber, which
The latter effect becomes extremely critical
is slightly negative in the longitudinal direc-
when carrying out hot bonded and/or
tion and strongly positive in the transverse
prepreg repairs that require the heating of the
direction, leading to excessive strain build-up
structure.
within the weave itself.

Atmospheric electricity 22-23

39.3 ENVIRONMENTAL DAMAGE AND
Atmospheric electricity, of negligible con-
DEGRADATION
squence to metal structures having inherent
conductivities, can have a crippling effect on
39.3.1 GENERAL EFFECTS OF AGING 24
non-metallics, which compels the operator to
All polymeric materials are subject to degra- place a high priority on periodic testing and
dation over time. For this reason, the proper maintenance of anti-static and light-
importance of maintaining protective coat- ning protection schemes, i.e. ground paths,
ings and sealants cannot be emphasized too bonding fasteners, bus strips, jumper cables,
strongly. The 'normal' operating environment conductive enamel and/ or flame-spray coat-
of an aircraft exposes composite structures ings, as well as discharge ports.
not only to considerable static and dynamic
loads, but also to significant temperature gra- Chemical contamination
dients, extreme variations in humidity
conditions, and to a number of chemical Aircraft fluids and chemicals that are harmless
agents necessary in aircraft systems, the most on metals can effectively destroy a composite,
detrimental being hydraulic fluid, a powerful if allowed to penetrate its outer protective lay-
solvent. ers. Chemical paint strippers routinely used
on metal aircraft are occasionally - albeit inad-
vertently - applied to composite surfaces with
39.3.2 EXPOSURE OF COMPOSITES TO THE destructive consequences, even if the exposure
'NORMAL' FLlGll ENVIRONMENT is but of short duration. Presently, composites
can only be stripped by abrasive, non-chemi-
There is abundant evidence that the combined
cal methods5•
effects of stress, temperature, water and other
fluids expose bonded and especially fiber-
reinforced composite structures to a far wider Overheat conditions
range of hazards than their baseline metal Heat, except for annealing temperatures, is of
analogs. Following is a brief description of the minor concern with metals; by contrast, the
most common environmental hazards com- heat resistance of composites is effectively lim-
posite materials are exposed to. ited by the maximum use temperature of the
polymeric matrix. Heat generated by lightning
strikes has been known to vaporize matrix
resins and create large areas of delamination
and fiber fracturing on composite rudders,
 Damage removal techniques 861

ailerons, wing and stabilizer tips, nose domes, from the backside, with the additional benefit
and nacelle cowling. When exposed to hot of causing only minimal disruption to the
gases over long periods, polymeric resin aerodynamic surface. Figure 39.1 illustrates
binders, irrespective of chemistry, can become this principle.
completely destroyed through a process some- If the backside is inaccessible, the damage
times described as thermo-oxidation. This must be repaired from the aerodynamic skin
condition may be found on all types of com- side, inevitably enlarging the repair surface
posites, including those with inorganic and making the repair more difficult to per-
matrices, such as metal matrix composites. form. With only one side accessible, the
Preventive maintenance may consist of the question of how best to apply vacuum pressure
application of heat-resistant, ablative or intu- is always problematic and requires consider-
mescent coatings. Extensive redesign of the able operator skills. (Applying vacuum
detail may be necessary, using metals or, if a pressure for a bonded repair is an art form that
fiber composite is to be used, choosing a poly- must be learned as any other.) As an alternative
imide or similar high temperature resistant to field repairs, panels are often removed from
resin system. the affected structure and routed to a repair
facility specially equipped to effect the appro-
priate restorations. It should be noted that the
39.4 DAMAGE REMOVAL TECHNIQUES
damages affecting the aerodynamic skin sur-
face normally require 'flush' repairs to
39.4.1 PLANNING THE REPAIR: 'THINK
preserve the original contour, particularly in
BEFORE CUTTING'
zones of the aircraft defined as aerodynami-
After determining the full extent of the damage, cally critical. Except for small damages, the
the repair technician must consider a range of tooling and skill levels required to effect proper
pOSSible approaches, based on such considera- repairs do not exist at field stations.
tions as damage location, access to the damage,
required disassembly to create better access, r-AERODYNAMIC
available tooling and repair materials, as well " ~ SKIN DAMAGE

trrrr1111111
as the allotted out-of-service time. Because
most repairs are 'on-condition', i.e. the result of
damage events affecting the structure at unpre- IIII1
dictable locations in a multiplicity of manners,

'7r
allowing only limited pre-planning, the techni- REPAIR PLIES FILLER

cian's experience and intuitive problem-solving

abilities are of paramount importance.

39.4.2 AERODYNAMIC SKIN DAMAGE

[ij "
REMOVAL REPAIR PLIES ~
CORE PLUG
If the damage affects the outer, aerodynamic LUSHSIDE
or 'tool' side of a panel and the backside (the
'bag side') is accessible, it is best to remove
material from the backside, and as much of the
core material as necessary, to gain access to the
II]I=IIIIUIII
NON-FLUSH SIDE
damage. Using this method preserves as much
of the aerodynamically 'clean' surface as pos-
sible. Repair work is more easily accomplished Fig. 39.1 Aerodynamic skin side repair.
862 Repair aspects of composite and adhesively bonded aircraft structures

39.4.3 REMOVAL OF METAL FACE SHEETS Aluminum core removal

AND DOUBLERS
Severely damaged aluminum core (crushed,
Metal-faced sandwich panels are used on corroded, failed node bonds, etc.) should
wing spoilers, wing and stabilizer panels, always be replaced. Removal is generally
flaps, slats, engine cowling, landing gear strut accomplished by using non-metallic scrapers
doors, as well as in a multitude of other appli- or chisels mounted in a pneumatic rivet gun.
cations, including aircraft interiors. Damaged, Care must be taken to avoid damaging the
corroded, or disbonded face sheets are gener- intact face sheet on the far side. The remaining
ally peeled away after the application of adhesive fillets on the far side should be
carbon dioxide pellets ('dry ice'). The dry ice is abraded with a rotary sander, provided the
allowed to dwell on the surface until the ther- adhesive is still firmly attached, thlis provid-
mal shock has weakened the bond strength of ing a good base for bonding in the
the adhesive sufficiently to allow the skin to be replacement core plug. If the far side adhesive
removed. If done properly, the face sheets sep- is plasticized or unbonded, or corrosion is
arate, leaving the core cells relatively found between the adhesive layer and the
undamaged. metal skin, the adhesive must be removed for
closer inspection and possible reconditioning
of the bonding surface. Corroded skins and
39.4.4 REMOVAL OF COMPOSITE FACE SHEETS
doublers are routinely replaced.
AND DOUBLERS
The outer surface of most non-metallic sand-
Non-metallic core removal
wich panels consists of only a small number
of prepreg fabric and/or tape plies co-cured Non-metallic core materials are generally
onto non-metallic honeycomb core, although replaced if crushed or split at the nodes, or if
precured laminates, secondarily bonded to irreversibly contaminated by oil, hydraulic
the core, are also found. Removing damaged fluids, or other contaminants that would
or disbonded face sheet materials requires inhibit subsequent repair resin adhesion and
either a rotary sander (physical abrasion) or cure. Repair shops often attempt to deconta-
the use of a hot air gun combined with peel- minate core by flushing out the cells with
ing action. Heating the skin laminate has the solvent, a method not always successful and
effect of weakening the resin fillets suffi- a potential environmental hazard. If the cont-
ciently to allow the technician to peel the face aminant is water, dehydration of the core by
sheet materials with only minimal damage to evaporation, placing the part in an oven at a
the core. low temperature, is often possible, allowing
the material to be salvaged (see Section 39.5).
When core replacement becomes necessary,
39.4.5 REMOVAL OF HONEYCOMB CORE
the affected sections are generally cut out
MATERIALS
with knives or rotary cutters; the resin fillets
After the face sheet material has been remaining on the far side are then removed
removed, the condition of the honeycomb core with rotary sanders, to create a proper sur-
must be determined. Because of the high cost face for bonding in the replacement core
and limited availability of some core materials, plug.
repair shops attempt to salvage the original
material if at all possible.
 Decontamination 863

39.5 DECONTAMINATION Effects of contaminants on structural

integrity
39.5.1 EVACUATION AND
Dimensional swelling and plasticization of the
DECONTAMINATION OF POLYMER MATRIX
resin matrix generally result from exposure to
COMPOSITE STRUCTURES
high humidity at high temperatures, exposure
to many aircraft fluids, to chemical paint strip-
Vulnerability of polymers to fluids
pers, and to a variety of common solvents.
Organic matrix composites typically absorb Absorbed moisture lowers the glass transition
between one and two percent of their dry temperature of the laminate4 and may be con-
weight in moisture under normal service con- ducive to additional microcracking within the
ditions. There exists a certain risk when such matrix which, in tum, increases the potential
assemblies are subjected to the elevated tem- for additional moisture absorption. Micro-
peratures routinely applied during bonding cracks are considered irreversible, since they
and laminating repairs. During hot bond remain after the laminate has been completely
repairs, the absorbed moisture volatilizes. The dehydrated. Absorbed chemicals mayor may
effect on the repaired structure may manifest not affect the structural or mechanical proper-
itself in the form of porosities in the bondline or ties of the composite, but generally render the
in the laminate. In severe cases, such as when affected part unrepairable because they inhibit
water is present in the core cells, the pressure repair resin adhesion and cure.
resulting from the entrapped steam often
results in uncontrollable skin-to-core disbond-
39.5.2 GENERAL PRECAUTIONS
ing. For these reasons, it is always advisable to
pre-dry composite panels when moisture cont-
Water evacuation under vacuum pressure at
amination exists in detectable quantities, or
elevated temperatures
may be presumed to exist, given the general
condition of the part. One should keep in mind Removing water is mandatory in all cases, but
that fibers, with the exception of aramid, do not the process becomes especially critical if the
absorb moisture. Moisture absorption is a phe- repair requires the application of elevated cure
nomenon that affects primarily the resin matrix temperatures under vacuum pressure, which
and, secondarily, non-metallic core made from is typical or many in situ heating blanket type
aramid fibers. As a general rule, resin systems repairs. The operator must be aware that,
cured at 170°C (350°F) or above are more resis- under a vacuum bag, lowering the vapor pres-
tant to moisture pick-up than resin systems sure also lowers the boiling point of the water;
cured at lower temperatures, which includes all at the same time, increasing the temperature
room temperature-cured repair resins and cold- increases the steam pressure inside the sand-
bond adhesives6-8. wich (Fig. 39.2). The result is often a failed
repair: blown core and disbonded face sheets.
Effect of contaminants on weight and
balance Removing moisture barriers (coatings and
films)
Fluids absorbed by, or otherwise introduced
into a structure, induce weight gains and may Evacuation of composite laminates is best
cause out-of-balance conditions in flight con- accomplished by first removing any protec-
trol surfaces. Contamination detected should tive coatings and moisture barrier films that
always be evacuated, the leakage paths identi- may still be present and intact. Barrier mate-
fied, repaired and the structure resealed. rials may be various enamel finishes, sealer
864 Repair aspects of composite and adhesively bonded aircraft structures

459.02

333.89 •
• R-~
,;s (vf:o P Steam
177.94 .~;p~~ 89.6
«I~ (psI)
~~9:-'t-
106.73 BOILING POINT, .~~-I(() 52
~«)
WATERATNOR~
ATMOSPHERIC
P
58.63
PRESSURE ® 28
Vapor
29. 14.7
(In Hg) 20.70 I \ 10
, ......\ '~~
- - - - -t r )
13.98 7
9.20 I ~VACUATION WINDOW 5
3
5.88
3.64
•
.'r~
I I EVACUATION WINDOW
1.8
2.18 • I I 2
(Dopencore
evacuation
1.25 • I I ®Through face
0.70 • sheet evacuation
0.36 • @)Zoneof
0- 0.18 ......--1f--__+--+-r-.....-+-~I____+-_+-......---I increasing risk
-~ - - 0 20 40 60 80 100 120 140 160 180 200
of __ 32 68 104 140 I 176 212 248 284 320 356 392
I
160

Fig. 39.2 Pressure and evacuation guidelines for honeycomb core repair.

coats, or bondable plastic films. Their solvents. Often, considerable quantities may
removal is essential to create a path for be necessary to purge the contaminated core,
volatiles to escape rapidly. creating potential environmental hazards. In
many cases, complete core replacement may
be the only appropriate action.
Removing face sheet materials
Evacuation of honeycomb sandwich panels is
39.5.3 SPECIFIC EVACUATION TECHNIQUES
most effectively done by removing one of the
face sheets. This method exposes the core and
Evacuation of fluids from core with face
allows the thorough flushing of any cQntami-
sheet removed ('open core' evacuation)
nants with an appropriate solvent. Complete
drying should be performed under vacuum Visible liquids should be evacuated by blow-
pressure at moderate heat. ing filtered, compressed air across the surface.
This should be followed up by flushing the
core with an oil-free solvent and then allowing
Flushing contaminants with solvents
the solvent to evaporate completely. Next, sev-
Evacuation of chemical contaminants may be erallayers of breather fabric are stacked over
accomplished by flushing !he core cells with the panel, the assembly is envelope-bagged
 Typical repairs 865

and a vacuum of approximately 67 kPa (20 the structure. Wet lay-ups normally involve
in Hg) is applied. The panel is then heated the use of the same type of fabric used in the
slowly to approximately 74°C (165°F) and original construction, in conjunction with a
allowed to remain at temperature for a mini- laminating resin capable of room temperature
mum of one hour. cure under vacuum pressure. The quality of
the repair is generally enhanced by applying
moderate heat (100°C max) by means of heat-
Evacuation of fluids from core with face
ing blankets, heat lamps or hot air.
sheets intact ('through-the-facesheet')
evacuation
Heating blankets
First, all protective coatings and moisture bar-
rier plies must be removed from the areas to be Heating blankets used in conjunction with
evacuated. Then the geleoat of the outermost vacuum pressure repairs should have an out-
ply should be abraded to expose the fibers. put (watt density) of no less than 7750 W 1m2
(Fibers inadvertently damaged during this (5 W lin2). To facilitate draping over curved
process require subsequent repair.) Next, sev- surfaces, heating blankets with silicone rub-
eral layers of breather fabric are applied and ber-embedded elements are preferred over
the assembly envelope-bagged. Then a vac- mineral fiber-insulated pads, because of their
uum of 34-40 kPa (10-12 in Hg) is applied and inherent flexibility. Stiffer pads should only be
the panel heated very slowly (5°C per minute used on flat surfaces.
maximum heating rate) to approximately 75°C
(165°F) and maintained at that temperature
Heat lamps
and vacuum pressure for a minimum of 24 h.
After this initial drying cycle, the temperature Heat lamps that are used either as the primary
should be increased to 107°C (225°F) and heat source, or as a means of augmenting other
maintained for an additional four hours. heat sources, should be 250-300 W tungsten or
quartz tube, explosion-proof types. When
using heat lamps as the primary source, the
Handling of dried details - inspection and
effective heat input is controlled by the stand-
storage
off distance, as shown in Fig. 39.3. To avoid
After drying, details should be re-examined overheating any portion of the assembly being
and, if satisfactory, stored in a clean, dry envi- repaired, thermocouples should be placed at
ronment until the appropriate repair actions several locations to monitor the temperature
can be taken. throughout the cure cycle. The stand-off dis-
tance or the positioning of the lamp should be
adjusted as necessary to maintain the cure tem-
39.6 TYPICAL REPAIRS
perature within specified limits.
39.6.1 WET LAY-UP REPAIRS AT AMBIENT OR
ELEVATED TEMPERATURES Hot air blowers
So-called 'wet lay-up' repairs are the most fre- Hot air blowers similar to hair dryers are often
quently recommended because they require used to accelerate resin cure; they may also be
only the most basic in terms of equipment, used for reticulation of unsupported film
tooling, and repair materials. On the other adhesives. Such devices are typically designed
hand, they are also the most limited in terms with 1000-2000 W heater elements and fan
of size and applicability because such repairs drive motors.
do not restore the full, pre-damage strength of
866 Repair aspects of composite and adhesively bonded aircraft structures

3. Sand off any protective or decorative fin-

ishes and coatings; scrape off sealants,
especially silicone sealants.
4 Inspect detail for presence of water or other
15 fluid contamination.
5. Evacuate panel using one of the methods

-
14 described in Section 39.5.
6. Remove damaged materials - face sheets,
U)
doublers, and core, using the appropriate
CD
.c 13 techniques described in Section 39.4.
u
c:
:;. Note: Steps (5) and (6) may be inverted,
TA 12 depending on the condition of the part.
is
11
Core plug repair
10 1. Obtain and prefit replacement core plug,
using same as original material, cell size,
and density (or an approved substitute).
2. Apply resin compound to edges of core
°c 55 66 77 plug to provide a shear tie and insert the
plug into the cavity.
°F 130 150 170 3. Apply release film, breather fabric, thermo-
couples, and vacuum bagging materials.
Fig. 39.3 Heat lamp stand-off distance against tem- Apply vacuum and check bag for leaks.
perature.
4. Cure core splice (shear tie) resin, observing
the appropriate time/temperature relation-
39.6.2 USE OF ADDmONAL PLIES OVER WET
ship specified for the core splicing resin.
LAY-UPS
Maintain vacuum pressure throughout the
Recognizing that wet lay-ups are inherently cure cycle.
inferior to autoclave-cured laminates, many 5. Remove bagging materials and thermocou-
repair specifications9 call for the addition of ples.
two or more plies of the type of material used
in the original construction, as a means of
Face sheet repair
compensating for the loss of stiffness implicit
in wet lay-ups. The added plies do, however, 1. Taper and splice joint area.
result in weight gain and some loss in aerody- 2. Sand core plug flush with innermost ply.
namic cleanness. 3. Vacuum up sanding dust, solvent clean
repair surfaces and allow solvent to dry
completely.
39.6.3 TYPICAL WET LAY-UP REPAIR
4. Using same as original fiber type and
PROCEDURES
weave style, and observing proper yarn ori-
entation, prepare and impregnate each
Damage assessment and removal
repair ply of fabric with an appropriate
1. Determine perimeter of damaged area by laminating resin mixture.
NDpo. 5. Apply repair plies, observing ply stacking
2. Clean area with solvent. sequence and fiber orientation.
 Typical repairs 867

6. Apply perforated release film, breather / 9. Oil-free solvents and clean cheesecloth
bleeder fabric, thermocouples, and vacuum wipers.
bag. Apply vacuum and check bag for 10. Heat lamps and/or blankets, hot air gun.
leaks. 11. Thermocouples and temperature monitor-
7. Cure laminate under 67-81 kPa ing equipment.
(20-24 in Hg) vacuum pressure, while 12. Compressed air and vacuum source capa-
observing the appropriate time/tempera- ble of being regulated.
ture relationship specified for the repair 13. Environmental conditions: Work should be
resin. Maintain vacuum pressure through- done indoors, under conditions of moder-
out the cure cycle. ate temperatures (ambient) and low
8. Remove bagging materials. relative humidity (40-65%).

Restoration of coatings, finishes and 39.6.4 WET LAY-UPS USING PRECURED

sealants PATCHES
1. Ensure resin is fully cured (must be hard Instead of repairing damaged face sheets 'ply-
when tapped and resistant to solvents when for-ply', using dry fabrics and laminating
wiped with solvent-soaked cheesecloth). resins, prepreg materials may be precured,
2. Reactivate surface by mild abrasion. between layers of peel ply fabric, under auto-
3. Clean surface and allow to dry. clave conditions, and stored for later use as
4. Reapply finishes, including any anti-static patching materials. Precured patches should
and lightning-protective coatings that may be perforated to facilitate resin flow and to
be required. provide vacuum contact. (Perforations should
5. Reapply any sealants or other coatings be of sufficient diameter and spacing to pro-
removed for repair. vide a vacuum path and resin bleed, without
causing resin starvation at the bondline.)
Perforated precured materials may be used
Materials, tooling, equipment and repair
with laminating resins, film adhesives, or
environment 25
adhesive paste. When using laminating resins,
1. A two-part epoxy laminating resin of the 3-5 wt% fumed silica ('CAB-O-SIL'TM, made
required chemistry. by Cabot Corporation, is generally specified)
2. A compatible core splice resin or com- or an eqUivalent thickener should be used to
pound. improve resin filleting on the honeycomb core.
3. Same as original type and style of fabric
(unidirectional tape may be replaced with
Limitations of precured materials
two plies of fabric of equivalent thickness,
if allowed by the local structural repair Precured carboni epoxy patches are normally
manual). applied only over flat surfaces. Precured
4. Same as original type, cell size, and density glass/epoxy patches (or similar low modulus
core material, or an approved substitute. fiber) may be applied over mild curves. Since
5. Release film materials, both solid and per- precured patches are basically 'scab' patches,
forated. they should not be used if the surface requires
6. Breather and bleeder fabrics. a high degree of aerodynamic cleanness. If
7. Vacuum bagging film, vacuum gage, and precured patches are needed for repairing sur-
vacuum sealer tape. faces having compound shapes, special
8. Abrasive discs, hand-held pneumatic contour molds must be fabricated and used as
motor. a strongback to precure the material, so as to
868 Repair aspects of composite and adhesively bonded aircraft structures

produce a precise contour match. The pre- 39.6.5 PREPREG REPAIRS

cured patch may be regarded as the composite
equivalent of a metal stamping. Autoclave repairs
Restoring damaged laminates by utilizing the
Necessity of reducing vacuum. pressure for same as original preimpregnated fabric or
bonding precured details and for co-curing tape, at the same as original cure temperature
prepregs and film adhesives and pressure, is normally recommended when
full restoration of the original design proper-
Available test data indicate that precured
ties is a requirement. In practical terms,
patches, as well as prepregs and film adhe-
however, full restoration should only be
sives being co-cured (unless cured in an
attempted by depot level facilities, since any
autoclave under positive pressure conditions)
such action necessitates:
should be processed under 33-40 kPa (10-12
in Hg) vacuum pressure only. Significant 1. removal of the affected part from its parent
reductions in bond strength have been assembly;
observed when such repairs are cured under 2. availability of strongback tooling to main-
heating blankets and at full vacuum pressure. tain contours;
Many combination repair techniques, utilizing 3. an autoclave capable of meeting the origi-
both precured and resin-impregnated dry fab- nal cure parameters;
rics, prepregs and film adhesives have been 4. availability of same as original materials of
developed for specific damage conditions and construction and requisite facilities, equip-
damage locations. All such repairs should be ment and NOT capabilities.
cured under reduced vacuum pressure.
Clearly, only major operators have the neces-
sary capabilities to conduct what can only be
Wet lay-ups cured at elevated temperatures described as a remanufacturing operation. At
the present time, only a limited number of the
Unlike low temperature (65°C maximum) cur-
major airlines have the requisite eqUipment to
ing resins, laminating resins capable of being
perform rebuilds to OEM specifications. To
cured under vacuum pressure and up to 150°C
satisfy market demands, a number of repair
(300°F), so-called 'room temperature setl ele-
facilities have been granted remanufacturing
vated temperature post-cure' resins, produce
authority under Part 145 of the Federal
high quality repairs and are therefore consid-
Aviation Regulations.
ered desirable alternatives to prepreg repairs.
Because of the hazards inherent in all elevated
temperature repairs, especially non-autoclave Non-autoclave repair methods
repairs performed under vacuum pressure
For limited damage requiring only partial
only, a cautious approach is necessary. It is
restoration, there are approved alternative
almost universally recommended to cure the
repair methods. Nearly all these utilize
repair under the lowest possible cure temper-
prepregs and film adhesives that are normally
ature at the expense of elapsed time. A given
cured by means of heating blankets under
resin may be curable in two hours at 150°C
'vacuum pressure only' conditions. Such
(302°F) and may require six hours at 85°C
repairs can be carried out with minimal capital
(185°F). To reduce the risk of part failure dur-
investments. Because such repairs yield lower
ing the cure, it is generally advisable to opt for
than original design strengths, size limitations
the longer cure cycle at the lower temperature.
usually apply. These limitations are contingent
upon the specific location of the damage as
 Typical repairs 869

defined in the Structural Repair Manual for part.) It is generally accepted that, before con-
the aircraft in question. As a rule, repairs in the templating a prep reg repair, the following
vicinity of a load path, as defined by finite ele- factors be given serious consideration.
ment analysis, are severely restricted. The
1. The part must be completely dry (see
allowable repairs in so-called 'field areas', i.e.
Section 39.5).
at some predetermined distance away from
2. If at all possible, the part should be enve-
spars, ribs, hinge and latch points, etc., are
lope-bagged to prevent backskin disbonding
more generous in terms of size as well as
during the cure.
repair method.
3. The cure should always be effected at the
lowest permissible temperature specified
Vacuum pressure/heating blanket repairs (using for the product.
prepregs and film adhesives) 4. If at all possible, a repair prepreg and/or
film adhesive should be selected that is cur-
In situ prepreg repairs are often preferred over
able at a temperature 40-60°C (104-156°F)
wet lay-up / elevated temperature repairs
lower than the original cure temperature.
because the resin content of the repair is more
This is of particular importance when
easily controlled by using a prepreg. One of
repairing structures originally cured in the
the risks associated with the use of production
170-180°C (33&-356°F) temperature range.
prepregs and adhesive films is that these prod-
(Several such products are becoming avail-
ucts were formulated for production and
able as a result of persistent industry
normally require high cure temperatures
demands. Representative products are
which, when applied to damaged parts likely
listed in Table 39.1. This listing is given for
to contain residual moisture, may cause severe
reference only and does not imply endorse-
disbonding of the remaining, thus far undam-
ment of any given product.)
aged, structure. (The repair action thus
severely damages or effectively destroys the

Table 39.1 Repair adhesives and resins curable at reduced temperatures

Product Class Manufacturer Min. cure temperature Max. use temperature

°C OF °C OF
FM300-2 Film AmCy 120 250 150 300
FM250 Film AmCy 112 230 80 180
FM73 Film AmCy 105 225 120 250
FM123-5 Film ArnCy 95 200 120 250
EA9680 Film Hysol 120 250 150 300
EA9394 Paste Hysol 95 200 150 300
PL795 Film BFG 120 250 177 350
CYCOM919 Resin ArnCy 112 235 70 160
SP377 Resin 3M 95 200 105 220
Epon9410 Resin Shell 80 175 See Note
DER329 Resin Dow Room temperature See Note
Epocast35A/927 Resin Furane R. T. + postcure See Note
Note: Postcure raises upper use temperature.
870 Repair aspects of composite and adhesively bonded aircraft structures

Voids and porosities in vacuum-pressure cured as panel edge close-out in preference over
laminates and bondlines metal stampings.
Major disadvantages of 'vacuum-pressure-
only' cures are a reduction in the compaction 39.6.6 SURFACE PREPARATION FOR NON-
of the laminate and the inevitable formation of METALLIC SUBSTRATES
porosities in the laminate and/or adhesive
bondline. The finished repair yields, as a rule Abrasion and cutting of plies
of thumb, approximately only 80% of the
Taper-sanding is the preferred method of creat-
strength of an autoclave-cured part in terms of
ing a scarf joint at the substrate/repair
shear and flexural properties. The problem of
interface, especially if the substrate material is
compacting thick laminates may be overcome
made from a woven fabric. Repairs in unidi-
to some extent by hot debulking each ply, or a
rectional tape laminates often use step joints,
stack of several plies of a laminate, under vac-
with each repair ply butted against the original
uum pressure before the final cure. This
ply. Instead of sanding, the splice joint is then
method is labor-intensive but useful; it draws
prepared by cutting each ply carefully with a
off entrapped gasses, improves resin flow,
sharp instrument such as an 'Exacto' knife. It is
fiber wet-out and therefore overall laminate
common to use a lap of 13-19 mm (0.5-0.75 in)
quality.
per ply, although there is lack of agreement
with respect to the optimum lap distance or the
Prepregs co-cured with film adhesives stacking sequence of fabric plies, i.e. whether
the smallest or the largest ply should be placed
Repair technicians often use a compatible film . first. Some authorities calculate the overlap as
adhesive together with a prepreg when mak- a function of materials thickness (e.g. L = 18T)
ing a repair. A layer of film adhesive is whereas others recommend a straightforward
especially desirable as a bond ply over honey- 13 mm (0.5 in) overlap per fabric ply and a
comb core because it enhances the honeycomb 25 mm (1 in) overlap per tape ply in the zero
peel strength by providing a deeper glue fillet degree orientation3,11-13.
than would be achieved with prepreg alone.
There is, however, beside the added cost, a
slight weight gain that must be considered Use of peel plies
when repairing a weight and/or balance criti- Multi-stage processes using precured lami-
cal part. nates often use peel ply fabrics which, upon
removal, yield a surface that requires no fur-
ther cleaning or abrading. Chapter 29
Prepregs applied over metal substrates
contains some important observations about
Prep regs applied over metallic substrates peel piles.
always require the use of a layer of film adhe-
sive between the metal and the non-metal. The
Grit blasting
metallic substrate also requires the normal
surface preparations applicable for metal Grit blasting followed by solvent wiping is
bonding, by one of the methods described in sometimes used to prepare non-metallic sub-
Section 39.6.8. Prepregs are often used to pro- strates for subsequent bonding and
vide debris protection in damage-prone areas laminating operations. Plastic media with a
of thin-skinned sandwich panels, notably Mohs hardness of 3.0-3.5 (US Plastic and
wing flaps and other panelling in line with the Chemical Corporation's Polyextra and
landing gear. Occasionally, prepregs are used Polyplus granulated plastics, sieve size 30/40,
 Typical repairs 871

propelled at a low incident angle (15-30°) and Cleaning, deoxidizing, anodizing, bonding primer
at moderate nozzle pressure (25-30 psig) have application /cure
been demonstrated to remove coatings effec-
For optimum joint strength and bond durabil-
tively without damage to fibers, and to leave
ity, all metal surfaces that are to be adhesively
surface conditions of high qualitYS.
joined require the following essential steps: (1)
degreasing; (2) alkaline cleaning; (3) deoxidiz-
39.6.7 BASIC REPAIR JOINT PREPARATION ing; (4) low voltage anodizing in chromic or
phosphoric acid; (5) application, and (6) pre-
Whatever the specific surface preparation
baking of a bonding primer. For other than
method, the focus must be on producing a
complete rebuilds, which imply complete tear-
smooth, contamination-free, activated bond
down of the bonded elements, stripping of all
surface capable of promoting adhesion and,
adhesive residues, and full reprocessing of
after the cure, capable of transferring the struc-
details through solution tanks, tank etch-
tural loads across the joint with minimal
ing/ anodizing and primer prebaking are often
disruption of the load path and minimal stress
omitted at the expense of repair quality and
build-up. Stress risers of any kind, abrupt
longevity. Comparable values of various pre-
changes in thickness, brittle adhesives, the
bond surface treatments are shown in Fig.
wrong scarf angle, poor detail fit-up, preloads,
39.4.
etc. should be avoided.

Non-tank anodizing
39.6.8 REPAIR OF METAL BONDMENTS
A process known as PANTA (Phosphoric Acid
Honeycomb panels with metal face sheets Non-Tank Anodizing) exists but requires
extensive preparatory work, equipment and
Because thin-skinned honeycomb sandwich
special skills, and has therefore not been fully
panels are the most easily damaged, structures
accepted by the industry at the present time.
of this type are most often in need of repair.
Parts processed in this manner have been
Several kinds of repair activity are considered
demonstrated to be almost equivalent to tank-
typical by the industry:
processed parts in terms of bond strength and
1. minor repairs consisting of the application durability14.
of cold or hot bonded metal patches;
2. partial skin and/or core replacement with
Surface preparation for in situ non-autoclave
or without the benefit of autoclave pres-
repairs
sure;
3. rebuild or remanufacture (considered depot Typically, the repair patches or partial
level repair). replacement skins are cleaned, acid etched,
anodized, primed and prebaked. Structure
not amenable to tank solution processing, i.e.
Aluminum surface preparations
the lap joint areas of the structure being
The quality of the repair is directly related to repaired, is typically prepared with an acid
the quality of the surface to which the adhe- paste, followed by a deionized water rinse,
sive is applied. Poorly or inadequately air drying, and spray application of a bond-
prepared bonding surfaces are the primary ing primer without, however, the benefit of
reason why bonded repairs fail. prebaking. Elevated temperature prebaking
is generally impossible without exposing the
structure to heat damage and is therefore
872 Repair aspects of composite and adhesively bonded aircraft structures

LOSS IN SHEAR STRENGTH OF 2024 -T3 SAMPLES

BONDED WITH FM 123·5 ADHESIVE, AFTER 30 DAYS
AT 1200 F AND CONDENSING HUMIDITY

UNEXPOSED I
EXPOSED I TANK ETCH
+ PHOS. ANODIZE + CIP

UNEXPOSED I
EXPOSED I TANK ETCH + CIP

I
EXPOSED I TANK ETCH

UNEXPOSED I
EXPOSED I SCOTCHBRITE, MEK + ALODINE

UNEXPOSED I
EXPOSED I SCOTCHBRITE, MEK + PASA • JEL

UNEXPOSED I
EXPOSED I SCOTCHBRITE" MEK

I I I I I I
o 1000 2000 3000 4000 5000 8000

SHEAR STRENGTH, (PSI)

Fig. 39.4 Effect of various alwniniwn surface treatments on repair bond strength and durability.

omitted, at some sacrifice in terms of bond primers, have been demonstrated to produce
strength and durability15. joints of considerable durability and should be
encouraged in preference over abrasion with
aluminum oxide paper only.
Abrasive cleaning of lap joints
For reasons of expediency, many repairs are
Application of bonding pressure
effected under conditions considered marginal.
One common practice is to abrade the joint
Vacuum bagging and bondline thickness control
area with aluminum oxide paper, followed by
solvent wiping and the application of the adhe- For non-autoclave repairs, the most common
sive. Repairs of this type, whether the adhesive method of applying bonding pressure is by
selected be a paste or a film, are rarely of long means of a vacuum bag. Film adhesives used
duration and should be considered 'interim' for repair are normally scrim-supported and
repairs only. On the other hand surface prepa- thus provide bondline thickness control. (A
rations using three-dimensional abrasives such listing of representative film adhesives avail-
as Scotchbrite®, a product of the 3M Company, able with supporting scrims is provided in
in conjunction with high quality bonding Table 39.2. This listing is for reference only and
 Typical repairs 873

Table 39.2 Scrim-supported film adhesives

Manufacturer Product designation Cure temperature range Max. use temperature

°C of °C of
Curing at 120°C (250°F)
ArnCy FM73 107-150 225-300 120 250
FMI23-2 107-120 225-250 120 250
FMI23-5, FM137 95-120 200-250 120 250
Hysol EA9628 113-120 235-250 95 200
3M AF126-2, AF163-3 113-120 235-250 120 250
AF3109-2 107-177 225-350 See Note
Narrnco' Metlbond 1113 95-143 200-290 See Note
Metlbond 1133 95-135 200-275 See Note
B.P. Goodrich Plastilock 717B 107-120 225-250 82 180
Curing at 177°C (350°F)
ArnCy FM61, FM150-2 163-177 325-350 150 300
FM96 160-177 320-350 177 350
FM300 163-177 325-350 177 350
FM400 163-177 325-350 204 400
FM300-1 150-177 300-350 150 300
FM350 171-182 340-360 177 350
Hysol EA9689 177-182 350-360 177 350
EA9649R 175-180 345-355 177 350
3M AF191 175-180 345-355 177 350
AF131-2 175-180 345-355 204 400
AF143-2 175-180 345-355 177 350
AF147 175-180 345-355 150 300
Narrnco' Metlbond 328 163-191 325-375 150 300
Metlbond 329 135-185 350-365 204 400
Metlbond 1515 163-177 325-350 150 300
B.P. Goodrich Plastilock 729-3 171-182 340-360 177 350
Note: Use temperature increases as a function of cure temperature.
a Now marketed by American Cynamid

not an endorsement for any given product.) line porosities that inevitably result in lowered
When using paste adhesives, scrim cloth is bond strength, which must be taken into
normally inserted between the adherends to account during the repair design. In an effort to
prevent adhesive squeeze-out and resin star- overcome these negative effects, a unique bag-
vation in the bondline. ging method called 'double-bagging' was
developed a number of years ago. This method
provides for an inner, 'low vacuum' bag 34 kPa
Bondline porosities resulting from vacuum
(under 10 in Hg) for expelling volatiles, and an
pressure
outer, 'high vacuum' bag 81-98 kPa (24-29
The repair technician must be aware that not in Hg) to provide the equivalent of 41-47KPa
all film adhesives are equally suitable for bond- (12-14 psig) bonding pressure on the assembly.
ing under vacuum pressure; indeed, most The intent of this method is to minimize the
products are formulated for positive (Le. auto- effect of full vacuum pressure on the resin dur-
clave) pressure applications. After curing ing cure by isolating the laminate within a
under vacuum, some adhesives exhibit bond- separate diaphragm.
874 Repair aspects of composite and adhesively bonded aircraft structures

Application of mechanical pressure Pressure damage affects primarily details made

from light-weight honeycomb core, which is
Mechanical pressure applications are some-
easily crushed if not adequately protected by
times used when both sides of the part are
well-anchored support blocks. Bondments are
accessible for clamping. Anacoustical (sound
especially vulnerable to damage when the
suppression) panels having perforated or oth-
integrity of the vacuum bag is breached and
erwise permeable skins make the application
compressed gasses enter the assembly. Well-
of vacuum impossible unless the panel is
designed fixtures, proper padding of potential
envelope-bagged. To do so generally requires
puncture sites, bagging films of high quality,
extensive tear-down of the assembly and
pressure levels appropriate for the materials
removal of the affected panel. Permeable face
and part configuration, as well as constant
sheet materials (perforated metal, feltmetal,
monitoring of the pressure cycle are imperative
permeable glass fiber/polyimide laminates, or
to prevent damaging parts during the cure.
various fine mesh wire cloth acoustic sheet
Table 39.3 shows the bonding pressures consid-
materials) may be successfully bonded under
ered typical. It should be noted that for
mechanical pressure, using liquid adhesives,
assemblies incorporating honeycomb (sand-
pastes, or unsupported reticulating film adhe-
wich structures) the recommended pressures
sives.
are predicated on the compressive strength of a
core being simultaneously subjected to both
Sand bags autoclave pressure and elevated temperatures.
Panel edges are occasionally collapsed during
Pressure application methods employing sand cure, if not properly supported against side
bags, shot bags, etc., exist but are cumbersome loads. Bevelled edges should have angles
and yield repairs of marginal quality and between 15 and 20°, as shown in Fig. 39.5.
questionable durability. Such repairs should Steeper angles require that the core be stabi-
always be rendered 'fail-safe' by the addition lized with additional resins or core fillers to
of mechanical fasteners to provide a secondary prevent collapse under pressure.
load path should an adhesive failure occur
during subsequent flight service.
Application of heat

Inflatable bladders How to introduce the proper amount of heat

for curing the repair adhesive or resin has long
A novel pressure application method using an been considered problematic. Unlike manufac-
inflatable rubber bladder system has report- turing processes, which can be optimized
edly been successfup6. The loads generated by through cure cycle verification by destructive
the inflating bladder must be reacted out testing, most repairs are rather unique and
against one or several hard points on the air- influenced by a multitude of factors not easily
craft structure, which requires equipment of controlled. Heating blankets of a constant watt
model-specific design geometry. density tend to overheat thin sections, e.g. the
trailing edges of a bondment, while undercur-
ing the bondlines located over a heat sink, e.g.
Specific risks associated with pressure
a heavy metal fitting or a spar. It has often
Parts may be damaged during repair through been found necessary to protect thin sections
improper pressure application. Most damages against overheating by inserting silicone rub-
occur in the autoclave and are the result of poor ber pads between the heating blanket and the
fixturing and of insufficient attention being part, thus reducing the effective cure tempera-
paid to proper vacuum bagging techniques. ture in selected areas, while allowing the
 Typical repairs 875

Table 39.3 Recommended bonding pressures

Core material Thiekness!cell size Density Max pressure

em in kglm3 p.ef kPa psig
Honeycomb Sandwich and Metal/Metal Panels (Aluminum Bondments)
Aluminum over 12.7 over 0.5 48 or higher 3 or higher 118 35
less than 3 84 25
Aluminum under 12.7 under 0.5 48 or higher 3 or higher 35 40
less than 3 101 30
No core n/a n/a 169-338 50--100
Laminates and Panels Containing Non-metallic Honeycomb Cores
Aramid" 3.1 1/8 48-64 3.0--4.0 118 35
4.7 3/16 48 or lower 3.0 or lower 84 25
56-88 3.5-5.5 101 30
635 1/4 all 84 25
9.5 3/8 all 84 25
No core n/a n/a 152 or 45 or
higher higher
,. Nomex HRH, HRP, HMX or similar core materials.

colder portions to reach the appropriate cure alarms may prevent part damage during
temperature. Occasionally, heat lamps or other the cure cycle.
auxiliary means must be employed in conjunc- 3. Heat sinks may drain away heat energy
tion with heating blankets to provide required for resin cure, leaving residual
additional heat inputs at critical locations to uncured materials of unacceptable struc-
make sure the resins are fully cured. It is tural value. Hot bonding should not be
imperative that thermocouples be used at as carried out during adverse atmospheric
many locations as necessary to monitor the conditions or while the aircraft is cold-
cure cycle and to ensure the repair meets spec- soaked.
ification requirements when completed. 4. Improper heat-up rate control may cause
resin flow and gel anomalies resulting in a
product of marginal quality. Heat-up rates
Specific risks associated with heat
must be monitored or appropriate control
The principal risks associated with repair devices used.
activities on structures that require the use of 5. Foaming adhesives may generate exother-
thermosetting resins and adhesives are: mic reactions resulting in irreversible
damage. This hazard can be avoided by
1. Water or residual moisture in any portion of
minimizing the width of splice gaps to be
the assembly may vaporize and cause addi-
filled through careful sizing and fit-up of
tional damage such as ply separation, core
details prior to and during lay-up. Non-
node bond separation, or skin-to-core bond
metallic core details should be joined by
failure (see Section 39.5).
crush splicing rather than by adhesive
2. Overheat conditions may develop under a
foams to reduce the amount of reactive
heating blanket, causing irreversible dam-
polymers present in the panel during the
age, occasionally a fire. Constant
cure cycle.
monitoring or the incorporation of overheat
876 Repair aspects of composite and adhesively bonded aircraft structures

vvvvv
5052
11111 0°
1000/.
15°
80%
30°
76%
45·
19%
90·
2%

HRH 100% 86% 70% 15% 2%

-
()
a::
0
II:
I-
0
100

80

60
. .
.... .... .....
......
""
'.,
\
•

•
AI 5052
3/16-5052-001

NON-METALLIC
::r: \ HRH 10-118-3.0
I-
a: \
0 40
\
~
° \
,
20
I, ...
..... ....
0° 15° 30° 45° 90·

OFF-AXIS BARE COMPRESSIVE STRENGTH

OF HONEYCOMB CORE
b

Effect of core edge bevel

on core stability under
bonding pressure
c

Fig. 39.5 (a) Bare compressive strength of honeycomb at various angles of loading; (b) Off-axis bare com-
pressive strength of honeycomb core; (c) Effect of core edge bevel on core stability under bonding pressure.
 Typical repairs 877

Lower cure temperatures enhance repair safety be the choice for repairs on aluminum, because
of its galvanic compatibility and its favorable
Problems associated with hot bonding
CTE with respect to aluminum. Graphite is
increase exponentially as a function of cure
inherently incompatible for the same reasons
temperature. Cure temperatures in excess of
and may be suitable only for repairing titanium
IS0°C (360 0 P) are several times more likely to
substrates. Composite repairs can be made by
result in a failed part than repairs performed
applying multiple layers of prepregged fabrics
at lower temperatures. Given the option,
or unidirectional tapes, which may be cured by
repairs should always be conducted at the
means of heating blankets under vacuum or
lowest practical cure temperature, using a
mechanical pressure. Overlays of this type have
suitable adhesive or resin system. See Table
been reported to enhance the fatigue life of con-
39.1 for typical products.
ventional metal structures by several orders of
magnitude21• One major disadvantage is the
39.6.9 COMPOSITE REPAIRS APPLIED TO need for a chemically prepared surface involv-
METAL STRUCTURES ing the use of acids which, if entrapped under
the repair, could cause corrosion and premature
Resin-impregnated fiberglass cloth repairs structural failure, making periodic NDT of the
on aluminum repair mandatory for the remainder of the air-
frame life. Por maximum effectiveness,
Wet lay-up, epoxy-impregnated fiberglass
phosphoric acid non-tank anodizing (PANTA)
cloth repair patches have been approved repair
and the use of bonding primers are essential.
methods via OEM Structural Repair Manuals
When using graphite as the backbone fiber, a
and Military T.O.s for many yearsl. Utilizing
barrier ply of fiberglass is necessary to prevent
room temperature curing resins in conjunction
galvanic coupling between the repair material
with fabrics, wet lay-up repairs can be applied
and the substrate. The risks involved have thus
over flat as well as curved surfaces with a min-
far inhibited the use of graphite on aluminum.
imum in equipment and under almost any
conditions. Experience has shown, however,
that in terms of overall quality and durability,
39.6.10 MECHANICALLY FASTENED REPAIRS
these repairs are the least desirable and should
be applied only when more advanced methods With increased use of composite materials in
are unavailable. All resin-impregnated cloth primary and principal structure not readily
repairs over metal require the use of primers removable from the aircraft after a damage
(typically nitrile rubber based liquids) to pro- incident, bolted repair concepts are being vali-
mote resin-to-metal adhesion. dated for major skin/stringer and skin/chord
damage repairs. Utilizing mainly precured
composite elements together with metal dou-
Repairs utilizing advanced fibers
blers and splice angles, such repairs can be
Repairs utilizing boron/epoxy and graphite/ effected where access is limited to one side of
epoxy prepregs over aluminum substrates have the structure only. Essentially, such repairs are
been under active consideration for use on mil- a logical extension of, and quite similar to,
itary aircraft for some time17- 21 • Only recently, a conventional mechanically joined metal
large freight carrier made the decision to apply repairs, except that both metallic and precured
boron/ epoxy patches on some of its large trans- composite elements are utilized. The Boeing
port category aircraft, making this the first time B-777 is the first major program to approve
that boron is being used for this purpose on a this type of repair on its primary structures,
commercial fleet in the USA. Boron appears to chiefly its all-graphite composite empennage.
878 Repair aspects of composite and adhesively bonded aircraft structures

39.7 TECHNICIAN TRAINING AND SKILL 39.8.2 INSUFFICIENCY OF TECHNICAL

REQUIREMENTS TRAINING

Personnel engaged in designing and carrying Technical school curricula as well as regula-
out repairs to bonded and composite aircraft tory guidance materials dealing with
structure should be familiar with the funda- composite aircraft repairs are suffering from a
mental concepts listed in Table 39.4. technology lag that can, ultimately, only be
bridged by greater emphasis on education and
39.S CONCLUSION AND SUMMARY training at all levels. Community colleges and
vocational schools should encourage the
39.8.1 TECHNOLOGICAL EVOLUTION active participation of people experienced in
OUTPACING TRADmONAL AIRCRAFT the field (even though they may not possess
MAINTENANCE SKILLS AT ALL LEVELS the requisite academic credentials) and seek to
enlist the help of subject matter experts.
It has been observed that with every new gen-
eration of commercial aircraft, there is an
increase in the utilization of composite materi- 39.8.3 DESIGN AIRCRAFT FOR
als and a corresponding increase in the MAINTAINABILITY
complexity of its design. The transition from
Design criteria focused on rnanufacturability
simple hand layed-up bonded aluminum hon-
without regard to maintainability may ulti-
eycomb sandwich and fiberglass-skinned
Nomex panels to monolithic carbon fiber mately result in compromising flight safety,
structures produced largely by means of auto- especially now that composites are finding
mated equipment has been a long, inexorable, increasing use in primary and principal air-
craft structure. New regulations mandating
and not altogether painless process from the
,damage-tolerant' designs should be of great
operators' point of view.
The end users, principally the world's value in the determining the design criteria of
commercial airlines, are finding it increas- future aircraft. Airline customer involvement
ingly difficult to keep up with the rapid in the design of new aircraft must go beyond
payload, range, and other marketing concerns;
technological changes thrust upon them by
the time has come for the designer to solicit the
the manufacturers. Despite the large volume
of technical literature available on the subject comments and suggestions of the maintenance
of composites, there is a dearth of practical engineer, the inspector and the mechanic.
information. One is tempted to say that the
industry is encumbered by a surfeit of highly 39.8.4 NEED FOR STANDARDIZATION
specialized data that is impenetrable to all
but the experts. There is no denying that this Standardization of repair methods, practices
research is both necessary and beneficial; and especially, repair materials is long over-
there is, however, a dire need to make this due. The emphasis must be placed on the
data amenable to all through thoughtful dis- typical and generic, rather than the peculiar
tillation. The worker in the field must know and proprietary. Cooperative efforts involving
the practical effects of this research on the manufacturers, materials suppliers, airlines,
daily exercise of his craft. He must be kept repair facilities, regulatory agencies, profes-
abreast of technological advances and given a sional societies as well as academia will be
chance to upgrade his skills in order to meet needed to ensure the long-term viability of
ever-changing demands. composites in aircraft structures.
 Conclusion and summary 879

Table 39.4 Composite repair training topics

Components of Composite Materials

Fibers and filaments Glass and quartz; carbon and aramid; boron
Product definitions Fiber; roving; strand; yarn; woven fabrics, unidirectional tape; milled fibers
Fiber mechanical Density; strength; modulus; coefficient of thermal expansion
properties
Finishes and sizings Chrome; silanes; plasma treatments; resin solutions
Specialty fabrics Scrims; peel plies; bleeders and breathers; ceramics
Bonding adhesives, Epoxies; polyesters; phenolics; polyimides; bismaleimides; catalysts and
resins, prepregs hardeners; cyanate esters; acrylics; anaerobics; liquid adhesives; primers;
coupling agents; film adhesives; prepreg fabrics and tape
Procedures of Fabrication and Processing
Laminating The 'Laminate Code'; isotropic, anisotropic, quasi-isotropic laminates; cross-plied
laminates; hybrid laminates; anacoustic laminates
Other structures Filament/tape winding; RTM; braiding; pultrusions
Curing methods Autoclave; non-autoclave; single and multi-stage cures; postcuring; cure
monitoring - flow / gel/ set
Vacuum bagging Bagging films; sealant tapes; breathers and bleeders; bagging techniques
Adhesive bonding Pastes; liquids; films; cements; pressure application
Sandwich panel Face sheets, doublers and close-outs/pans; core; properties of sandwich
construction construction: static strength and rigidity; adhesive IDleting; shear ties
Core materials Metal/non-metal honeycomb; cell sizes and shapes: hexagonal, overexpanded,
flexcore; core density/weight; directional properties; compressive and shear
strengths
Surface preparation Cleaning, etching, anodizing, primer application; aluminum, steel, titanium,
for metals other metallic adherends
Surface preparation Abrasion; grit blast; taper sanding lap joints; step joints; use of peel plies
for non-steps
Joining and fastening Bonded joints; mechanical joints; fastener types and alloys; hole spacing; edge
distances; hole sizes for composite joints
Machining of Sawing; routing; drilling; sanding and grinding; water jet cutting; laser cutting
composites
Protective Coatings and Sealants
Organic polymers Pinhole fillers; sanding sealers; primers; surfacers; enamels: epoxy, polyurethane;
polysulfide coatings and sealants
Anti-static and Carbon-IDled enamels; flame spray coatings; EMI shielding materials
lightning protection
Heat and fire Silicone coatings; ablative and intumescent coatings; heat-resistant enamels
protection
Environmental Effects
Moisture Effect of moisture on uncured resins
Effect of moisture on cured systems
Temperature Effect of temperature on uncured resins
Effect of temperature on cured systems; glass transition
temperature; heat deflection temperature
Corrosion Galvanic corrosion: carbon/metal couples
Corrosion prevention
Erosion Effect of rain and particulates
Radiation Ultraviolet; thermal; nuclear
Continued on next page
880 Repair aspects of composite and adhesively bonded aircraft structures

Table 39.4 (Continued)

Ahnospheric Static charges; lightning strikes

electricity
Chemicals Aircraft system fluids: oils, hydraulic, deicing
Accidental exposures: paint stripper, solvent spills
Aging Fatigue and embrittlement effects on composites
Standard Tests for Adhesives and Prepregs
Metal adhesion Tensile shear; T-peel; honeycomb peel; crack extension (wedge) test
Prepregs Volatile and resin content; resin flow; gel time; tack
Cured laminates Interlaminar shear; short beam shear; flexure; 45° in-plane shear; tensile and
compressive strength and modulus; sandwich beam; hot/wet strength
Inspection and Quality Controls
Non-instrumented Visual inspection; tap test; penetrants
Instrumented X-ray; moisture detector; pulse-echo ultrasonic; through-transmission ultrasonic;
resonance ultrasonic; eddy current
Process quality Raw materials handling and storage; environmental controls; processing
controls materials controls; facilities and equipment controls; tool design and alteration;
detail preparation; in-process sampling inspections and witness coupons
Post-repair NDI Verification of compliance with specifications: cure cycle chart review; physical
tests and checks
Damage Assessment, Failure Analysis, Preventive Maintenance
In-service damage Foreign object impact; environmental degradation
Failed repairs 'Lesson learned'
Periodic inspections Coatings and sealants
Electrical continuity check
Specific Repair Methods
Repair categories Construction type; original materials; original cure temperature
Repair preparation Stripping; damage assessment and removal; decontamination
Repair materials Adherends and adhesives; auxiliary/processing materials
selection
Surface preparations Metals; non-metals
Adhesive/resin cure Applying heat and pressure; cure and postcure; cure monitoring
Assembly completion Reassembly; finishes; weight and balance; final inspection
Health and Safety Aspects of Composites
Chemical exposure Dermal; ocular; inhalation; ingestion
routes
Hazard levels Acute vs. chronic toxicity
Material safety data The MSDS and how to interpret
Basic industrial Engineered control systems; personnel protection
hygiene
Toxicology of Resins and catalysts; solvent and diluents; fibers and fiber dust; sealants and
composite materials coatings
 References 881

REFERENCES 13. Kuperman, M.H., Graphite/Epoxy Repair Program

Test Results, Internal Report for United Airlines,
1. Anon., Advanced composite repair guide, Contract San Francisco, CA 94128, 1983.
No. F33615-79-3217, Air Force Wright
14. Locke, Melvin c., Non-Tank Phosphoric Acid
Aeronautical Laboratories, Wright-Patterson
Anodize Method of Surface Preparation of
AFB, OH 45433, 1982.
Aluminum for Repair Bonding. Report pre-
2. Anon., Environmental Durability of Speed Tape.
pared for inclusion in Adhesive Bonded Aerospace
Summary Report by Boeing Materials
Structures Standardized Repair Handbook,
Technology, Renton, WA, 1982. .
Technical Report AFML-TR-77-206, Air Force
3. Anon., Guidance Material for DesIgn,
Wright Aeronautical Laboratories, Wright-
Maintenance, Inspection and Repair of Patterson AFB, OH 45433,1978.
Thermosetting Epoxy Matrix Composite Aircraft 15. Kuperman, M.H., Bond Strength and Bond
Structures. Montreal/Geneva: lATA, Doc. Durability Study for the Evaluation of Various
Gen/30431991.
Surface Treatments for Aluminum Adherends.
4. McKague, Lee, et al., Test of graphite-fiber siz-
Internal Report for United Airlines, San
ing effects upon laminate properties. SAMPE '., Francisco, CA 94128, 1975.
Nov /Dec, 1979.
16. Molent, L. et al., Design of an All-Boron/Epoxy
5. Egan, William, Composite Paint Stripping
Doubler Reinforcement for the F-111C Wmg
Development. Manufacturing Development Pivot Fitting: Structural Aspects. In Composite
Report No. 6-35081. Boeing Commercial
Structures Oxford: Elsevier Science Publishers,
Airplane Company, Seattle, WA 92124, 1985.
1989.
6. Hertz, Julius, Moisture effects on the high tem-
17. Anon., Adhesive Bonded Aerospace Structures
perature strength .of fiber-reinfor~e.d. resin Standardized Repair Handbook Air Force
composites. ConvaIr Aerospace DIVISIon of
Materials Laboratory (AFSC) , Wright-Patterson
General Dynamics/Hercules Inc. Joint Study
AFB, OH 45433. Technical Report AFML-TR-77-
NAS 8-27435, 1972.
206,1978.
7. Dexter, H. Benson and Donald J. Baker, Flight
18. Baker, A.A., Fibre composite repair of cracked
Service Environmental Effects on Composite metallic aircraft components - practical and
Materials and Structures. NASA Langley basic aspects. Composites, 1987, 18(4).
Research Center, Hampton, VA 23556. 73rd
19. Baker, A.A., Boron Fibre-Reinforced Plastic
AGARD Structures and Materials Panel
Patching for Cracked Aircraft Structur~s.
Workhop, San Diego, CAOctober 7-8,1991.
Lecture delivered to the Melbourne, Australia,
8. Crossman, EW, and Flaggs, D.L., Dimensional
Branch of the Royal Aeronautical Society.
stability of composite laminates during envi-
Aircraft, September 1981.
ronmental exposure. Lockheed Palo Alto
20. Sandow, Forrest A. and Raymond K. Cannon,
Research Laboratory, Palo Alto, CA. SAMPE J.
Composite Repair of Cracked Aluminum Alloy
July/August, 1979. Aircraft Structure. Final Report AD-A190-514,
9. Anon., Structural Repair Manuals (All Models),
Flight Dynamics Laboratory, Wright-Patterson
Boeing Commercial Airplane Company, Seattle AFB, OH 45433, September 1987.
WA,98124.
21. Kelly, Larry G., Composite Repair of Cracked
to. Seidl, A.L., Inspection of composite structure. Aluminum Structure (Fatigue Life Extension
Report prepared for ATA!IATA/SA~
Study). Air Force Wright Aeronautical
Commercial Aircraft ComposIte RepaIr
Laboratories, Wright-Patterson AFB, OH 45433,
Committee (CACRC), Washington, D.C.
Undated Report
Meeting, December 3-5, 1991.
22. Anon., Atmospheric Electricity - Aircraft
11. Anon., Bonded Component Repair Manual
Interaction. AGARD Lecture Series No. 110.
(BCRM), Boeing Commercial Airplane
NATO Publication, printed by Technical Editing
Company, Seattle, WA 98124. Document D6-
and Reproduction, Ltd., Harford House, 7-9
51169, 1983.
Charlotte St., London, UK, 1980.
12. Anon., Repair Procedures for 250/350 d~g. F Cur~d 23. Fisher, Franklin, and Plumer, J. Anderson,
Aramid Fabric/Epoxy and Aramld/Graphlte Lightning Protection of Aircraft. NASA Reference
Fabric/Epoxy Hybrid Composite Structures. Boeing Publication 1008, National Aeronautics and
Commercial Airplane Company, Seattle, WA Space Administration, Washington, DC 20456,
98124. Document D6-48908, 1982. 1977.
882 Repair aspects of composite and adhesively bonded aircraft structures

24. Springer, George S., (Ed.) Environmental Effects

on Composite Materials. Westport, CN:
Technomic, 1981.
25. Anon., DoD/NASA Structural Composites
Fabrication Guide, prepared under Contract No.
F33615-79-C-5125 by Lockheed-Georgia
Company for Air Force Wright Aeronautical
Laboratories, Wright-Patterson AFB, OH 45433,
1982.
REUSE AND DISPOSAL 40
Harry E. Pebly

40.1 INTRODUCTION developed but the problem of markets for the

reprocessed composites remains largely
As the United States and the world become
unsolved.
more conscious of the shortage of landfill
Plastics recycling is generally classified into
space and the need for conservation, they will
four categories by final product or by-product,
be driven to find ways to solve or alleviate the
as follows:
problem by developing methods for reuse or
other disposal of materials. Composite materi- 1. Primary: Recycling a plastic product into an
als offer a particular challenge because of the identical or similar new product;
nature of the reinforcements and fillers. In the 2. Secondary: Recycling a plastic product into
case of thermoset composites, the matrix or a new product that has less demanding
resin cannot be remelted or reprocessed as can physical or chemical properties;
be done with thermoplastic matrices. Many of 3. Tertiary: Converting plastic wastes into
the non-composite thermoplastics are rou- basic chemicals or fuels through a process
tinely collected and recycled in towns and such as pyrolysis;
cities throughout the world. This chapter will 4. Quaternary: Burning the plastics and recov-
primarily deal with thermoset composites ering their heat energy.
because of the challenge of finding methods to
Composites are recycled in all of these cate-
put these materials back into use, and in a
gories, with the fourth, burning for energy
viable economic scenario. Most of the recy-
cling effort involves reuse of sheet molding recovery, of the least value.
compound (SMC) because of its considerable
potential in the automotive market and its 40.2 HISTORICAL PERSPECTIVE
emerging economic viability. Technology
involving SMC is usually applicable to bulk Recycling of certain commodity thermoplas-
molding compound (BMC) also. Thermoplastic tics, particularly polyethylene (PE) and
composites will also be discussed. Hybrid polyethylene terephthalate (PET), has
composites, presumably because of their rela- become a routine matter in thousands of
tively small usage and more complex makeup, cities and towns in the past few years. There
have had no reported recycling attention, but have been upturns and downturns in recent
some of the reuse technologies described years in balancing availability of PE and PET
herein should be useful. Several practical tech- with demand for products and markets. One
nologies for recycling and reuse have been example is the demand for plastic soda bot-
tles for conversion to polyester for garments
(Disenhouse, 1994). The situation illustrates
Handbook of Composites. Edited by S.T. Peters. Published
an important point, common to all industries:
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 it is not only a question of developing tech-
884 Reuse and disposal

nologies for reuse but making it economically Carbon fiber

viable to do so.
Carbon fibers can be recovered from uncured
Activity in recycling thermoset composites
epoxy prepreg scrap by two methods. One is
in Europe, particularly Germany, precedes
by use of an organic solvent agitated bath
that in the USA. The action is driven by legis-
which washes out the resin. The second
lation and public opinion, as is the case in
employs thermal degradation, below 400°C
Japan also.
(750°F), to eliminate the epoxy. The resulting
fibers are cut to length ±3 mm (0.10 in) for use
40.3 RAW MATERIAL FORMS with aligned discontinuous fibers and provide
improved drapability for complex shapes
40.3.1 REINFORCEMENTS (Norris, 1990). Because fiber and fiber scraps
are randomly oriented, the cutting operation
Glass fiber should be carried out in two directions, neces-
sitating a special cutting device (Richter and
Glass fiber is the most widely used reinforce-
Brandt, 1987). Aligning short fibers results in
ment in the world, with nearly 1.2 million
composites with mechanical properties that
metric tons used in 1992 to reinforce 50 or more
are below those of long fiber reinforced mate-
different engineering thermoplastic and ther-
rials, but considerably above those of short
moset resins. Various processes used to recycle
fiber reinforced injection and compression
or separate the glass from the resin affect the
molded composites.
nature and quality of the recovered glass. First
quality glass can be reused directly, without
further processing, in a new composite mater- 40.3.2 RESINS
ial. Other processes yield a glass that will
require chemical resizing to renew its interfacial Work has been done to develop epoxies, poly-
bonding to the resin. When the glass suffers imides and unsaturated polyesters for
deterioration in strength or contamination in recovery (Section 40.4.2). The reversal of
recycling, the glass fiber must be refibered or chemical reactions utilized to produce ther-
remelted before it can be used again. moset and thermoplastic polymers to
Commercially viable technology that will regenerate monomers or beginning chemicals
allow widespread use of recycled and contam- is covered below (Section 40.4.4). It is possible
inated glass in reinforcement manufacturing to extract polyester resin from uncured SMC
processes is now being developed. and reuse the glass fiber in both SMC and
Implementation of this technology will open BMC. The extraction is done by solvent
up glass remelting and fiberization as a viable extraction with mechanical agitation.
recycling approach, making it possible to con-
template truly closing the loop for glass
40.3.3 PRElMPREGNATED MATERIALS
reinforcement (Graham, Jutte and Shipp, 1993).
(PREPREGS)
In thermoplastic recycling, a mixed stream
of resin often occurs. Addition of virgin glass A multi-year program, beginning in 1991, is
fibers has been shown to provide improved underway on carboni epoxy prepreg scrap
properties in mixed thermoplastics. Use of recycling at the Center of Excellence for
recycled glass fibers would therefore benefit Composites Manufacturing Technology
both thermoplastic recycling as well as glass. (CECMT), operated for the US Navy by the
The glass could be obtained from recycled Great Lakes COmposites Consortium (GLCC).
thermoplastics or thermosets, either as a The priority of the scrap program at CECMT is
regrind or from one of the other approaches. illustrated by the fact that for every
 Reuse technologies 885

0.45 kg (lIb) of composite materials used in flakes and molded into test panels. Mechanical
aircraft, ship or missile manufacture, and physical properties were found to be 80%
0.90 kg (2Ib) of material ends up in landfills. of that obtained with virgin prepreg. Work
The average ratio is 1.3 to 1. This waste was then begun to use material directly from
includes both material generated by the mate- the shredder, to avoid cutting the high tensile
rial supplier and material generated by the strength material into 50 mm (2 in) flakes.
end item manufacturer. In time, composite Early indications are that approximately the
scrap produced at repair facilities will add to same strengths can be obtained with the ran-
this total. Approximately $1 billion of all types domly oriented material. Preliminary analysis
of raw prepreg is wasted annually and $25 indicates that most epoxies can be comingled
million is spent on its disposal, according to and coprocessed.
information prepared for the Naval Industrial Niche markets are being sought for uses of
Resources Support Activity (Broudy, 1993). the reclaimed prep reg, based on the 80%
Another concern is with the legal status of strength obtained and the comparatively lim-
prepreg waste: is it considered hazardous? ited amount of carboni epoxy produced in the
Both state and federal regulations must be US, estimated to be 5.5 x 106 kg (2.5 million lb)
considered. California, for one, has require- annually. Markets are sought which would
ments for cured and uncured waste and its provide highest profitability for the manufac-
disposal. Other state requirements may vary turer; sporting goods and medical prosthesis
for both material forms. devices are possibilities.
The objective of the CECMT program is to
develop a plan for reclamation of
40.4 REUSE TECHNOLOGIES
carboni epoxy prepreg scrap and demonstrate
the recyclability of the material, based on a Cured thermoset composites undergo the fol-
survey of users which revealed that carboni lowing steps in preparation for reuse. Cutting
epoxy is the major prepreg contributing to the complete items (e.g. boats) into small pieces
waste stream. This scrap or waste material is suitable for handling for shredding is the first
sometimes referred to as 'offal'. The aerospace step in the process. Items 1.35-1.80 m 2
industry generates 60-80% of the prepreg (15-20 ft2). in size or scrap (cured or uncured)
offal, with 50% of the total in broad goods may go directly to shredding. Shredding
form. The most widely used epoxide is reduces the material to sizes suitable for the
tetraglycidylmethylenedianaline (TGMDA) next step, whether it be particle preparation
cured with diaminodiphenylsulfone (DDS). (grinding, milling, granulating) or pyrolysis.
Other accomplishments of the program, as Shredding will not be necessary for pyrolysis
reported by Lause (1993) of GLCC, are as fol- if the item can be accommodated in the pyrol-
lows. A low speed, high torque shredder was ysis furnace. Classification of the shredded
found to be best for cutting high tensile and material or of the resulting particles will be
high shear strength carbon fibers. Other cut- required if more than one product is desired
ting blades evaluated were guillotine, carbide, (e.g. glass fiber and powder). Shredding for
slitter and shear types. A low torque machine most experimental work provides chips (or
did not give a clean cut. The shredder employs pieces)50 mm x 400 mm (2 in x 8 in) to 50 mm
two cutting blades, rotating inwards or x 50 mm (2 in x 2 in); granulation gives 9 mm
counter to one another, powered by a 30 hp (3/8 in) and smaller particles; and grinding or
motor. Shredder input was the scrap prepreg milling a powder, suitable for filler, of 50 J.Un
collected from Gerber cutters, i.e. random size or less. Size of particles and powder is deter-
and shape. The offal was cut into 50 mm (2 in) mined by screen size.
886 Reuse and disposal

40.4.1 MECHANICAL GRANULATION OR granulator, a hammer mill pulverizer and a

MILLING classifier. Each piece of machinery is commer-
cially available; custom design is not required
Mechanical granulation is probably receiving
(Butler, 1991).
the most attention as a method for reuse of the
In the USA, the SMC Automotive Alliance,
final composite, whether it be waste from the
Owens-Corning and Premix are among the
molding process or the cured product.
leaders in fiberglass composite recycling, but
Polyester (unsaturated) SMC, because of its
other proprietary recycling activity is also
predominant place in the total spectrum of
being carried on (Section 40.10).
thermoset composites, is the material form of
Canadian work is being done by Phoenix
most interest. For example, the total automo-
Fiberglass Inc. and Plastiglas Industries
tive usage for SMC in 1993 is estimated to be
(Darrah, 1993). Phoenix Fiberglass, through a
342 x 106 kg (155.6 x 106 1b) and the SMC
mechanical process, is able to separate the two
Automotive Alliance projected a 55% jump to
components of a laminate that include the
mO.re than 594 x 106 kg (270 x 106 1b) by 1995
glass reinforcement in the form of fibers and
(WIgotsky, 1993). Phenolic and urethane-based
the thermoset resin in the form of a powder. In
thermoset composites have also been demon-
the pilot plant, the waste laminate was
strated to be recyclable by mechanical milling
inspected for contaminants before it was intro-
to a reinforcing filler. Further, it appears that
duced into the recycling process, with the
fully milled thermoset can be reused as filler in
result that long and short fibers as well as two
an indefinite number of cycles without loss of
types of fillers or resin extenders were recov-
performance.
ered.
Mechanical recycling methods are based on
Plastiglas Industries Ltd recycles its own
using SMC scrap directly without alteration of
waste composites, but has developed the
its chemical nature. 'The SMC scrap is shred-
m~thodology to combine waste composites
ded and/ or ground into a form which can be
WIth other materials from other sources, as
used as a filler. Ball or hammer milling, cryo-
well as with other virgin materials. 'These
genic grinding and knife granulating are used
other materials may come from the blue box
for mechanical granulation. A typical granula-
program (a curbside collection effort), from
tor uses double cross-angled blades rotating
sandb~asting operations that want to dispose
against an opposing angled blade, while a
of therr waste from sanding operations, and
hammer mill reduces material size by the
even crushed concrete.
hammer revolving at high speeds and cutting
Materials that have been identified for use
by impact in mid air. 'The entire SMC article
a~cording to ~ certain recipe will be sorted by
may be ground to a fine powder (including all
SIZe. Dependmg on the end-use, pieces that
the glass fiber). 'The other approach is to shred
range in size from 6-200 mm (0.25-8 in) can be
and mill the SMC article in a manner which
used. When the material that is to be used in a
will recover some glass fibers which are sepa-
recipe is larger than 200 mm (8 in), it is
rated from the particulate material (powder).
reduced to the required size. 'This part of the
Recovery of some glass fiber is important
operation may use a specially built water-
from a cost-effectiveness basis because fibers
cooled diamond wire saw, which can handle
have more commercial value as a reinforce-
pieces up to and including 3 m (10 ft) in diam-
ment filler (versus particulate filler). Also, it
eter and/ or 3 m (10 ft) square. If further size
requires more energy to grind all of the fiber
reduction is required, shredders and grinders
into a powder, then to recover a portion of the
are used to achieve the proper particle size
glass fibers. Premix found the most economi-
A significant German program was
cal process to regrind scrap consisted of a
sparked by public and political pressure upon
 Reuse technologies 887

the plastics industry to become responsible Milled SMC is being used in several automo-
for the reuse and cost of their products at their tive parts. Up to now there has been no
end-of-life (Schaefer and Plowgian, 1993). The technical development reported of a fiber
major companies from the SMC/BMC market recovery method. Pyrolysis is another impor-
founded ERCOM Composites Recycling tant reuse technology.
GmbH. The proposed German Refuse Act for The Swedish Institute of Composites con-
car recycling, with its priority for material centrated on evaluating several proprietary
recycling, also provided impetus for action. formulations of SMC using two reground frac-
ERCOM offers a complete system to close the tions: one powdered fraction with sizes in the
loop between used parts from automotive ser- range of 200 J..lm - 1 mm and one fibrous frac-
vice and disassembly plants and the reuse of tion with sizes of up to 500 J..lm (Pettersson and
fibrous reinforcing material in new SMC com- Nilsson, 1993). The reground SMC came from
pounds: automotive bumpers supplied by ERCOM.
Existing grinding and fractioning techniques
• a mobile shredding truck that crushes used
were used.
parts at disassembly and production sites;
• transport the compacted material to a cen-
tralized fractionizing plant; 40.4.2 REVERSIBLE CROSSLINKING
• produce a range of fiber rich recyclate mate-
The Department of Energy has sponsored
rial to sell back to SMC producers and other
work at Polytechnic University, Brooklyn,
end users.
New York, to develop experimental thermoset
The mobile shredder and fractionizing plant polymers designed for recovery (Tesoro et al.,
have been in operation since the beginning of 1992). Tesoro has shown that it is feasible to
1992. The mobile shredder reduces the size of recover chemical compounds and polymers
large parts to a chip size of about from epoxies, polyimides and unsaturated
50 mm x 50 mm (2 in x 2 in). This is a volume polyesters. Designing the polymers for recy-
reduction of a factor of five. The commonly cling by introducing labile bonds is utilized for
used metallic inserts in automotive parts do epoxy resins crosslinked with disulfide-con-
not have to be removed beforehand. In fact, taining curing agents and for disulfide-
preshredding to this predetermined chip size containing polyimide copolymers. Stress-
is an essential requirement to assure removal strain curves for epoxy resins cured with a
of metallic parts in the next processing step. disulfide-curing agent, dithiodianiline
The preshredded material is then processed (DTDA) and a commercial aromatic diamine
automatically through a series of pneumatic curing agent (MDA) show the equivalence of
and mechanical fractionizing steps. It is essen- the DTDA to MDA. Solubilization of the cured
tial to remove metallic parts in this sequence to thermosets by reduction of disulfide bonds
assure a metal free product. A hammer mill is provides sites for re-curing. Epoxies may also
used for the grinding step. The resulting fiber- be crosslinked with a polyfunctional reagent
rich recyclate can be produced in a series of which reacts with thiol groups formed in
fractions containing a mixture of glass fibers, reduction. Tesoro has also worked on obtain-
fillers and resin. ing chemical products from cured unsaturated
The emphasis in Japanese composite recy- polyesters (UP) (Tesoro and Wu, 1993). The
cling has been in dismantling or cutting up focus of the investigation was on chemical
large fabricated parts, particuarly fishing ves- reactions of well established UP structures to
sels and boats, bathtubs, and tanks (Kitamura, yield soluble, well-characterized chemical
1993). Shredding is best done by a non-gearing products that would be suitable as raw mate-
type or a rotary shock shearing type shredder. rials for synthesis. Neutral hydrolysis was
888 Reuse and disposal

found to be a viable approach through cleav- carried out experiments in cooperation with
age of ester bonds, giving phthalic acid in Michimae using a batch type external heating
good yield and soluble carboxyl-containing pyrolysis furnace (Kitamura, 1993). The sys-
oligomers of potential utility were isolated. tem was conducted for scrapped reinforced
Recovered oligomers for acid curing of plastic (RP) ship material. When material is
epoxy resin systems have been explored with heated for about 10 min, gas is produced and
encouraging results. Smooth transparent films more gas can be collected as the temperature
perhaps suitable for coatings have been rises. The amount of gas which can be col-
obtained. The use of recovered oligomer as lected reaches a peak at 27SoC (S27°F). The gas
saturated acid in the ex novo synthesis of UP produced is either water cooled or run
resins has also shown promise. through a heat exchanger to yield some pyro-
oil and the rest, gas. The RP used in the
experiment had a 40 wt % of glass content as is
40.4.3 PYROLYSIS
common for RP used in ships. The residue is
Pyrolysis is the controlled thermal degrada- fiberglass and carbon, i.e. 40 wt % and 12
tion of an organic material into one or more wt %, respectively.
recoverable substances through the applica- The Japan Marine Recreation Association
tion of heat in an oxygen-free environment. under the Ministry of Transport and Sasakawa
Pyrolysis should not be confused with inciner- Foundation has been developing a mobile
ation, which is the total oxidation of all the pyrolysis/incineration system for RP ships
organic materials. Pyrolysis has been used for since 1990. It is economically advantageous for
years to reduce plastic and rubber wastes, par- large structures like RP ships to be treated as
ticularly the stockpiles of used tires. they are without pre-processing. The mobile
Pyrolyzing SMC in the USA, however, is rela- pyrolysis/incineration system comprises two
tively new. trailers, each of which has a body 7 m (23 ft)
Primary emphasis to date has been applied long, and 2.6 m (8.6 ft) wide, holding a
to scrap materials in the manufacturing destructive distillation furnace and a rotary
process. Ultimately, recycling of parts at the kiln, respectively. The system first pyrolyzes
end of their application life cycle must be an RP ship into pieces less than 6 m (20 ft) in
addressed if elimination of all SMC from land- length under a temperature condition of about
fill is to be achieved. 400°C (7S0°F). After shredding into small
Pyrolysis has a major advantage in that it is pieces, removing metals and moving to the
well suited to handling contaminated scrap, second trailer on which a kiln is loaded, the
i.e. SMC which contains paint, adhesives, fas- residue is then incinerated in the rotary kiln. A
teners or is mixed with other materials. pilot plant on land has been constructed and
The SMC Automotive Alliance of the SPI tested and mobile pyrolysis/incineration sys-
Composites Institute (1991), a group of mater- tem trailers will be completed in 1993.
ial suppliers and molders, has worked on
pyrolysis of SMC auto scrap. In its early
40.4.4 CHEMICAL DEGRADATION
stages, the main problem was shredding and
feeding materials into the pyrolysis equip- Under the proper conditions, the chemical
ment. A discrete chip with minimum polyester reactions that are utilized to produce ther-
breakup is needed to make the process cost moset and thermoplastic polymers can be
effective. By-product utilization is the other reversed to regenerate the various monomers
factor which will influence the feasibility of involved. Among these are the 'chem-olysis'
the process for SMC waste disposal. processes: hydrolysis, glycolysis, methanoly-
In Japan, The Ship Research Institute (SRI) sis, aminolysis, as well as polyurethane (PUR)
 Properties and products of recycled materials 889

foam and reaction injection molding (RIM) ical solvent which is blended both in the waste
depolymerization, saponification, refinery paste and waste resin, to keep them in a
cracking, and thermal 'unzipping' of poly- pumpable state, Premix has used the waste to
olefins. fuel cement kilns (Butler, 1991). This waste
Since these approaches take the organic will be used to fire the kiln and then the ash
polymers back to more basic chemical build- from the firing is put into the cement. This
ing blocks, they have the potential advantage process also eliminates the need to landfill the
of yielding known chemical streams that can ash. Some bathtub manufacturers in Japan use
be converted, using existing purification and special incinerators to recover the steam for
polymerization technology, to new polymer boilers from the waste. However, in this case,
with properties equivalent to the virgin poly- secondary treatment equipment is necessary
mer. The depolymerization processes will also to prevent air pollution.
free up the glass reinforcement used in the
composite. In the polyester field, chemical
40.4.7 DISPOSAL/RECYCLING MODELS FOR
degradation is being utilized primarily for
REUSE TECHNOLOGIES
polyethylene terephthalate (PET). Inasmuch as
this thermoplastic is primarily used without A disposal/recycling model for composite
reinforcement, e.g. bottles and film packaging, waste material has been developed by Vital
no detailed attention will be given here. Visions Corp. to help evaluate practical and
Typical products obtained in chemical degra- cost effective disposal or recovery methods
dation or depolymerization are dimethyl (Wood, 1991). The State of Minnesota has
terephthalate (DMT) and ethylene glycol (EG). funded a study for development of an eco-
Pyrolysis, another form of chemical degrada- nomic model to determine if collection of fiber
tion, has been discussed previously; it is a reinforced plastic scrap from processors can
practical recycling process for unsaturated produce profitable products, either materials
polyester (SMC). or end-use goods (McDermott, 1992).

40.4.5 HEAT CLEANING 40.5 PROPERTIES AND PRODUCTS OF

RECYCLED MATERIALS
Heat cleaning has been utilized in recovering
useful products from scrap circuit board lami- Because of the big supply and the big market
nates (Hanson, 1991). Pennisula Copper (automotive), the most effort in thermoset
Industries, after removing the copper from the recycling has been in particle recycling of glass
board, has found it can heat finished or inter- fiber reinforced polyester, predominantly
mediate thermoset composite products, SMC. It appears that fully milled thermoset
completely volatilize the epoxy resin, and can be reused as a filler in an indefinite num-
recover the glass fibers with much of the ber of cycles without loss of performance.
strength of new fibers. The key is a carefully Pyrolysis degrades the mechanical properties
controlled temperature profile as the scrap of the glass fiber but fiber is still useful as a
moves through the furnace (called a calciner), particulate filler.
and control of the process gases. The solidified The approaches for the recycling of glass
gases are also potentially useful by-products. reinforced thermoplastic composites utilize
the technology already developed for the
non-reinforced polymers used in packaging.
40.4.6 FUEL
However, the thermally induced chemical
liquid waste from producing SMC and BMC degradation of thermoplastics and the physi-
has found use as fuel. By introducing a chem- cal attrition of glass fibers that occurs during
890 Reuse and disposal

repeated extrusion and injection molding erties with the regrind. The density
require attention. decreases, as found by other researchers, in
this case about 10% at the highest loading.
LORIA® (a registered trademark of Ashland
40.5.1 SHEET AND BULK MOLDING
Oil Inc.) and Diffracto analyses were also
COMPOUNDS
conducted. The outcome of both tests show
Owens-Corning reports in Table 40.1 some that 30 phr material gave the best surface.
properties of SMC containing filler made from Beyond the 30 phr loading, the surface deteri-
recycled SMC (Graham, Jutte and Shipp, 1993). orated.
Replacing the calcium carbonate filler with In the BMC evaluation, Premix found very
ground SMC actually improves most mechani- little effect on the mechanical properties, with
cal properties. Fibers that have been recovered slight decreases in tensile strength, flexural
from SMC by mechanical separation show modulus and notched Izod as the regrind
even greater potential reinforcement value. loading increased. Owens-Corning has shown
Premix Inc. evaluated several formulation that unfractionated, granulated SMC, still con-
of SMC and BMC, using scrap materials in taining fiber with considerable length and
varying ratios (30%, 58% and 88%) (Butler, aspect ratio, provides some reinforcement
1991). Beyond 88%, the material viscosity was value in BMC (Table 40.3).
unmanageable. The SMC formula used for Work by ERCOM in Germany evaluated
the regrind studies was a typical automotive properties at three recyclate contents obtained
formulation (Table 40.2). Their data show that in SMC compounds using SMC recyc1ate
only slight changes occur in mechanical prop- obtained from scrap and used parts (Schaefer

Table 40.1 Properties of SMC containing composite filler made from recycled SMC (20% carbon carbonate
replaced with composite filler)

Standard formulation Low Density formulation

Control Recycled Control Recycled
Tensilestrength,MPa 77.6 97.9 105.6 95.5
(psi) (11250) (14195) (15310) (13850)
Tensile modulus, MPa 11590 12970 8927 10670
(psi x 10 6) (1.68) (1.88) (1.29) (1.53)
Flexural strength, MPa 186.0 210.0 163.2 195.1
(psi) (26970) (30450) (23665) (28290)
Flexural modulus, MPa 8970 9867 8683 9325
(psi x 10 6) (1.30) (1.43) (1.26) (1.35)
Notched izod, Jim 953 1018 824 888
(ft Ib/in.) (17.8) (19) (15.4) (16.6)
Unnotched izod, Jim 1164 1388 981 958
(ft Ib/in.) (21.8) (26) (18.4) (17.9)
Water absorption (%) 0.31 0.28 0.20 0.19
Specific gravity, gl cm3 1.76 1.81 1.32 1.33
(lb/ft3) (110) (113) (82.4) (83)
 Properties and products of recycled materials 891

Table 40.2 Typical automotive formulation control

PHR Basted on paste % BOT %

Polyester resin 55 1.3 14.13
Polyvinyl acetate LPA 40 13.69 10.27
Styrene 5 1.71 1.28
Tertiary butyl perbenzoate 1.5 0.52 0.38
Zinc stearate 4 1.37 1.03
Wetting agent 8 2.60 1.95
Calcium carbonate 175 59.91 44.93
Magnesium hydroxide 4 1.37 1.03
Glass 25% 25.00

30 phr regrind PHR Basted on paste % BOT %

Polyester resin 55 18.83 14.13
Polyvinyl acetate LPA 40 13.69 10.27
Styrene 5 1.71 1.28
Tertiary butyl perbenzoate 1.5 0.51 0.38
Zinc stearate 4 1.37 1.03
Wetting agent 8 2.60 1.95
Calcium carbonate 145 49.91 37.43
Regrind 30 10.00 7.50
Magnesium hydroxide 4 1.37 1.03
Glass 25% 25.00

Table 40.3 Properties of BMC reinforced with recycled glass reinforced SMC

Tensile Elongation Flexural Izod impact

SMC (%)
Stress, Modulus, Stress, Modulus, Notched, Unnotched,
Added
(%)
MFa MPa MFa MPa Jim Jim
Sample (psi) (psi x 1(j) (psi) (psi x 1(j) (jt lblin) (jt lb/in)
BMC 0 27.9 13099 0.44 96.7 10548 270 361
control (4045) (1 9) (14020) (1.51) (5.1) (6.8)
Stnd. BMC 6 16.1 9858 0.68 68.6 9789 270 278
+SMC (2334) (1.43) (9947) (1.40) (5.1) (5.2)
Stnd. BMC 12 17.3 12685 0.22 71.4 10203 209 274
+SMC (2508) (1.84) (10350) (1.46) (3.9) (5.1)
BMCResin 70 25.8 7859 0.36 55.4 6411 89 114
+SMC (3740) (1.14) (8033) (0.98) (1.6) (2.1)

and Plowgian, 1993). Table 40.4 shows that Treatment Council) of the Japan Reinforced
strength can be maintained and there is a Plastics Society, as the final step in a complete
reduction in density. The latter implies a lower recycling evaluation beginning with cutting
weight for a given car part, an added benefit up and shredding large RP parts, obtained
for the auto recycling evaluation underway. physical and mechanical property data for
In Japan, the RTC (RP Recycling and SMC where CaC03 was replaced with varying
892 Reuse and disposal

Table 40.4 Comparison of properties of SMC com- (SICOMP) performed mechanical testing on a
pounds using recyclate low-shrink SMC with 20% 25 mm (1 in) glass
reinforcement and a 50% reduction of the cal-
Base 5% 10% 15% cium carbonate filler, down to 80 phr
Flex-strength, 194 222 185 208 (Pettersson and Nilsson, 1993). Improved flex-
MPa (psi) (28130) (32 190) (26825) (30160) ural strength together with slightly reduced
Impact, kJ/m2 102 107 126 145 flexural modulus was obtained for the SMC
(ft lb/in) (19.4) (20.4) (24.0) (27.6) containing 10% regroWld material.
Density, g/ cm3 1.90 1.85 1.80 1.72
(lb/ft3) (118.6) (115.5) (112.4) (107.4) Evaluation of a virgin PET polyester and
one prepared from reclaimed PET has been car-
Note. Recyclate content calculated as a percent of total ried out by DeMaio (1991). Although the resins
weight of compound.
did not contain reinforcements and therefore
amoWlts of regrind (Kitamura, 1993). Tables may not truly belong in a book on composites,
40.5 and 40.6 present these data and confirm the test data are of interest while considering
results by other investigators that mechanical SMC and BMC above. Tables 40.7 and 40.8
properties are not harmed (in fact, in some show that no significant differences of liquid or
cases improved) and the density is lowered. physical properties exist between the two
The Swedish Institute of Composites materials once processed into a polyester resin.

Table 40.5 Physical/mechanical properties of recycled SMC - standard type

Formulation
Materials 1 2 3
Resin 100 100 100
CaC03 125 78 36
Reground 0 32 60
GF (1 in) (%) 30 30 30

Properties
Shrinkage (%) 0.06 0.07 0.08
Specific gravity, g/ cm3 1.73 1.64 1.59
(lb/ft3) (108.0) (102.4) (99.3)
Flexural strength, MPa 200 213 186
(psi) (29000) (30 885) (26970)
Flexural modulus, GPa 12.3 11.8 11.4
(psi x 1<J6) (1.78) (1.71) (1.65)
Tensile strength, MPa 78 77 79
(psi) (11310) (11165) (11455)
Tensile modulus, GPa 11.8 11.5 11.2
(psi x 1<J6) (1.71) (1.67) (1.62)
Izod impact, mJ/mm2' 82.0 87.4 86.0
(ft Ib/in) (15.6) (16.6) (16.4)

*edgewise, unnotched
 Properties and products of recycled materials 893

Table 40.6 Physical/mechanical properties of recycled SMC - class A SMC

Formulation
Materials 1 2 3
Resin 100 100 100
CaC03 180 123 76
Reground 0 38 70
GF (1 in), % 28 28 28

Properties
Shrinkage, % 0.00 0.00 0.00
Specific gravity, g/ cm3 1.85 1.77 1.68
(lb/ft3) (115.4) (110.5) (104.9)
Water absorption, % 0.44 0.39 0.30
Surface quality 3.6 3.8 5.6
average deviation (p.)a
Flexural strength, MPa 190 178 167
(psi) (27550) (25810) (24215)
Flexural modulus, GPa 11.5 10.9 10.4
(psi x 106) (1.66) (1.58) (1.51)
Tensile strength, MPa 85 82 68
(psi) (12325) (11890) (9860)
Tensile modulus, GPa 11.0 10.7 10.2
(psi x 106) (1.60) (1.55) (1.48)
Izod impact, mJ/mm2' 76.5 81.5 78.5
(ft lb/in) (14.6) (15.5) (14.9)
"edgewise, unnotched;
"Takeda method.

40.5.2 PHENOUC (PF) AND POLYURETHANE to 30% recycled content. The recycled PUR
(PUR) COMPOSITES must be ground into a fine powder before
While the major activity in thermosets has been reprocessing to obtain the Class-A surface.
with unsaturated polyesters (SMC and BMC),
property data have also been reported on recy-
40.5.3 RECYCLED THERMOPLASTIC
cled glass reinforced PF, Table 40.9 (Olson,
COMPOSITES
1992) and PUR, Table 40.10 (Graham, Jutte and
Shipp, 1993). Work to recycle reinforced PUR Glass-filled polypropylene (PP) has received
will increase as structural reaction injection the most attention, perhaps because of its
molding (SRIM) finds more use in structural potential for automotive use. Owens-Corning
applications and PUR presumably continues to studied the effects of repeated recycling by
be the resin most used. Krauss-Maffei granulation of the PP and injecti()n molding
(Germany) has a new RIM system that can pro- (Graham, Jutte and Shipp, 1993). Table 40.11
duce exterior auto body panels with a Class-A shows a loss in properties caused by polymer
surface from reinforced PUR that includes up degradation and a decreasing fiber length.
894 Reuse and disposal

Table 40.7 Mechanical property comparison data; Table 40.8 Liquid property comparison data
recycled PET
Recycled Virgin
~easured property
Recycled Virgin PET PET
~easuredpropert;v
PET PET Viscosity, cps 470 420
Flexural strength, MPa 117.9 121.4
% Solids 60.1 60.8
(psi) (17400) (17600)
Flexural modulus, MPa 3586 3586
Specific gravity 1.098 1.102
(106 psi) (0.52) (0.52) Acid number, solids basis 17 20
Tensile strength, MPa 65.5 67.6 180°F SPI gel test
(psi) (9500) (9800) (catalyzed with 1% BPO)
Tensile elongation, % 2.5 2.8 Gel time, min 6.5 6.0
Barcol hardness 49 47 Cure time, min 1.7 1.5
HDT,oC 419 414 Peak 225 229
(OF) (215) (212) temperature °C (OF) (438) (444)
Izod impact, kJl m 0.16 0.18
(ft lb/in) (3.1) (3.4)
Water absorption, % 2.0 1.8
glass fiber. The results of additional glass in
injection molded recycled PP are contained in
Two viable approaches to overcome the prop- Table 40.13.
erty loss are the introduction of polymer Jutte and Graham (1991) considered the
additives and the use of additional glass. Table effect on properties of PP containing granu-
40.12 presents data to show the improvement lated SMC. Both coarse and fine recycled SMC
in recycled PP with the addition of Po1ybond were evaluated at three filler levels. Modulus
3001 (BP Chemicals), an acid functionalized was improved and unnotched Izod impact
PP. The recycled PP was Himont Pro-Fax 6523 properties decreased; the data are tabulated in
with 6 mm (0.25 in) Owens-Corning R-34B Table 40.14.

Table 40.9 Mechanical properties of recycled glass reinforced phenolic

Regrind typical
40% resin + 50% resin + 60% resin +
properties of
60% regrind 50% regrind 40% regrind
virgin material
Density, g/cm3 1.45 1.43 1.40 1.61
(lb/ft3) (90.5) (89.3) (87.4) (100.5)
Compressive strength, MPa 252 263 273 230
(ksi) (36.5) (38.1) (39.6) (33.4)
Flexural strength, MPa 100.0 102.0 83 110
(ksi) (14.5) (14.8) (12) (16)
Notched Izod, kJ /m2 2.05 1.98 2.03 3.6
(ft lb/in) (0.39) (0.38) (0.39) (0.69)
Unnotched Izod, kJ/m2 7.8 8.6 6.4 8.0
(ft lb/in) (1.49) (1.63) (1.22) (1.52)
 Properties and products of recycled materials 895

Table 40.10 Properties of 10% regrind RIM: in glass dation. NBC SRIM systems can be ground,
filled polyurethane dried and injection molded either neat or
blended with nylon-6. Recycling evaluations
Standard + utilized materials flaked to a granular size of
Standard 10% regrind
approximately 2 mm (0.08 in). Comparison of
system <200",m
NBC SRIM properties to recycled (via injection
Flexural modulus, N/mm2 900 836 molding) part properties masks the possible
(psi x 106) (0.13) (0.12)
effects of the injection molding process due to
Tensile strength, N/mm2 28.0 24.6 the inherent differences in reinforcement
(psi) (4060) (3570) length. For this purpose, comparison of unre-
Tensile modulus, N/mm2 750 569 inforced NBC RIM and injection molded,
(psi x 106) (0.11) (0.08) recycled NBC provides a better understanding
of the effect of the recycling process on the
Elongation, % 160 133
resin matrix (Table 40.15). As the test data indi-
Hardness, Shore D 59 63 cates, the initial injection molding cycle does
Density gl cm3 1.18 1.20 not adversely affect the properties of the NBC
(lb I ft3) (73.6) (74.9) system. When an SRIM composite 35-45%
fiber reinforcement package was recycled, the
recycled SRIM composite yielded properties
In the course of developing a thermoplastic comparable to virgin injection molded impact-
matrix for SRIM, DSM RIM Nylon Inc. devel- modified nylon, indicated in Table 40.16.
oped a family of nylon block copolymers
(NBC) named Nyrim (Reitz, 1992). Nyrim
40.5.4 YIELDS FROM PYROLYSIS
chemistry combines AP-caprolactam with an
activated 'rubbery' prepolymer. NBC compos- Because pyrolysis occurs in an oxygen-free
ites can be readily recycled by conventional environment, SMC decomposes into three
injection molding techniques without degra- recoverable substances: pyro-gas, pyro-oil and

Table 40.11 Mechanical properties of recycled 30% glass reinforced polypropylene

Sample Tensile Elongation Flexural Izod Impact

(%)
Stress, Modulus, Stress, Modulus, Notched, Unnotched, Glass
MPa MPa MPa MPa Jim Jim length,
(psi) (psi x 1(J6) (psi) (psi x 1(J6) ift lb/in) ift lblin) (mm)
Control 80.0 6688 2.19 126.6 7562 103 421 361
(11600) (0.97) (18355) (1.10) (1.93) (7.90)
Recycled 69.6 6329 2.28 106.4 6749 80 321 329
IX (10 090) (0.92) (15430) (0.98) (1.49) (6.01)
Recycled 63.5 6026 2.29 100.0 6306 74 292 305
2x (9210) (0.87) (14500) (0.91) (1.38) (5.47)
Recycled 55.9 5461 2.47 89.8 5557 64 253 280
3x (8105) (0.79) (13020) (0.81) (1.20) (4.73)
Recycled 52.0 5261 2.50 86.1 5288 60 241 271
4X (7540) (0.76) (12485) (0.77) (1.12) (4.51)
896 Reuse and disposal

Table 40.12 Properties of polybond modified 30% recycled glass reinforced polypropylene

Sample Tensile Elongation Flexural Izod impact

(%)
Stress, Modulus, Stress, Modulus, Notched, Un notched,
MPa MPa MPa MPa Jim Jim
(psi) (psi X 1(0) (psi) (psi x 1()6) (jt lblin) (jt lb/in)
Control 76.5 6405 2.31 121.7 7239 101 400
(11 090) (0.93) (17650) (1.06) (1.89) (7.49)
Recycled Ix 62.5 5798 2.45 97.6 4888 73 314
(9062) (0.84) (14150) (0.71) (1.36) (5.88)
+ 5% Polybond 73.1 6143 3.64 116.1 5681 112 595
(10660) (0.89) (16830) (0.82) (2.10) (11.14)
+ 10% Polybond 75.8 6260 3.69 120.9 5632 121 638
(11 000) (0.91) (17530) (0.81) (2.26) (11.94)
+ 15% Polybond 77.2 6316 3.45 124.2 5729 126 679
(11200) (0.92) (18000) (0.83) (2.36) (12.72)

Table 40.13 Properties of increased glass content recycled reinforced polypropylene

Sample Tensile Elongation Flexural Izod impact

(%)
Stress, Modulus, Stress, Modulus, Notched, Unnotched,
MPa MPa MPa MPa Jim Jim
(psi) (psi x 1()6) (psi) (psi x 1(0) (jt lblin) (jt lb/in)
Control 76.5 6405 2.31 121.7 7239 101 400
(30% glass) (11 090) (0.93) (17645) (1.05) (1.89) (7.49)
Recycled Ix 62.5 5798 2.45 97.6 4888 73 314
(30% glass) (9062) (0.84) (14150) (0.71) (1.36) (5.88)
+ 2.5% glass 67.2 6260 2.36 101.5 5460 81 317
(9744) (0.91) (14715) (0.79) (1.52) (5.94)
+ 5.0% glass 70.3 6964 2.13 108.8 6101 82 302
(10 200) (1.01) (15775) (0.88) (1.52) (5.66)
+ 7.5% glass 73.8 7377 2.08 112.2 6722 85 305
(10701) (1.07) (16270) (0.97) (1.59) (5.71)
+ 100% glass 75.8 8067 1.87 117.4 8169 83 281
(11 000) (1.17) (17020) (1.18) (1.55) (5.26)

solid by-product (essentially glass fiber and pyrolyzed SMC can be recycled successfully
filler). In one test by the SMC Automotive into general purpose and Class A automotive
Alliance, the resulting composition was gas SMC, at up to 30% of the calcium carbonate
14%, oil, 14% and inert solid 72%. The gas gen- filler loading with no adverse effects on pro-
erated was sufficient to fuel the pyrolysis unit, cessing or physical properties (Rusch, 1993).
making it self-sustaining. The energy content The pyrolysis experiments performed by
was close to natural gas. The solid by-products Michimae in Japan on scrapped RP ship mate-
can be processed into fillers for SMC, BMC or rial gave the yields in Table 40.17. The pyro-oil
thermoplastics. When milled into a filler, the which was recovered had a strong acid content
 Properties and products of recycled materials 897

Table 40.14 Properties of polypropylene containing granulated SMC

Sample Tensile Elongation Flexural Izod impact

(%) HO-
SMC Stress, Modulus, Stress, Modulus, Notched, Unnotched,
TUL
(%) MPa MPa MPa MPa Jim Jim °C(OF)
(psi) (psi x 106) (psi) (psi x 106) (ft Iblin) (ft lblin)
Polypro- 0 22.13 1035 29.17 965 77.9 968.5 63.3
pylene (3210) (0.15) 9.72 (4230) (0.14) (0.14) (18.14) (146)
Polypro 15 21.56 1724 33.51 1310 84.4 236.5 67.2
pylene + (3120) (0.25) 3.50 (4860) (0.19) (1.58) (4.43) (153)
coarse
30 21.03 2207 34.83 1655 89.2 155.4 78.3
recycled
(3050) (0.32) 2.34 (5050) (0.24) (1.67) (2.91) (173)
SMC
50 22.43 3034 36.21 2276 101.4 148.4 94.4
(3250) (0.44) 1.61 (5250) (0.33) (1.90) (2.78) (202)
Polypro- 15 20.55 1448 32.41 1310 74.2 205.0 51.9
pylene + (2980) (0.21) 4.78 (4700) 0.19 1.39 (3.84) (151)
fine recycled
30 20.96 2069 33.93 1.655 82.2 163.9 84.4
SMC
(3040) (0.30) 2.47 (4920) (0.24) (1.54) (307) (184)
50 17.17 2828 3.076 2483 73.14 104.1 82.7
(2490) (0.41) 1.25 (4460) (0.36) (1.37) (1.95) (181)
Polypro- 0 37.93 2000 50.69 1448 150.6 361.4 114.9
pylene + (5500) (0.29) 3.45 (7350) (0.21) (2.82) (6.77) (239)
9% chopped glass

Table 40.15 Properties of recycled neat NBC RIM Table 40.16 Properties of recycled NBC SRIM part

Property Virgin Recycled Tensile modulus, MPa 10400

(psi x 106) (1.51)
Tensile modulus, MPa 2445 2480
(psi x 106) (0.35) (0.36) Tensile strength, MPa 120
(psi) (17400)
Tensile stress, MPa 58 60
(psi) (8410) (8700) Elongation, % 3
Elongation, % 40 125 Flexural modulus, MPa 8000
(psi x 106) (1.24)
Flexural modulus, MPa 2485 2450
(psi x 106) (0.36) (0.35) Notched Izod, kJ/m2 10
(ft lb/in) (1.9)
Izod, kJ/m2 10 10
(ft lb/in) (1.9) (1.9) Values are dry, as molded

Mold shrinkage, % 2 1.2 with a viscosity which resembled heavy

Melting point, °C 214 220 naptha, a flash point close to that of gasoline
eF) (417) (428) and was highly flammable.
The Government Industrial Research
Density, g/m3 1.10 1.14
(lb/ft3) (68.7) (71.2) Institute and Shikoku/MITI conducted a pyrol-
ysis experiment using steam in a gas-heated
Water absorption, % 3.1 3.1 continuous-batch method, with temperatures
898 Reuse and disposal

of typically 500°C (932°F). RP from ships con- 40.5.5 HEAT CLEANED REINFORCEMENIS
taining 40% fiberglass was used. The system Peninsula Copper Ind ustries (PC!) has
produces hardly any smoke or gas. The resin obtained data on their heat-cleaned glass fiber
yields pyro-oil and fiberglass can be recovered incorporated into polyester, polypropylene
without loss of strength. Table 40.18 shows the and nylon matrix laminates (Hanson, 1991).
yield (Kitamura, 1993). The performance of recovered glass fiber rein-
forcement is comparable to virgin glass in
many respects in polyester and nylon. Table
Table 40.17 Yield from pyrolysis - SRI/Michmae 40.19 compares PC! recovered fiber with PPG
fiber with a polyester resin. Table 40.20 pre-
Materials Yield wt.%
sents polypropylene data, comparing the base
Gas (liters) 90.7 resin properties to those using 22 wt % bare
Pyro-oil (g) 344 34.4 reclaimed glass and two formulations of 22 wt
Remainder (g) 524 52.4 % silane-treated glass. The difference between
Note: per 1000 g of RP Formula A and Formula B is not explained, but

Table 40.18 Yield from pyrolysis - GIRIS/MITI Table 40.19 Average mechanical properties of poly-
ester/PCI-Glass composites and PPG composite.
Materials wt.%
Property pel PPG
Gas (CO) 18
Pyro-oil: Tensile strength, MPa 25.0 24.9
Styrene monomer 25 (psi) (3626) (3600)
Solids (phthalic) 11 Compressive 89.7 n/a
Remainder: strength, MPa (psi) (13000) n/a
Fiberglass 40 Flexural strength, MPa 81.4 84.5
Carbon 6 (psi) (11800) (12300)
FIber length, em 10.2 6.4
Note: per 1000 g of RP
(in) (0.4) (0.25)

Table 40.20 Average mechanical properties of polypropylene/PCI-

glass composites

Property Filled, Filled,

Unfilled PP untreated treated AlB
Tensile strength, MPa 24.70 26.68 27.20/31.67
(psi) (3582) (3869) (3945/4593)
Tensile modulus, MPa 873 2120 2540/1370
(psi x 106) (0.127) 0.308) (0.369/0.199)
Yield stress at break, MPa 25 20.1 17.9/28.4
(psi) (3630) (2910) (2602/4121)
Strain at break 556 391 313/16.7
(%, at 50 nun/min)
Glass content 0.0 22.0 22.0
(volume%)
 Disposal of nonrecyclables 899

Table 40.21 Average mechanical properties of adding to landfill. Incineration leaves a

nylon 6/6 PCI-glass composites residue which normally is deposited in a land-
fill, but increasing legislature action as well as
Property w/PCI public opinion is reducing the use of landfills.
glass fiber
The energy content in SMC can be recovered
Tensile strength, MPa 102.8 by incineration, but the relatively low energy
(psi) (14905) content together with the high amount of ash
Flexural strength, MPa 184.2 residue from the incineration makes incinera-
(psi) (26690) tion of SMC quite unprofitable. The ash
Flexural modulus, MPa 5292 residues from incineration are not useful as
(psi x 106) (0.77) fillers in new SMC since the high temperatures
when incinerating, together with the presence
Izod impact strength of oxygen, converts the calcium carbonate to
(notched 1/8 in), kJ/m 0.04
(ft lh/in) (0.8) calcium oxide and that will adversely influence
(unnotched), kJ/m 0.36 the maturation process (chemically thickening)
(ft lh/in) (6.7) when making new SMC. There are no disad-
vantages, however, either in the form of added
Elongation, % 3.0
pollution or toxic emissions.
Mold shrinkage, (in/in, %) 0.006,0.6% The important aspect that SMC consists of
Glass content (vol %) 14.8 only 20-30% organic materials makes inciner-
ation and chemical degradation less suitable
compared to particle recycling (regrinding)
both show improvement over the base resin and pyrolysis which are using the inorganic
and the bare reclaimed glass. The values over- part of the material as well.
all fall far short of most of the commercially The non-metal mixture that remains,
available glass filled PP, where tensile known as 'fluff', after such items as automo-
strengths range from 29 MPa (4200 psi) to 82.5 biles and white goods (refrigerators, washing
MPa (12 000 psi) with a mean value of 44.7 machines, etc.) are processed in shredders, is
MPa (6500 psi.). normally disposed of in landfills. Seventy-five
The performance of reclaimed silane-treated percent of shredder plant feed is from cars.
glass fibers in nylon is closer to that obtained This feed mixture has a composition of
for commercial compounds than for approximately 15-25% plastics, 25-75% inerts
polypropylene. Properties obtained using PeI and 2-35% moisture.
glass are given in Table 40.21. Depending on Research by General Motors, Ford and
the resin formulation, manufacturer's data Chrysler and the American Plastics Council is
show tensile strengths for virgin glass rein- being conducted on recovery, recycling and
forced nylon range between 75 MPa reuse of plastics from scrapped cars and
(11 000 psi) and 151.8 MPa (22 000 psi). trucks. Although plastics residue from recy-
Flexural properties show a similar relationship. cling vehicles represents less than 1% of the
solid waste sent to landfills in the USA each
year, auto industry officials estimate that it still
40.6 DISPOSAL OF NONRECYCLABLES
accounts for more than 1 million tons of the
nation's solid landfill waste annually.
40.6.1 INCINERATION AND LANDFILL
Research efforts are focused on materials that
The more desirable forms of disposing of com- currently are not recovered, primarily plastics,
posites are described in the previous section, glass, fluids, sealers, fabric, adhesives, paint,
both for economic payback and avoiding and rubber (fluff). Fluff accounts for less than
900 Reuse and disposal

2 wt % of municipal solid waste. The joint lished infrastructure for recycling could head
effort targets the plastics portion of the fluff. off severe government restrictions on plastics
The recovery of plastics from shredder residue use.
is hampered by frequent contamination from As with durables in general, aids for recy-
paint, other plastics, metals, and adhesives. cling vehicles include design for easy
Thus another goal of the effort is to develop separation and collection, parts consolidation
better methods to purify scrapped plastics. to simplify separation, material identification
The Department of Energy's Argonne on parts, restriction of plastic parts to recy-
National Laboratory has developed a process clable resins, and avoidance of paints, coatings
to separate plastics and other recyclable mate- and heat sensitive additives. Automakers
rials from the 'fluff' of shredded automobiles. worldwide have agreed on an identification
The process utilizes a fluidized bed process code for all plastic parts heavier than
and yields PUR, fines, iron oxide and mixed 100 g (3.5 oz). Ford Motor Company has
plastics. issued guidelines on design for for recyclabil-
Germany appears to be moving away from ity (Miller, 1993)
its rules forbidding incineration of many kinds Germany has been a leader in design for
of waste. The Environment Ministry report- reuse. Porsche, on assignment from the
edly will introduce legislation that would Automotive Engineering Research
allow incineration of composites and certain Association, investigated designs for medium-
plastics if recycling proves too expensive priced autos in which plastic parts could be
(Protzman, 1993). readily dismantled and recycled. The best
designs for meeting these objectives were
those in which combinations of easily disman-
40.7 DESIGN FOR REUSE
tled parts were manufactured from a single
recyclable resin. An example of such a combi-
40.7.1 AUTOMOBILES
nation is a PP bumper system with a PP
An obvious approach to assist in the economic sheath, an expanded PP core and glass-mat-
success of recycling of multi-material parts is to reinforced PP beam (Nir, Miltz and Ram,
design the item originally for ease of disassem- 1993). Many reclamation projects focus on the
bly. Of particular interest are automobiles and bumper system. Bumpers are collected in pilot
white goods (e.g. refrigerators). The major auto programs, ground into particles and blended
makers of the US Council for Automotive with virgin resins. Use of a single recyclable
Research (USCAR), through their Vehicle resin has been practiced by Peugeot, Opel (sin-
Recycling Partnership, are studying ways to gle resin bumpers) and BMW (body panels).
enhance compatibility and dismantling of plas- The use of one resin in different forms, as
tic parts (Wigotsky, 1993). Some bumper parts described above, is also under development
now need only four bolts instead of twelve. by Reko and DSM in auto dashboards.
Potential amendments to the US Resource In Landshut, Germany, BMW has a facility
Conservation and Recovery Act, passed in capable of dismantling 25 cars a day. Eighty
1976, may specifically name automobile solid percent of the parts in the BMW's low end '3
waste. Federal regulations could possibly series' can be reutilized in some form. Plastic
require certain car design features, if only by bumpers are turned into interior wall panels.
excluding certain materials. The VRP is work- Seat covers and PU foam are used as sound-
ing to establish a strong infrastructure for deadening material for the floor in the '3
recycling plastic parts, starting with things series'. BMW also has three dismantling plants
that are easy to get at, such as fender liners, fan in the USA (New York City, Los Angeles and
shrouds and radiator supports. A well-estab- Orlando) (Protzman, 1993).
 Applications and markets 901

40.7.2 APPLIANCES roof-rack module which uses recycled filler.

Introduced in 1992, the rack may have been
Design for recyclability, of appliances and
the first use of recycled filler in new SMC.
other white goods, is underway at several
Germany's ERCOM and the SMCAA have
companies. Such designs allow for easy disas-
signed an agreement to 'enhance the use of
sembly, e.g. snap fit assembly with no
SMC/BMC in automotive parts by providing
fasteners, with each part preferably manufac-
an organization in each country that will
tured from a single, readily recycled resin and
accept scrap parts for recycling'. SMC is well
labeled as to its composition. Polymer
ahead of thermoplastics in automotive recy-
Solutions, a joint venture of GE Plastics and
cling use for major parts. This is because SMC
Richardson Smith, has already designed an
fillers appear to be continually reusable with-
easy-tn-disassemble refrigerator and a circuit
out sacrificing physical properties; replacing
breaker box (Nir, Miltz and Ram, 1993).
half the CaC03 with SMC filler reduces
part weight and cuts compound cost.
Thermoplastics have limitations on the
40.8 APPLICATIONS AND MARKETS
amounts or frequency of reprocessing permit-
ted before significant mechanical property
40.8.1 AUTOMOTIVE
degradation occurs. There is also concern
The automotive industry is evaluating many with possible mixing of incompatible resins.
applications of recycled composite materials, Plastic material suppliers are working to pro-
particularly SMC, in new auto parts. The tech- duce resins with recycled plastic content
nology to recycle SMC, as discussed earlier, is (Miller, 1993).
available and apparently viable. The major There has, however, been reported one
barrier to recycle SMC is the lack of an infra- successful application of a recycled thermo-
structure, i.e. collection, processing, and plastic composite. Chrysler will feature
distribution, but this network will develop recyclable fenders for all of its 1993 LH pas-
because of environmental pressures, etc. senger cars. The fenders are made of
The first production application of recy- DuPont's Bexloy K, a new glass-reinforced
cled SMC filler in the USA was by General PET polyester based composite. They can
Motors on the 1993 Corvette, an inner panel either be reground and molded again or
for the rear panel assembly (Rusch, 1993). The returned to the two pure feedstock ingredi-
part does not require a Class A surface. ents through a patented DuPont
Corvette anticipates expansion of recycled methanolysis process (American Plastics
SMC to eight more underskin panels in ensu- Council, 1993).
ing models. General Motors spokesmen predict that
In other applications, the SMC Automotive within five years there will be the necessary
Alliance (SMCAA) reported Ford Motor Co. infrastructure for recycling such parts as
expected to start using SMC with recycled fender liners, HVAC ducts, fan shrouds, radia-
composite filler on the Econoline Van interior tor supports and many interior components
engine cover during the 1993 model year, and (Wigotsky, 1993).
Chrysler will use recycled SMC on interior
trim on the RAM Van, the first parts that are
40.8.2 BUILDING CONSTRUCTION
visible. And in Europe, the SMCAA listed
three production auto parts that contain recy- A Canadian company has developed a prod-
cled SMC as filler (Audi spare-tire well, VW uct Stratum (Stratum@ is a registered
Polo engine module and VW Passat front-end trademark for the recycled materials produced
panel). Japan's Toyota Carib has an interior by Plastiglas Industries) composed of a plastic
902 Reuse and disposal

binder, usually polyester, filled with up to 85% 40.9 LEGAL AND ENVIRONMENTAL
by volume of recycled composite scrap ASPECTS
(Darrah, 1993). One application encapsulates
The major federal waste legislation in the USA
the scrap-resin mixture between two FRP
addressing solid waste issues is the Resource
skins. The resulting panel has passed the full
Conservation and Recovery Act of 1976. It
scale fire test for exterior panel systems.
requires the Environmental Protection Agency
Work in Japan has been done to develop a
(EPA) to set guidelines for government pro-
gypsum filled with milled composite ther-
curement of recycled products and it
moset scrap and a mortar with milled
mandates Federal agencies and contractors to
composite scrap, both with mechanical prop-
implement affirmative procurement programs.
erties suitable for construction (Kitamura,
Most states and many local governments have
1993).
established programs to procure recycled
materials. The 1976 Act was reauthorised in
40.8.3 ELECTRICAL PARTS 1988 and continues in effect (Donnelly, 1993).
An industry concern is that the federal reg-
In Germany, two parts made with new SMC ulations might venture into the areas of car
formulations containing 10--15% by weight design, if only by excluding, within certain
fiber rich recyclate are already in production time periods, certain materials and design
or approved for production. These are an elec- approaches.
trical distribution cabinet and a cable In the meantime, Senator Baucus (D-MT),
distribution base (Schaefer and Plowgian, Chairman of the Senate Environment-Public
1993). Strength and modulus can be main- Works Committee, unveiled in April, 1993, a
tained with a reduction in density. four-part blueprint for recycling to be part of a
recycling bill to be introduced later. Two parts
40.8.4 OTHER APPLICATIONS concern composites: the Federal Government
should take the lead in procuring goods made
Peninsula Copper Industries has identified an from recycled materials and second, manufac-
application for small chips of printed circuit turers should be responsible for their products
board in an epoxy/coal tar base (Hanson, when they become waste.
1991). This composite is used as a cushion In Germany, increasing concern for the
material between rails and the steel approach environment has led to proposed laws for the
plates at railroad crossings. Initial tests indi- reuse of post consumer and post industrial
cate that this composite is an effective waste. Legislation for recycling of consumer
replacement for rubber-based materials goods packaging was enacted in 1992, setting
presently used. Several parts from SMC, made minimums for collection and recycling
from recycled SMC, have gone into production through 1995. Now proposed is the German
or are approved for production in Germany. Refuse Act for Car Recycling, with its priority
These include sinks, seats and chairs. for material recycling (Schaefer and Plowgian,
In the thermoplastic field, a commercial 1993).
application for 'Nyrim' nylon block copoly- Other proposed legislation would levy a tax
mer-based SRIM is a manhole cover (Reitz, on new cars to pay for their eventual
1992). It utilizes a 35-45% fiber reinforcement disposal. Another proposal by Germany's
package. The recycled SRIM composite Environmental Minister would make automo-
yielded properties comparable to virgin injec- bile manufacturers responsible for the final
tion molded impact-modified nylon. disposal of their cars. Still another stipulated
that by 1993, plastics in new cars produced in
 Organizations active in composites recycling 903

Germany would be required to contain 25% by and construction/furniture.

weight recycled materials. Legislators and The SMC Automotive Alliance, Southfield,
industry will be following Germany's new MI, is an activity of the SPI's Composites
environmental laws, expecting to learn from Institute, composed of 35 SMC plastics mater-
that country's experience, especially since the ial molders and SMC raw material suppliers.
current Democratic administration has One of their objectives is to prove that pyroly-
expressed interest in a progressive environmen- sis and grinding are viable methods of
tal policy, i.e. one of reduce, reuse and recycle. recycling automotive components made of
Two new laws have been enacted in Japan SMC.
in 1991; one is 'Law Concerning the Utilization The Center of Excellence for Composites
of Recycled Resources' and the second Manufacturing Technology is operated for the
'Amendment of Waste Processing and Navy by the Great Lakes Composites
Cleaning'. Earlier, the National Diet amended Consortium, Kenosha, WI. One of the first pro-
a part of the Marine Safety Act to address the jects undertaken was a prepreg (carbon/
problem of treatment for scrapped ships. epoxy) scrap reclamation project.
Research and development on dismantling, The Big Three Vehicle Recycling Partnership
cutting up, shredding and various recycling of the US Council for Automotive Research has
methods of other large items followed. Such teamed with the Automotive Group of the
items include bathtubs or bath units, tanks American Plastics Council to create an infra-
and automotive parts. These enactments made structure for plastics recycling of scrapped cars
people strongly aware that plastic waste recy- and trucks.
cling had become an important social Suppliers of Advanced Composite Materials
problem. Association (SACMA), Arlington, VA, is a
trade association of composite material suppli-
ers, parts fabricators and organizations that
40.10 ORGANIZATIONS ACTIVE IN
provide ancillary support. Waste management,
COMPOSITES RECYCLING
including recycling composite waste, is one of
The American Plastics Council, Washington, their areas of activity.
DC, is a joint initiative with the Society of the ERCOM Composite Recycling GmbH,
Plastics Industry Inc. (American Plastics Karlsruhe, Germany, is composed of SMC pro-
Council, 1993). It was earlier known as the ducers and processors, resin suppliers and
Council for Solid Waste Solutions. The Council glass fiber producers. It was founded to
mission is to develop and implement a pro- develop and prove the recyclability of SMC in
gram for the responsible use, recovery order to maintain SMC market share in light of
(including recycling and energy recovery) and the strict German attitude towards waste man-
conservation of plastics. The Council is com- agement.
posed of 25 of the leading plastic resin Reinforced Plastic Waste Recycling and
producers, downstream customers and repre- Treatment Council, Chuo-Ku, Tokyo, Japan, a
sentatives of the plastics processor committee of the Japanese Reinforced Plastics
community. One of its four task forces is Society was organized to seek technical and
Product Stewardship, devoted to advancing social solutions in disposal and reuse of large
all options of integrated waste management. composite items such as ships, tanks and bath
This task force, under the Durables Program, units.
has identified four program groups to receive The Swedish Institute of Composites, Pitea,
recycling (life-cycle management) attention. Sweden, is doing experimental work to show
They are automotive, major appliance, com- that reground SMC can be incorporated into
puter and business equipment, and building new SMC and BMC products.
904 Reuse and disposal

ACKNOWLEDGEMENTS Nir, M., Miltz J. and Ram, A. 1993. Update on

Plastics and the Environment. Plastics
The author wishes to acknowledge Mrs Peggy Engineering, 49,(3), 87-88.
Renkel for her word processing expertise, Norris, Donald. 1990. Proc. 1990 ASME/ESD
helpful suggestions, diligence and patience in Advanced Composites Conference.
chapter preparation and Dr David Graham for Olson, Barbara and H. DeKeyser. 1992. Recycling
his careful review of the text. Cured Phenolic Material, Rogers Corporation,
Molding Materials Div No. 5215-022-1.5E.
Presented at SAE International Congress,
Session Code M25B (February).
REFERENCES
Pettersson, Joakim and P. Nilsson. 1993. 48th
American Plastics Council. 1993. Private Annual Conference, SPI Composites Institute,
Communication, T. Donnelly. What Industry Is Session 15-F.
Doing? Protzman, Ferdinand. 1993. Germany's Push to
Broudy, Phil. 1993. NIRS. Private Communication. Expand the Scope of Recycling. The New York
Butler, Kurt. 1991. 46th Annual Conference, SPI Times. July 4, sec 3:8.
Composites Institute, Session 18-B. Reitz, James. 1992. 47th Annual Conference, SPI
Darrah, Basil. 1993. 48th Annual Conference, SPI CompOSites Institute, Session 4-B.
Composites Institute, Session 1S-B. Richter, Heinz and Jurgen Brandt. 1987. Recycling
DeMaio, Anthony. 1991. 46th Annual Conference, Carbon Fiber Scrap, Engineered Materials
SPI Composites Institute, Session 18-C. Handbook 1: Composites, pp. 153-156.
Diesenhouse, Susan, 1994. Polyester Becomes Rusch, Ken. 1993. 48th Annual Conference, SPI
Environmentally Correct. The New York Times, Composites Institute, Session 15-G.
February 1994, Sec 3.5. Sahut, Jan 1994 Reinforced Plastic Making Some Big
Donnelly, Thomas III. 1993. Status Report on Fancy Parts. Plastic World, 52(10), 24
National Legislation. SPE Recycling Schaefer, Peter and A. Plowgian. 1993. 48th Annual
Conference, June 16. Chicago, Illinois. Conference, SPI Composites Institute, Session
Graham, W. David, Jutte R.B. and Shipp, D.L. 1993. 15-D.
48th Annual Conference, SPI Composites SMC Automotive Alliance. 1991. 46th Annual
Institute, Session 15-H. Conference, SPI Composites Institute, Session
Hanson, David. 1991. 46th Annual Conference, SPI IS-D.
Composites Institute, Session 18-E. Tesoro, Guiliana et al. 1992. 47th Annual
Jutte, Ralph and W. D. Graham. 1991. 46th Annual Conference, SPI Composites Institute, Session 4-
Conference, SPI Composites Institute, Session C.
18-A. Tesoro, Guiliana and Y. Wu. 1993. Chemical
Kitamura, Tatsundo. 1993. 48th Annual Conference, Products from Cured Unsaturated Polyesters.
SPI Composites Institute, Session 15-E. Advances in Polymer Chemistry, 12(2), 185-196.
Lause, John. 1993. GLCe. Private Communication. Wigotsky, Victor. 1993. Plastics In Automotive.
McDermott, Joseph. 1992. Thermosets Recycled, Plastics Engineering, 49(4), 36-38.
What Now? Reinforced Plastics, 36(11). Wood, J. 1991. A Disposal/Recycling Model for
November. Composite Waste Materials. First International
Miller, B. 1993. Recycling: What the Big 3 say. SAMPE Environmental Conference. May 1.
Plastics World, 51 (10), 32-37 pp. 388-390.
LAND TRANSPORTATION APPLICATIONS 41
Douglas L. Denton

41.1 INTRODUCTION pro~esses must be used to manufacture parts

rapIdly enough to meet the production rates of
The use of polymer composite materials in
the assembly plants. As a result, injection and
land transportation is steadily increasing
compression molding processes are used
because of the cost and performance advan-
extensively for these applications. Composites
tages .that composites offer over competing
typically used in transportation applications
matenals. Today, composites are extensively
consist of low-cost grades of thermoplastic or
used in well-established passenger car, van
t~ermoset polymers reinforced with E-glass
and truck applications. New applications
fIbers. Often these composites also contain
demanding higher structural performance lev-
mineral particulate fillers. Composites con-
els are under development. Railroad cars,
taining high-modulus fibers, such as carbon,
mass transit vehicles and a wide range of mil-
~nd higher-performance resins such as epox-
itary ground transportation systems offer
Ies, are used only where the higher cost can be
expanding opportunities for composite mate-
justified to meet special product requirements.
rials.
~etal. matrix composites are used very spar-
Driving forces for the use of polymer com-
mgly m ground transportation.
posites in ground transportation applications
Over the past 50 years, the use of polymer
include low manufacturing investment cost,
composites generally has progressed from
cost reduction through parts consolidation,
low-performance applications towards more
weight savings, good mechanical properties,
demanding applications requiring excellent
'Class A' surface quality, excellent durability
s~rfac~ appearance, high mechanical proper-
characteristics, inherent dent and corrosion
~Ies, mcreased temperature stability or
resistance, good noise and vibration dampen-
Improved durability. With increasing
ing, styling flexibility and dimensional
demands to reduce vehicle weight for
stability. Factors which tend to mitigate
improved fuel economy, and to reduce invest-
against their use are high materials costs, low
ment costs for greater competitiveness, the use
modulus and the reluctance to use materials
of composites in ground transportation is
perceived to be 'new and unproven'.
expected to increase for many decades to
Only those materials and processes which
come.
provide required performance at the lowest
cost achieve long-term commercial success in
transportation applications. For high-volume 41.2 ECONOMICS AND MARKET GROWTH
automobile and truck applications, high-speed
The steady growth in the use of composites for
land transportation is attributed primarily to
Handbook of Composites. Edited by S.T. Peters. Published the development and acceptance of new appli-
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 cations. From the late-1960s to the mid-1990s
906 Land transportation applications

composite usage in USA transportation more be equal to or lower than stamped steel parts.
than quadrupled (Fig. 41.1), and since the A primary factor offsetting high materials cost
early 1980s it has expanded at a faster rate is the lower investment needed to make parts
than the total composites market. Land trans- from composites compared to steeL To pro-
portation use of polymer composites duce a composite part generally only one
represents the largest segment (over 30%) of mold and one press is needed whereas most
the total market and reached 430 000 metric steel parts require multiple stamping tools and
tons in 1994 according to the SPI Composites presses to form the part. In addition, compos-
Institute. ites offer the opportunity to integrate several
parts formed in steel into one part. This fur-
ther reduces the number of tools and presses
~~--------------------------~
needed, and eliminates the welding opera-
tions required to join the stampings.
The cost advantage achieved or cost penalty
incurred by composites depends upon the
total volume of an application. As the number
of parts produced becomes very high, the
reduced investment in tools and presses is off-
set by the higher materials cost for the
composite. In the automobile industry the use
of a composite part produced at a volume of
several hundred thousand units per year may
not be not cost effective. But production of the
Fig. 41.1 Growth in the shipments of polymer com- same composite part at volumes of tens of
posites for use in land transportation. (Source: SPI thousand units per year may be very competi-
Composites Institute.) tive. The specific cost 'cross-over' point
between composites and steel depends on a
large number of factors and must be deter-
Composites have generally displaced metal mined for each specific application.
alloys such as steel to gain new applications.
In order for a material substitution to be suc-
41.3 HISTORICAL DEVELOPMENT OF
cessful. composites must provide all of the
APPLICATIONS
functions required by the part at a competitive
cost. Because of strong competitive forces in Polymer composites began to appear on cars
the transportation industry original equip- shortly after World War II in small compo-
ment manufacturers (OEMs) seldom pay nents under the hood and inside the
higher cost for a new material to achieve desir- passenger compartment. The first major mile-
able functions beyond the part requirements. stone in exterior parts was the introduction of
Therefore, composites must achieve cost effec- the Corvette in 1953 which sported body pan-
tiveness for long-term use in transportation els made of glass fiber reinforced polyester
applications. (Fig. 41.2). The body panels were produced
On a price per unit weight basis, composite with open mold and preform molding
material costs are generally several times processes up to 1972, when the production
higher than the metals used in automobile and was converted to compression molded sheet
truck applications. However, when the total molding composite (SMC). The invention of
cost of component production and vehicle the SMC process and the development of 'low
assembly is considered, composite parts can profile' polyester resins, which provided
 Historical development of applications 907

Fig. 41.2 Glass fiber reinforced polyester body panels on the 1953 Corvette was the first major application
of composites in the automotive industry. Other composite applications introduced on the Corvette include
bumper beams, leaf springs, radiator support, seat backs and rear floor pan. (Photo courtesy of General
Motors Corporation.)

improved surface appearance, allowed the Fuel Economy (CAFE) standards led to intense
proliferation of composites into higher vol- development of new composite applications
ume body applications. and a significant upswing in composite usage
SMC grille opening panels (GOP) were in the late 1970s (Fig. 41.1). Composites not
introduced in the late 1960s and rapidly only expanded into additional cosmetic parts,
spread throughout the industry because fiber but began to be seriously considered for use in
glass GOPs saved weight and reduced cost structural components. Intense programs to
through significant parts consolidation. With develop radiator supports, transmission sup-
the introduction of composite panels on auto- ports, leaf springs and wheels were initiated.
mobile assembly lines, the dimensional Efforts to find additional under-the-hood and
consistency of molded parts became well interior applications of composites were also
known. Manufacturers also recognized that intensified.
the SMC panels could not be deformed to Improvements in composite surface quality
adjust for variations in the steel body struc- and productivity lead to the introduction of
ture. Modification of assembly plant the 1984 Fiero, the second high-volume vehicle
procedures to accommodate the unique char- with all exterior body panels made of compos-
acteristics of composite parts continues to be ites. Using innovative body construction,
an issue with new part introduction. exterior panels produced by the SMC and
The need to reduce vehicle weight through RRIM (reinforced reaction injection molding)
materials substitution and downsizing to meet processes were mechanically fastened onto a
government mandated Corporate Average steel spaceframe. A similar design approach
908 Land transportation applications

was used for the construction of General 41.4 CURRENT APPLICATIONS

Motor's all purpose vehicles (APV).
Introduced in 1989, these minivans used 41.4.1 AUTOMOTNE APPLICATIONS
approximately 120 kg (260 lb) of SMC ~er
On average, cars produced in USA in 1994
vehicle (Fig. 41.3). With annual production
contained about 50 kg (110 lb) of fiber rein-
volumes ranging from 90 000 to 148 000 vehi-
forced composites. The broad usage of
cles per year this represented the largest sin?le
composites in the automobiles and trucks is
application of composites in the automotive
illustrated by the list of applications in Table
industry.
41.1. Distinctions can be made between appli-
The Dodge Viper was the first American
cations that are experimental, have been in
vehicle to have all body panels produced by the
limited production or are well-established in
resin transfer molding (RTM) process (Fig.
production. In many cases, breakthrough com-
41.4). The low production volumes for the
posite applications are introduced by
Viper (under 5000 per year) made this an ideal
manufacturers of low volume specialty vehi-
application for composites. The selection of ~ cles before they appear in high volume car
RTM process with low cost molds resulted m
lines produced by larger companies such as
greater investment savings than could be
Chrysler, Ford or General Motors. These lar.ger
achieved with SMC and RRIM. This milestone
companies often will evaluate a new apphca-
application was introduced in the early stages
tion on a low-volume specialty vehicle or on a
of an industry-wide effort to develop the RTM
customized segment of a high-volume vehicle
and SRIM (structural reaction injection mold-
before committing to high-volume production
ing) processes for the production of structural
with a composite part.
and Class A parts for the transportation market.

F·g 413 GM's all purpose vehicle (APV) is the automotive industry's largest application of ~omposites.
E~ch v~hicle uses approXimately 120 kg (260 lb) of SMC in addition to other polymer compOSItes. (Photo
courtesy of General Motors Corporation.)
 Current applications 909

Fig. 41.4 The Dodge Viper is the first American car to have all Class A composite body panels produced
by the resin transfer molding (RIM) process. (Photo courtesy of Chrysler Corporation.)

Automotive body applications Newly developed composites made from com-

pression molded glass fiber mat thermoplastic
Composites have proven to be very successful sheet (GMT) or liquid molded (RTM/SRIM)
in a wide range of exterior body panels and thermoset resins may also compete with SMC
are used in hundreds of vehicle applications. for horizontal body panels.
Excellent surface finish, light weight, and a The broader use of composites in body pan-
thermal coefficient of expansion near that of els is restricted by their limited cost-
steel have made these applications successful. effectiveness at very high production volumes.
Customers appreciate the dent and corrosion Composite panels are well-suited for niche
resistance of composite panels. vehicles and for providing economical oppor-
SMC dominates composite applications in tunities to achieve styling differences within a
horizontal panels such as hoods, roof and deck body line, but they are seldom employed in
lids, and competes with RRIM for vertical pan- high-volume applications. Another factor that
els such as fenders, doors and quarter panels. precludes the use of some polymer composites
The use of unreinforced thermoplastic body in body panels is sensitivity to high tempera-
panels on vehicles such as GM's Saturn and ture. In most assembly operations all body
Chrysler's Intrepid, Concorde, and Vision panels are subjected to a heat soak at
demonstrate the potential for the displacement 175-205°C (350-400 0 P) for about 30 min fol-
of composites in some body applications. lowing the electrodeposition of a corrosion
910 Land transportation applications

coating ('E-Coat' or 'ELPO') to the steel body and extensive testing, a transverse rear leaf
structure. Only a limited number of low-cost spring was introduced on the 1981 Corvette.
polymer composites can resist dimensional This 3.6 kg (8Ib) filament wound E-glass/ epoxy
change or surface distortion during this heat monoleaf spring replaced a ten-leaf steel spring
treatment. weighing 19 kg (42Ib). A front transverse
While composites are used in the body spring was added to the Corvette in 1984 and
structure for several low-volume sports cars, the first composite longitudinal spring
this opportunity remains essentially untapped appeared on the General Motors Astro van in
in the auto industry. The Lotus Esprit has an 1985. The use of composite springs rapidly
all-composite body produced by the VARI spread across GM car lines in the late 1980s and
process (vacuum assisted resin injection) that early 1990s. In less than a decade from its intro-
is mounted on a steel frame. Lotus uses a sim- duction, the Inland Division of GM (now
ilarly produced composite floor pan and body Delphi Chassis Systems) was manufacturing
panels in the Elan. Consulier manufactures more than one million Liteflex® composite
sports cars with all-composite mono co que springs per year (Fig. 41.5) to meet the demand.
bodies produced by vacuum bag molding of
epoxy prepregs with E-glass, S-glass, carbon
and Kevlar® fibers. The entire body structure
weighs only 125 kg (275 lb) and takes all struc-
tural loads. Higher volume applications in
body structure includes the Corvette which
has used composites in the floor pan, rear
underbody and radiator support.
Since the early 1980s an evolving series of
composite materials and processes have been
used to produce front and rear bumper beams.
SMC containing random chopped and/ or con-
tinuous unidirectional fibers was initially
used. Subsequently, stamped thermoplastic Fig. 41.5 Composite leaf springs reduce vehicle
composites and SRIM bumper beams have weight, resist corrosion and outlast traditional steel
been commercialized. Special design consider- leaf springs. (Photo courtesy of Delphi Chassis
Systems.)
ations and parts consolidation have lead to the
use of glass-reinforced polypropylene
(Azdel®) in the front bumper beams of
Chrysler minivans, which are produced in vol- The exceptional weight savings and outstand-
umes of more than 500 000 units per year. ing durability have been the keystones to the
success of composite springs in cars and more
recently, in heavy trucks. Composite springs
Automotive chassis applications
last at least five times longer than steel coun-
While relatively few composite chassis compo- terparts in laboratory fatigue tests and in field
nents have been commercialized, notable service. Added advantages are improved
applications have provided evidence of the packaging due to smaller size, and improved
performance level and durability offered by ride and handling characteristics. Their long-
composites. term field success has also dearly
One of the most successful automotive struc- demonstrated the survivability of polymer
tural applications to date is the composite composite materials in the harsh under-vehi-
leafspring. After decades of development work cle environment.
 Current applications 911

Another suspension application demon- by the 'lost-core' process in which a low tem-
strating the durability and load carrying perature melting alloy such as tin-bismuth is
capability of composites is wheels. The com- used as a form to mold the hollow sections.
mercial introduction of polymer composite The core is then melted away from the cured
wheels was on the 1989 Shelby CSX modified part in a subsequent operation. An alternative
version of the Dodge Shadow. Developed by is to 'weld' two thermoplastic pieces to form
the Motor Wheel Corporation, the Fiberide® the manifold. To reduce cost, some fuel
wheel uses a combination of structural SMC intakes are being integrated with the air
and XMC. These materials provided a mix- intake manifold.
ture of chopped random and oriented Injection molded phenolic compounds are
continuous E-glass in a vinyl ester resin. At used in many engine and transmission appli-
reduced weight, the composite wheel out- cations because of dimensional stability and
performed both steel and aluminum wheels creep resistance at higher temperatures.
in fatigue. The development of polymer com- Current applications include pulleys, torque
posites that retained acceptable lug nut converter reactors, thrust washers, water out-
torque after wheel heating caused by extreme lets, valve covers, radiator end caps, motor
braking conditions was a technical advance commutators and fuel rails (Fig. 41.6).
that made this application possible. While the performance advantages of com-
A front suspension stabilizer link was posite drive shafts and propeller shafts (lower
introduced on the 1994 Ford Taurus and weight, better NVH, increased durability) has
Mercury Sable. Made from a glass fiber rein- been established, cost remains a barrier to
forced copolymer polyacetal, this was the their wide spread use. A composite drive shaft
first use of a structural thermoplastic com- has an economic advantage, however, when it
posites in such an application in North replaces two-piece metal shafts. The most suc-
America. This durable part does not require cessful designs have used glass and/ or carbon
painting and reportedly provides a 42% fiber composite overwrapped on an alu-
weight reduction and 33% cost reduction minum tube using pultrusion or filament
over the replaced steel part. winding processes.

Automotive powertrain applications

The application of composites in the power-
train is extensive and is growing. In the past,
only metals were considered for use in the
demanding environment of the engine and
transmission. Increasingly, composites are
being selected to reduce weight and cost, and
to improve NVH (noise, vibration and harsh-
ness), engine efficiency, packaging and
corrosion resistance.
Air intake manifolds are rapidly being con-
verted from metal to injection molded
thermoplastics reinforced with short glass
Fig. 41.6 Powertrain applications of phenolic com-
fibers. Introduced on European cars in the posites include transmission torque converter
1970s, composite manifolds are predicted to reactors and thrust washers, poly-V pulleys, and
be used on most vehicles by the turn of the water outlets. (Photo courtesy of Rogers
century. Most intake manifolds are produced Corporation. )
912 Land transportation applications

Experimental engines made almost exclu- 41.4.2 TRUCK APPLICATIONS

sively of polymer composite materi~ls, Composites are widely used in the medium-
including block, head, pistons and connecting
duty and heavy-duty truck industry to
rods, have been tested, but commercialization decrease weight, reduce manufacturing and
is unlikely in the near future. Many metal
maintenance costs, and extend the longevity of
matrix composite (MMC) parts have also been
the vehicles. Lighter trucks allow increased
prototyped for engine applica~ons, howe,:,er, payload capability and improved fuel econ-
few have gone into commerClal production
omy. Since trucks are generally not produced
(Table 41.1). Aluminum reinforced with silicon
in high volumes, the lower tooling inv~s~e~t
carbide, alumina or carbon fibers offers the
favors composite parts. The durabIlIty IS
potential for good mechanica! and the~mal important since trucks are used for many
properties with large. weIght savIn~s. years.
However, factors such as high cost, low ductil-
Composites are commercially used for
ity and machining difficulties have retarded essentially all skin surfaces on truck cabs, air
the commercialization MMCs in transporta-
deflectors, hoods, fenders, roofs, side closure
tion components.
panels, sleeper box and doors. Mack Tru~k
introduced the first structural SMC door In
Automotive interior applications 1978 and two years later GM Truck and Coach
made the first use of continuous fiber SMC in
Although manufacturers make extensive use
the door of the Astro and General truck cabs.
of polymeric materials on the interiors of cars In 1983 Mack Truck became the first manufac-
and trucks, most do not require fiber rein-
turer to introduce a cab with composites used
forcement to meet performance requirements.
in all exterior surfaces. These panels covered a
Some notable exceptions are seats, load floors, structural frame made of steel, aluminum, and
knee bolsters and instrument panel supports. structural composite beams. Several all-com-
Thermoplastic composite seat supports
posite truck cabs using monocoque structures
were first introduced on the 1975 Corvette. In have been prototyped and shown to meet all
addition to stamped polypropylene, ther- performance requirements.
mosetting polyurethanes and polyeste~ .are Weight savings of 20% are achieved in truck
used in both seat back and seat pans. An InJec- trailers when floor, wall and ceiling panels are
tion molded nylon seat with long fiber
composite. In addition, refrigerated trailers
reinforcement premiered on the 1993 Dodge
made from composites are 25-30% more ther-
Viper. Stamped thermoplastic sheet is cur-
mally efficient than metal trailers. In a recent
rently preferred for load floors in station application, Trail King has introduced a flat
wagons and extended pick-up truck cabs.
bed trailer with a fiber reinforced epoxy deck
Stamped polypropylene has been used in to achieve lower weight, corrosion resistance
numerous knee bolsters. This structural part at
and lower cost.
the bottom of the instrument panel helps to
One of the fastest growing application in
control occupant movement during a frontal
heavy truck is composite leaf springs for la~ge
crash. Composites and unreinforced thermo-
weight savings and greatly extended sprIng
plastics are also being used in instrument life.
panel supports to provide lower ~ost through
parts integration and reduced w.eIg~t. ~e u~e
of composites in the front of vehicle Intenors IS 41.4.3 RAILROAD APPLICATIONS
expected to expand.
Durability and light weight make composites
attractive for railroad applications, yet few
 Current applications 913

Table 41.1 Composite applications in automobiles and trucks!

Experimental Limited production Well established

Body Front rails Radiator support Grille opening panel
A, Band C pillars Floor pan Grille opening
Roof frame Cowl panel reinforcement
Rockers Wheel housing Rear end panel
Rear underbody Windshield surround Hood
Pickup truck box Door surround Deck lid
Truck cab structure Center tunnel Lift gate
Truck trailer bed Tanker truck trailers Rear hatch
EV battery trat Roof panel
Hardtop cover
Fenders
Quarter panels
Doors outers
Door inners
Spoiler
Headlamp cover
Fuel filler doors
Front bumper beams
Rear bumper beams
Truck air defector
Truck trailer walls
Truck trailer roof
Chassis Front cross member Wheels Leaf spring
Transmission support Stabilizer bar links Disk brake pistons
Brake rotors!
Powertrain Throttle body Drive shaft Air intake manifold
Oil pump Propeller shafts Battery tray
Transmission valve body Water pump housing Fan shroud
Hydraulic clutch actuator Water pump impeller Radiator end caps
Engine - Pistons, block, Fuel pump components Transmission torque
head, piston pins3 Fuel rail converter reactor I stator
connecting rods 3 Fuel tank supports Transmission thrust washer
rocker arms2 Camshaft sprockets Motor commutators
bed plate2 Oil pan Diesel electronic unit injector
Flywheel4 Differential cover Valve cover
Drive shaft2 Valve lifter guides Timing chain covers
CNG storage cylinders5 Poly-V pulleys
Cylinder liners2 Idler pulley
Pistons2 Distributor cover
Interior Car jack Knee bolster Seat back
Steering wheel Instrument panel support Seat support
Window frame/trim Load floor
Plenum
Glove box

1 Polymer composite unless noted otherwise 4 Electric powered vehicles

2 Metal matrix composites 5 Natural gas vehicles
3 Metal matrix or polymer matrix composites
914 Land transportation applications

have been commercialized. Pultruded glass (300 miles per hour), may offer an opportunity
fiber reinforced thermoplastic composite pan- for composites because of the need to mini-
els are used to construct a 'Secured Modular mize weight. Weight reduction is especially
Automotive Rail Transport' (SMART) for important for magnetic levitation (MAGLEV)
Union Pacific Railroad. Each three-tier struc- systems where the vehicle is suspended above
ture serves as a 'car rack' to protect eighteen the guideway to provide friction free move-
automobiles from damage and theft during ment.
transport to the dealer. The use of composites
also significantly reduces the maintenance cost
41.4.5 MILITARY APPLICATIONS
of the car rack.
Composite hopper car covers protect grain Limited applications of composites are found
and other dry materials that need protection in combat and non-combat ground vehicles in
from moisture. The covers, which range from the US military, but if current development
9-15 m (30-40 ft) in length, are fabricated by a programs are successful, much greater use of
hand lay-up process. Two prototype composites should result. Modem warfare
glass/polyester filament wound railroad cars requires rapidly deployable, survivable vehi-
termed the 'Glasshopper' have been in service cle systems. Thus, weight reduction in all
since 1981 without failure. Produced in a joint vehicles, including armored vehicles, is desir-
venture by ACF Industries, Cargill and able. In addition to weight savings,
Southern Pacific, the corrosion resistant, light- composites can potentially offer increased
weight composite car can carry a greater durability, improved signature management,
payload, but is significantly more expensive better personnel protection and lower produc-
than a steel hopper car. tion cost.
The goal of an ongoing Army project, the
Composite Armored Vehicle Advanced
41.4.4 MASS TRANSIT APPLICATIONS
Technology Demonstrator (CAV AID), is to
Buses and passenger rail systems offer many establish the feasibility of using polymer com-
more opportunities for the application of com- posites in the primary structure of a 20 tonne
posites than are presently in service. Some (22 ton) combat vehicle to achieve a 33%
cosmetic and semi-structural applications weight saving over a traditional metal vehicle.
have been successfully implemented, but few Composites are also being considered for com-
examples of structural components are found. bat vehicle armor - used either alone or in
The durability of polymer composites has conjunction with ceramics and/or metals - to
lead to extensive use in seats for buses, sub- provide significant weight savings over cur-
ways, people movers and trains. Composite rent materials. The High Mobility
sandwich panels with glass fiber-phenolic Multi-purpose Wheeled Vehicle (HMMWV),
skins over aramid or aluminum honeycomb which is currently in production, utilizes an
are used in walls, ceilings and floors of many integrated hood/fender assembly molded
European mass transit cars. These rigid panels from SMC. In upgraded versions, composite
are very lightweight and the use of phenolic armor is attached to the HMMWV.
resins allows attainment of fire/ emission stan- Phenolic spall liners containing Kevlar or
dards. The end caps of transit cars are often 5-2 Glass® fibers are used in the M113A3
molded composites. Pultruded exterior panels Armored Personnel Carrier and Bradley
are being substituted for aluminum on buses Fighting Vehicle to provide troop protection.
to reduce weight and decrease assembly cost. Longer service life and chemical protection
The development of high speed rail sys- were the motivation for the incorporation of
tems, which attain speeds up to 480 km/h composite seats and side racks in 2.2 tonne
 Conclusions 915

(2.5 ton) and 4.4 tonne (5 ton) trucks. Other and technical expertise acquired in defense
potential composite applications such as M1 programs to the development of composite
Abrams Main Battle Tank components (dri- technology needed to improve the competi-
ver's seat, air intake plenum, stowage box, tiveness of American industries with a
and power pack container), tactical vehicle particular emphasis on the automotive indus-
body, fuel tanker and HMMWV drive shaft try. Numerous federal programs are directing
are under consideration for production after dollars and technical expertise resident in the'
the tum of the century. National Laboratories into ground transporta-
tion programs. A notable example is the
'Partnership for a New Generation of Vehicles'
41.5 FUTURE DIRECTIONS
program, which teams the government with
The growth of the transportation market is the automotive industry to produce the tech-
expected to continue, and potentially acceler- nology to make safe, comfortable, and
ate, into the 21st century. As composites affordable cars that achieve up to 30 km/l (82
become better understood by designers, and miles per gallon) of gasoline with low emis-
as the reliability and advantages of these mate- sions. Composite materials are expected to
rials are more dearly demonstrated, they will play a key role in meeting this challenging
be used increasingly in more demanding goal. Successful implementation of the tech-
structural applications. Body structure, chassis nologies developed in these cooperative
and powertrain offer tremendous opportuni- programs could have an enormous impact on
ties for the utilization of composite materials. the usage of composites in future vehicles.
Improved manufacturing capability to rapidly
produce composite parts will boost their eco-
41.6 CONCLUSIONS
nomic viability in high volume applications.
The development of a commercial infrastruc- Ground transportation is the largest and one
ture and market for the large scale recycling of the fastest growing segments of the polymer
and reuse of polymeric materials will also composites market. While transportation use
increase the growth opportunity for compos- of composites is expected to expand, the rate
ites. of growth depends on a number of factors.
The potential economic advantages and Improvements in technology are needed to
weight reductions afforded by the u~ of com- increase high volume production capability
posites in integrated body structdre and and increase the cost competitiveness of com-
chassis components has driven development posites relative to other materials. The
programs since the 1980s. In 1988 Chrysler, performance and durability of composites in a
Ford and General Motors formed the wide range of structural applications must be
Automotive Composites Consortium (ACC) to demonstrated. More designers and engineers
conduct joint R&D on structural polymer com- need to become familiar with the unique char-
posites with a focus on these structural acteristics of composites and learn to develop
applications. Operating under the United States designs that use the full potential of these
Council on Automotive Research (USCAR), the materials. An economically viable infrastruc-
ACC works with supplier companies and uni- ture for dealing with post-consumer waste
versities to develop the processing, materials, must be established. With these advances,
design and joining technology needed to composites are expected to play an important
achieve production worthiness and cost effec- role as industry meets the increasing world-
tiveness of composite structures. Currently, the wide demand for safe, clean, energy efficient
US Federal Government is providing resources land transportation.
MARINE APPLICATIONS 42
Wayne C. Tucker and Thomas Juska

42.1 INTRODUCTION resins, but their superior resistance to hydrol-

Use of composites in marine applications is ysis and blistering Gacquemet and LaGrange,
widespread. The two major advantages of fiber 1988) makes isophthalic polyester a better
reinforced plastics over metals are resistance to choice for applications requiring long term
the marine environment, particularly the elim- exposure to water.
Use of vinyl esters is becoming more com-
ination of galvanic corrosion and the ease of
mon. Although they are more expensive than
tailoring structures, which are fabricated by
polyesters, vinyl ester laminates generally
molding processes. In addition, composites
have better properties. In addition, vinyl esters
have high strength-to-weight ratios.
This chapter is an overview of the materials have excellent resistance to matrix hydrolysis.
and fabrication processes used in marine appli- Also available are mixtures of vinyl ester and
polyester, which offer some of the benefits of
cations of composites. More comprehensive
studies of the use of composites in marine con- vinyl ester but at intermediate cost.
struction have recently been published (Smith, Epoxy laminating resins, rarely used in
1990; Greene, 1990; Davies and Lemoine, 1992). marine applications due to cost and the need
for elevated temperature cure, are used by
some boat builders in high performance one-
42.2 MATERIALS off racing craft and some production boats.
Among the advantages of epoxy are extended
42.2.1 RESINS out-time compared to room temperature cur-
Resins used in marine applications generally ing resins and low volatile organic content.
cure at room temperature, both for the low The ~aterial form is wet epoxy prepreg
(sometimes referred to as a wet-preg to distin-
fabrication costs and because elevated temper-
guish it from conventional prepreg) made by
ature performance is not required.
the boat builder (Gougeon, 1992). Fabrication
General purpose polyester is the most com-
mon laminating resin. These materials, based is normally by vacuum bag, room temperature
cure followed by oven post-cure at about 50°C
on orthophthalic acid (phthalic anhydride),
(122°F).
are the least expensive, but long term immer-
sion without a barrier coat will probably result
in blistering (Burrell, Herzog and McCabe, 42.2.2 REINFORCEMENTS
1987). Polyesters based on isophthalic acid are
slightly more expensive than general purpose E-glass fabric is the primary reinforcement in
marine construction, of which there are numer-
ous forms. There are two basic styles: woven
Handbook of Composites. Edited by S.T. Peters. Published and knitted. Woven fabrics are further subdi-
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 vided into woven roving and woven yam.
 Materials 917

Woven roving are the most common fabrics Nguyen, 1992). Their use is limited because of
due to their low cost. They are available in a cost and also because they cannot bear much
variety of weights and weave patterns, but a load in compression.
800 g/m2 (24 oz/yd2) plain weave is probably
the most frequently used. In small boat con-
42.2.3 CORES
struction, the fabric is usually mat-backed, in
which a chopped strand mat has been stitched Linear PVC foam is used extensively in boat
or powder bound to one side of the fabric. hulls, at a density of 80kg/m3 (5Ib/ft3). The
Woven yarn, also called textile fabric or resilience of the material allows hull deforma-
cloth, is more expensive and therefore used tion during impact with no skin/ core
infrequently. These materials are finer in tex- debonding, skin damage or core shear failures.
ture than woven rovings and are used as Crosslinked PVC foam is used in decks, bulk-
surfacing plys, particularly in tooling. heads and cabins because it has better
Knitted glass fabrics are becoming more structural properties and will not creep at the
common. The rovings are stitched together, higher topside service temperatures. It is
which keeps the strands relatively parallel (no available in a wide range of densities, but is
crimp as in weaves) and stabilizes the fabric usually used at 80 and 100 kg/ m3 (288 and
against distortion during handling. The orien- 360Ib/yd 3). Both types of PVC foam are
tations of the separate layers and the number closed cell, absorb very little water and are
of separate layers, which compose knitted fab- moderately priced.
rics can be varied, although a 2-ply 0/90 is End grain balsa, also widely used as the
most common. Knitted fabrics are currently core in boat sandwich structures, is inexpen-
available up to 2500 g/m2 (72 oz/yd2). sive and has excellent structural properties.
Chopped strand mat is most commonly Under some circumstances, however, balsa
used where a resin-rich layer is desired, such as can rot. If skin/ core debonding occurs and
between the core and fabric in sandwich struc- water is allowed to penetrate this interface, the
tures. Usually the first ply next to the gel coat wood will eventually degrade. It has been
is a layer of mat, which reduces print-through shown, however, that balsa will provide a long
of the reinforcement as well as providing addi- marine service life with proper maintenance
tional protection from water permeation. (Baltek Corporation, 1986).
5-2 glass is about six times the cost of E- Composite core is also available. The mate-
glass, so it is rarely used in marine structural rial is made in two forms, either sinusoidal
applications. 5-2 made in the G filament is nested-wave, or in a closed cell configuration
sized for epoxy and when made with the K where planar webs alternate with sinusoidal
filament is sized for polyesters and vinyl webs (Plunkett et al., 1992). This type of core is
esters. essentially a collection of sine wave stiffeners
Carbon fiber is also rarely used because of fused to skins with a laminating resin. It
cost. America's Cup racing sailboats use car- appears to have great potential for marine and
bon throughout the boat and some racing civil engineering structures and has in fact
powerboats use carbon as well. Carbon has an been used to make about 200 prototypes,
additional problem in that it can corrode met- including deckhouses, storage tanks, build-
als in contact with it in the presence of water ings and parabolic solar energy concentrators.
(Tucker, Brown and Russell, 1990; Aylor and
Murray, 1992).
42.2.4 SKIN/CORE BONDING
Polymeric fibers, namely Kevlar and
Spectra, are sometimes used in marine appli- A good bond between skin and core is critical
cations where weight is critical (Sloan and to the performance of sandwich structures. In
918 Marine applications

boat building it is general practice to use

vacuum bag pressure to achieve the necessary
quality. There are two basic methods: one is to
use a paste adhesive between skin and core
and the other is to use the laminating resin as
the adhesive.
Paste adhesives for a polyester or vinyl ester
laminate are filled polyesters. The adhesive is
catalyzed and troweled onto the surface of pre-
cured laminate (nonnally the outer skin) and
the foam core is sprayed with catalyzed resin
Gohannsen, 1990). The core is then bedded into
the paste, the vacuum bag is positioned and
the assembly is allowed to cure at room tem-
perature. The core either has holes to allow air Fig. 42.1 Boat hull blisters. (Courtesy of Dr. Tom
to escape, or contour core is used. In both cases, Rockett, University of Rhode Island.)
observation of the paste seeping through the
spaces is evidence for a successful bonding
operation. The inner skin is laminated directly
3. More water is drawn into the laminate
onto the surface of the core, with the first ply
through osmosis, creating a localized pres-
nonnally a chopped strand mat.
sure. The solution in the osmotic center
Some sandwich structures are made by
hydrolyzes the resin, resulting in an
bonding the core with a resin rich layer. In this
increase in water-soluble materials, which
method, a wet-out ply of chopped strand mat
then draw more water into the laminate
is applied to the outer skin and used as a bed-
and increase the pressure.
ding for the core. In a recently discovered
variation of this approach, which appears to There are materials solutions to the blistering
offer some advantages, a ply of polyester mat problem of orthophthalic polyester laminates.
is used to bed the core (Lewit, 1990). One is to coat the hull below the waterline
with an epoxy paint formulated to reduce
water penetration (Interlux, 1993). Another is
42.3 ENVIRONMENTAL EFFECTS to use a permeation barrier between gekoat
and laminate, as mentioned previously. A
42.3.1 BUSTERING more expensive solution to blistering is to
change to a vinyl ester laminating resin. Use of
Blistering can occur in gel coated marine lam-
isophthalic polyesters is another alternative,
inates (Fig. 42.1). The causes are complex and
which offers a compromise between cost and
blisters may occur due to the selection of mate-
resistance to hydrolysis.
rials or due to the process used in fabrication
Blistering of carbon fiber reinforced plastics
or both. The fonnation of blisters basically
(CFRP) in seawater can occur for another rea-
involves three steps (Marino, Rockett and
son. When galvanically coupled to a metal
Rose, 1985):
fitting, the CFRP becomes cathodic. The result-
1. Water penneates into the laminate. ing electro-chemical reaction produces
2. Water-soluble components in the resin are hydroxyl ions near the carbon fibers. To bal-
dissolved in this water and are concen- ance the charge, sodium ions are drawn in.
trated in cracks or voids in the laminate, The eventual result is blisters with high alka-
creating an osmotic center. linity (Tucker, 1991; Miriyala et al., 1993).
 Fabrication 919

42.3.2 MICROBIAL DEGRADATION production rates. These machines perform

three important functions: they wet out the
All engineering materials become colonized
fabric, meter the resin and place the wet out
by microorganisms within hours after expo-
fabric via an overhead crane or gantry
sure in natural waters (Little and Wagner,
(Venus-Gusmer, 1989). Laminators must still
1986). Microorganisms grow and produce a
roll out the fabric after it is placed to remove
viscoelastic layer or biofilm. The environment
air. Resin contents are reduced to about 50
at the interface is radically different from the
wt%.
bulk medium in terms of pH, dissolved oxy-
Portable impregnators are also occasionally
gen and organic and inorganic species (Little et
used. Some high performance racing sailboats
al., 1991). In addition to biofouling, the mater-
are laminated with wet epoxy prepreg made
ial properties may be adversely affected.
by the fabricator.
Epoxy and nylon coating on steel can be
breached by mixed cultures of marine bacteria
Gones, Walsh and Mansfield, 1991). Sulfate 42.4.3 RESIN TRANSFER MOLDING
reducing bacteria (SRB) degrade marine
Vacuum assisted resin transfer molding
caulks and polymeric coatings Gones, Walsh
(VARTM) has been shown to have great poten-
and Mansfield 1991; Jones et al., 1992a,b).
tial for the fabrication of ship hulls and
Organic surfactants on S-2 glass fibers have
secondary structures. The Naval Surface
reportedly been attacked in the presence of
Warfare Center, Carderock Division, has had
SRB as well (Tucker et al., 1993).
several prototype structures made with a
Microorganisms can be a threat to material
VARTM method known as SCRIMP (Seemann,
performance in the marine environment.
1989), including deckhouse modules, masts
and a Swimmer Delivery Vehicle (NSWC
42.4 FABRICATION Carderock Division, 1993). The advantages of
VARTM over hand lay-up are that very low
42.4.1 HAND LAY-UP resin contents are achieved (30-35 wt %), the
p¢cess is less labor intensive and styrene
Hand lay-up is the most common fabrication emissions are almost eliminated. Laminate
method in marine construction. Catalyzed mechanical properties of panels made by
resin is first applied to the gel coated mold VARTM show significant strength and stiff-
surface, usually by spraying. A layer of dry ness improvements over panels made by hand
mat or fabric is positioned onto the resin and lay-up Guska, Mayes and Seemann, 1993).
by pressure of hand-held rollers, resin gradu-
ally soaks through the fabric and most of the
trapped air is forced out. This process is 42.4.4 VACUUM BAG PROCESSING
repeated until the desired number of plys is Vacuum bag processing is used extensively in
attained. The full thickness of the structure is the boat building industry to facilitate bonding
rarely built up all at once, but is usually done of the core onto pre-cured skins, as described
in several lay-up / cure stages. Resin contents earlier. Vacuum bags are rarely used to compact
are normally about 55 wt % and void contents polyester or vinyl ester wet lay-ups because
about 1%. there is little time to place the bag before gela-
tion. Vacuum bag consolidation is used,
however, to laminate wet epoxy resins and
42.4.2 IMPREGNATORS
oven-cured prepregs, but the practice is limited
Impregnators are sometimes used in the fabri- to the production of high performance boats
cation of large composite structures to increase such as the America's Cup racing sailboat.
920 Marine applications

Autoclaves are rarely used in marine con- and can be used as a guide to the properties of
struction, mainly because the process is too typical marine laminates (made by hand lay-
expensive for marine structures. There are a up) with those attainable by alternative
few exceptions, however. Submarine sonar fabrication methods.
bow domes are autoclave cured, as are car- The properties of cored panels are becom-
bon/ epoxy sailboat masts. ing increasingly important to the design of
marine structures, particularly boats. Recent
studies have indicated that cored panel prop-
42.4.5 FILAMENT WINDING
erties as determined with standard beam
Filament winding is very often used to make bending tests should be supplemented by
structures and machinery for marine applica- measuring the response of plate specimens
tions. It is inexpensive, largely automated and subjected to uniform pressure loading (Huss,
low resin contents are achieved. It is most 1990; Reichard, 1992; Gougeon and Bertelsen,
appropriate for parts with a circular cross sec- 1993).
tion, such as shafting, piping and pressure
bottles.
42.6 STRUCTURES

42.5 MECHANICAL PROPERTIES 42.6.1 MINE COUNTER MEASURE VESSELS

Materials with a wide variety of fibers, resins, Mine counter measure vessels are among the
resin weight contents and void contents are largest GRP boats fabricated. These military
used in marine construction. Extensive testing vessels are made of non-magnetic material
at NSWC, Carderock Division, has led to a (GRP or wood) to reduce the threat from mag-
database of marine construction materials netic mines. Several foreign Navies have built
made by hand lay-up, VARTM and vacuum composite minehunters, including the Soviet,
bag consolidated prepregs Guska, Mayes and British, Swedish, Italian, French and
Seemann, 1993; Juska 1993). The values in Australian, usually 45-48 m (150-160 ft) long
Table 42.1 were abstracted from this database (Greene, 1990).

Table 42.1 Properties of marine laminates

Fabric Woven Woven Style 7781

resin roving roving epoxy
vinyl ester vinyl ester
Fabrication Hand lay-up VARTM Prepreg
Resin content (wt %) 50-60 30-35 35-40
Void content (%) 0.5-1.5 0-0.5 0.5-4
Compression strength, MPa 172-207 379-414 379-414
(psi X 103) (25-30) (55-60) (55-60)
Tensile strength, MPa 172-207 379-414 379-414
(psi X 103) (25-30) (55-60) (55-60)
Tensile modulus, GPa 11.7-13.8 21.4-22.8 21.4-22.8
(psi X 106) (1.7-2.0) (3.1-3.3) (3.1-3.3)
Flexural strength, MPa 276-345 483-552 483-552
(psi X 103) (40-50) (70-80) (70-80)
 Structures 921

The US Navy, whose Mine Counter 42.6.3 LARGE POWER YACHTS

Measure (MCM) ships have a wooden hull Large custom private power yachts, typically
construction, is currently building the MHC- 21-39 m (70-130 ft) long but up to 48 m
51 class coastal minehunter with GRP (160 ft), are made with GRP by several boat
(Hepburn, Magliulo and Wright 1991). The builders. Most of the builders make hulls
Navy selected to modify the Italian Lerici class cored with linear PVC, with decks, bulkheads
design for the MHC-51, which has an unstiff- and superstructures cored with crosslinked
ened single skin hull. The resin is an PVC. The cores are bedded in paste adhesive
isophthalic polyester formulated to have a and vacuum bagged to the outer skin. Blister
long gel time (about four hours). The rein- protection is usually provided by several lay-
forcement for the hull, decks and transverse ers of epoxy paint applied to the gel coat after
bulkheads is DF1400 (1400 g/m2; 40 oz/yd2) the boat is demolded.
composed of spun roving in the fill direction. Westport Shipyard developed the concept
The superstructure is constructed of Rovimat of using the same (female) mold to make sev-
1200, an 800 g/m2 (24 oz/yd2) woven roving eral non-identical boats. They have pulled 35
stitched to a 400 g/m2 (120z/yd2) mat. The hulls from a single mold, which they alter in
reinforcement is wet-out and placed with an length and width to accommodate the design.
impregnator. Westport uses an overhead crane impregnator,
which they believe saves 50% in labor over
42.6.2 SMALL BOATS hand lay-up. They laminate with a general
purpose polyester resin and a 800 g/m2
Small boats and craft are usually constructed (24 oz/yd2) woven roving (not mat backed).
from GRP and probably dominate the marine Other builders have also adopted the vari-
composites market. There are numerous styles able mold approach. The Christensen Motor
of power and sail boats. In spite of the diver- Yacht Company makes boats by hand lay-up
sity in hull shape and size there are fairly up to 43 m (140 ft) long. They prefer knitted
standard fabrication techniques and materials reinforcement (2408) for its handling charac-
used in boat building. A comprehensive teristics and more uniform resin content and
review of the methods and materials used in use vinyl ester in the outer skin. North Coast
Navy craft fabrication and repair has recently Boats has a mold capable of producing boats
been published (Russell, 1992). up to 40 m (131 ft) long. They make boats by
Female molds are more common than male hand lay-up with a relatively heavy knitted
plugs in production boats due to the impor- reinforcement (3205) and a vinyl ester/poly-
tance of unvarying hull shape and ester resin mixture. The Nordlund Boat
appearance. After the mold surface is coated Company laminates with a 600 g/m2
with release agent, it is sprayed with several (18 oz/yd2) knitted reinforcement for the outer
layers of gel coat (usually an isophthalic acid, skin, for reduced print through and woven
neopentyl glycol polyester) until a roving in the inner skins.
0.5-0.76 mm (20-30 mil) wet thickness is Some builders of large yachts do not use a
achieved. The gel coat is allowed to harden variable mold. Admiral Marine specializes in
prior to lamination. The first ply of the lami- one-off female tooling. They used this concept
nate is usually a chopped strand mat to to fabricate a 48 m (160 ft) private motor yacht
reduce print-through of the subsequent layers (by hand lay-up). An E-glass/Kevlar hybrid
of woven or knitted reinforcement. Boats are fabric was used in the decks and cabin.
usually fabricated by hand lay-up, as Jones-Goodell Yachts also make a unique
described earlier. female tool for each boat, producing boats up
to 33 m (110 ft) long.
922 Marine applications

Delta Marine produces displacement, semi- typical of marine construction, however; since
displacement and planing hulls using a the designs minimize weight and maximize
variety of techniques. The displacement hulls stiffness of structural components, the yachts
are built using a stiffened single skin, where use aerospace materials. For example, the
the semi-displacement hulls in the larger sizes boats built for America3 Foundation's success-
utilize a balsa cored sandwich construction in ful defense of the 1992 Americas Cup had
the sides and a stiffened skin bottom. The pri- carbon/epoxy tape in the hull (cored with
mary materials used are a mat backed knitted aluminum honeycomb), frames, keel, mast,
fabric (3205) and a combination of vinyl ester pole and boom, carbon winch drums and
and general purpose polyester laminating shafting and pultruded Kevlar standing rig-
resin. In addition to private motor yachts, ging (Kramers, 1993). The design philosophy
Delta builds commercial fishing vessels and was to keep the weight as low as the lACC
small charter boats. rules allowed, while maximizing stiffness and
being just strong enough to avoid failure. The
processing conditions for the hull are limited
42.6.4 AMERICA'S CUP RACING SAILBOATS
to 1 atmosphere pressure (i.e. vacuum bag) at
The America's Cup class yachts use compos- 95°C (203°F), while those for the mast are 3 x
ites more extensively than any other marine 105 N/m2 (3 atm) and 135°C (275°P).
structure. The materials in the boats are not

RECOVERY STROBE MAtN THRUSTERS

LIGHT

RFBEACON

TITANIUM HEMISPHERE
ACOUSTIC LINK
TRANSDUCER RETRACTABLE LIFTING
EYES

GRAPHITE COMPOSITE
PRESSURE HULL

DESCENT
WEIGHT
RELEASE

SIDE·LOOKING
SONAR
AFT VERTICAL
THRUSTER

Fig. 42.2 Advanced Unmanned Search System. (Courtesy of Naval Command and Ocean Surveillance
Center, RDT & E Division.)
 Structures 923

42.6.5 PRESSURE HULLS 42.6.7 FAIRINGS

The potential for composites in pressure hulls The submarine fairwater, or sail, (Fig. 42.3) is
has been demonstrated (Garvey, 1990; Stachiw an excellent example of a large complex com-
and Frame, 1988). Actual use of composites in posite marine structure. A typical fairwater is
these structures has been limited, however. 12 m (40 ft) long, 73 m (24 ft) tall and com-
One exception is the Advanced Unmanned posed of over 1000 component sections. These
Search System (AUSS), developed and oper- massive structures were retrofitted on numer-
ated by the Naval Command, Control and ous classes of submarines replacing steel
Ocean Surveillance Center, RDT&E Division fairwaters. The composite fairwaters provide
(NRad, 1992). It consists of a vehicle, a control the Navy with lighter weight and non-corro-
van, a maintenance van, a launch and recovery sive properties that steel could not match. The
ramp and tow fish for acoustic communica- construction of submarine fairwaters consists
tions. As shown in Fig. 42.2, the vehicle has a of E-glass/polyester panels and secondary
graphite/epoxy cylindrical pressure hull with stiffeners bolted to the metal superstructure.
titanium hemispherical endclosures. The Mast fairings on Navy submarines are fitted
design depth is 6095 m (20 000 ft) and it has
had 134 successful operations, including dives
to 3557 m (12 000 ft). The AUSS locates and
inspects objects on the ocean bottom with
sonar and cameras and transmits the informa-
tion to the surface acoustically. It is an
untethered vehicle, receiving commands
through the acoustic link.

42.6.6 SONAR DOMES

There are numerous submarine and surface
ship sonar domes. The largest sonar domes in
service for the Navy are submarine bow
domes. All bow domes made to date were
fabricated at the Terminal Island facility of
Hitco's Defense Products Division. The struc-
tures are fabricated with glass/epoxy
prepreg and are autoclave cured in one stage,
which is remarkable because of the size,
weight and thickness of the parts. The mater-
ial characteristics necessary to allow the
one-stage cure are stringent and only four
materials have been qualified for use in sonar
bow domes (Hitco, 1986). The mechanical
properties specified are also restrictive, par-
ticularly the value of Mode I fracture
toughness, where an incipient value of
2627 J/ m 2 (15 in lb / in) is specified.
Fig. 42.3 Submarine fairwater. (Courtesy of Lunn
Industries, Inc. VVyandanch,~.)
924 Marine applications

on periscope masts, antenna masts and 42.6.8 CONTROL SURFACES

snorkel masts. The masts are constructed from
The NR-l nuclear powered research subma-
epoxy / fiberglass and some types are filament
rine was targeted by NAVSEA and the David
wound.
Figure 42.4 illustrates a propulsion shaft Taylor Research Center as a test platform for
the first all-composite diving planes. The div-
fairwater for the US Navy's CG-47 class cruis-
ers. These hybrid-composite components were ing planes are hybrid composite construction
with a syntactic foam core.
filament wound with continuous carbon and
glass fibers on a six-axis filament winder. They
provide a 50% reduction in weight over their 42.6.9 OIL PLATFORMS
metallic counterpart while exhibiting superior
stiffness and strength. These cantilevered The offshore oil industry is making use of com-
structures are subjected to large hydrody- posite materials on their platform structures.
namic loads. The composite system resists Examples of applications are: composite drain
corrosion and biofouling due to a co-cured lines, composite cable trays, composite well
external jacket which also improves acoustic head access platforms (Fig. 42.5) and compos-
characteristics. ite firewater systems (Fig. 42.6). The structures

Fig. 42.4 Propulsion shaft fairwater. (Courtesy of B.E Goodrich Aerospace, Engineered Polymer Products
Business Group, Jacksonville, FL.)
 Machinery 925

42.6.11 DECKING

Aside from the application mentioned in the

section on oil platforms (Fig. 42.5), composite
decking is used where lightweight non-cor-
roding decking is needed. Examples are ship
superstructures, work platforms and catwalks.

42.6.12 SHIPBOARD ARMOR

Composite deckhouse armor is used by the
Navy in the LHD class. LHD-1, the USS Wasp,
has panels composed of Kevlar reinforced
Fig. 42.5 Composite wellhead access platform. vinyl ester. For LHD-2, LHD-3 and LHD-4, the
(Courtesy of EXXON Production Research material was changed to S-2 fabric-reinforced
Company.) phenolic, each vessel of which carries
64 000 kg (140000 lb) of armor plate (Owens
Coming Fiberglas, 1991). The change was
made to reduce cost without sacrificing ballis-
tics protection. The fiber/matrix adhesion in
the 5-2 material was deliberately engineered to
be poor, which results in superior resistance to
penetration by projectiles.

42.7 MACHINERY

42.7.1 PROPULSION SHAFTING

Use of composites in propeller shafts is being
investigated at the Naval Surface Warfare
Fig.42.6 Composite firewater systems on EXXON's Center, Carderock Division (Wilhelmi,
BATT platform. (Courtesy of Ameron Inc.) Appleman and Loo, 1986). The advantages
over steel are reduced galvanic corrosion,
are fabricated from pultruded vinyl ester with weight, cost and bearing loads, while allowing
an outer scrim of continuous glass mat while an increase in fatigue stresses, flexibility and
the pipelines are mament wound glass/epoxy. vibration damping. A small diameter 63.5 mm
(2.5 in) mament wound propulsion shaft has
been evaluated in a 4-year trial aboard a Naval
42.6.10 CONCRETE FORMS
Academy Yard Patrol Vessel (YP-654 Class).
Repairs to bridges, piers, docks and other Encouraging results from the YP trial, labora-
structures with deteriorating pilings are tory torsion and fatigue testing and analytical
accomplished by pulling a fiberglass form studies employing finite element stress analy-
around the old piling underwater. The form is sis techniques have led to current R&D efforts
set on a precast concrete footing which fits to develop a Navy-Standard composite shaft-
around the old piling. The fiberglass form is ing 'base laminate', design and procurement
then pumped full of concrete. The form may specifications and a metal! composite cou-
be stripped and re-used or left in place as pling technique suitable for the largest
added protection from ice and floating debris. diameter shafting systems.
926 Marine applications

42.7.2 PIPING SYSTEMS permit measurement of the radial thermal con-

Composite piping systems are gaining accep- ductivity of experimental candidate composite
tance in marine applications (Fig. 42.6). The heat exchanger tubes (Korczynski, 1993).
components of piping systems, piping, fit- Installation of composite centrifugal pumps
tings, valves, heat exchangers and pumps, are (Suitt and Girona, 1993) and piping on Navy
all being investigated for replacement with ships has proven encouraging to date. There
fiberglass (Wilhelmi, 1988). The major reason are some unresolved issues involved with
for replacement is the elimination of corrosion more extensive utilization of composite piping
and galvanic effects, resulting in an increased components in Navy ships, namely, concerns
service life with less maintenance. In addition, about impact, shock and fire resistance.
composite pump components and valves have
improved erosion resistance compared with
42.7.3 DIVING EQUIPMENT
metallic counterparts. A fluid flow loop was
designed and built by NSWC Carderock Composite materials in the diving industry
Division to evaluate the performance of the have proven advantages in numerous applica-
composite components under typical operat- tions. Figure 42.7 illustrates a fundamental
ing flow rates, pressures and temperatures. shift from the standard Navy Deep Sea Diving
This facility also incorporates a single-tube dress consisting of a copper and brass helmet
counterflow heat exchanger configuration to and breast plate bolted to a canvas suit with

Fig. 42.7 Navy standard deep sea diving dress, old and new. (Courtesy of Coastal Systems Station,
Dahlgren Division, Naval Surface Warfare Center, Panama City, FL.)
 Machinery 927

lead shoes and lead weight belt. Out of water

the dress weighed over 90 kg (200 lb). The
modem dress consists of a fiberglass helmet
with neoprene suit and weighs less than 45 kg
(100 lb) on land. Many of the buckles and fas-
teners used to attach lifelines, hot water lines,
emergency backpacks and other accessories
are made of a thermoplastic with a chopped
fiber filler, replacing metal fittings which tend
to corrode. Surface supplied gas bottles for the
diver's breathing air are made from fiberglass
overwrapped cylinders (Fig. 42.8). These gas
bottles can be charged to 20.6 Mpa (3000 psig)
and represent a 50% weight reduction over
conventional steel cylinders.
Figure 42.9 illustrates another type of div-
ing suit. The torso is a carbon fiber reinforced
vinyl ester. It is made by hand lay-up on a
mandrel and room temperature cured while
rotating on a spit. The diver inside the suit
remains at one atmosphere of pressure and
can dive to 610 m (2000 ft) remaining for sev-
eral hours without the need for
decompression.
Portable recompression chambers for trans-
porting divers in need of treatment from
remote areas are made from filament wound Fig. 42.9 Navy One Man One Atmosphere Diving
System (NOMOADS). (Courtesy of Coastal Systems
composite materials. Once inside and pressur- Station, Dahlgren Division, Naval Surface Warfare
ized, the diver may be transported to a Center, Panama City, FL.)
medical recompression facility.
42.7.4 CABLES
Because carbon is conductive, some interest has
been generated in making light weight conduc-
tive cables for underwater instrumentation
using specially treated carbon fibers. Due to the
weight of copper or aluminum wire, towed
electronics underwater are difficult to engineer.

42.7.5 BUOYS AND FLOATS

Sophisticated buoys for oceanographic mea-
surement have been developed for long term
mooring of instruments at depths up to 914 m
Fig. 42.8 Special diver air support system. (3000 ft) of sea water. The buoys are made of
(Courtesy of Structural Composites Industry, a syntactic foam and are deployed in a subsur-
Harsco Company.) face mode. They can provide over 680 kg
928 Marine applications

(1500 lb) of buoyancy. In addition, buoyancy New marine applications of composites are
systems are used on risers for offshore oil rigs. also under development. A particularly chal-
Because drilling takes place in excess of 1.6 km lenging and intriguing application is deep
(1 mi) deep on some rigs, the drill sections diving man-rated pressure hulls. The
become extremely heavy and require flotation Advanced Research Projects Agency (ARPA)
to offset the enormous weight. Syntactic foam is sponsoring the Thick Composites
with a tough outer composite skin is used for Technology Development program to design,
durable flotation. Many Coast Guard buoys fabricate and evaluate a Man-Rated
are now being deployed with syntactic foam Demonstration Article constructed of carbon
cores and composite skins. reinforced composite materials (Kelly, 1993).
The objective of this program is to develop the
technology, experience and confidence neces-
42.8 CURRENT AND FUTURE
sary to demonstrate the feasibility of using
DEVELOPMENTS
composites in primary structure in underwa-
Use of composites in the marine environment ter vehicles (Hoffman and Kelly, 1992; Leon
continues to evolve. There are several recently and Coffenberry, 1992; Camponeschi et al.,
developed resins, prepreg fabrics, cores and 1994).
fabrication processes whose implementation
will significantly change the industry. REFERENCES
Perhaps the most significant change is the
development of effective VARTM processes, Aylor, D. and Murray,J.1992. The Effect of Seawater
Environment on the Galvanic Corrosion
most notably SCRIMP (Seemann, 1990). Using
Behavior of Graphite/Epoxy Composites
conventional resins, fabrics, cores and tooling, Coupled to Metals. Carderock Division, Naval
VARTM increases fiber content over hand lay- Surface Warfare Center SME-92/32.
up and almost eliminates VOC emission. Baltek 1986. Baltek Data File 15I.
Resin development efforts include the for- Burrell, P.P., Herzog, D.J. and McCabe, R.T. 1987. A
mulation of polyesters and vinyl esters which Study of Permeation Barriers to Prevent Blisters
cure upon exposure to ultraviolet light (Pfund, in Marine Composites and a Novel Technique
1992). The major advantage is that these mate- for Evaluating Blister Formation. Froc. 42nd
Annual SFI Conference.
rials cure at room temperature but have an Camponeschi, E.T., Jr., Bohlman, R.E., Hall, J. and
indefinite working life. Carr, T.T. 1994. The Effect of Assembly Fit-Dp
The activity in prepreg development is low Gaps on the Compression Response of Thick-
temperature curing epoxies (Pfund, 1993). Section Carbon/Epoxy Composites. ASTM
Materials are available which cure at 85°C STP-Compression Response of Composite
(185°P) and have several weeks working life, Structures, 1994.
Davies, P. and Lemoine, L. 1992. Nautical
or cure at about 60°C (140 0 P) with a few days
Construction with Composite Materials, Froc.
working life. Excellent properties can be International Conference Paris, France, 7-9
attained with vacuum bag consolidation December 1992.
Ouska, 1993). Garvey, R.E. 1990. Composite Hull for Full-Ocean
In core technology, low cost thermoplastic Depth. MTS Journal, 24(2):49-58.
honeycombs are being investigated to replace Gougeon 1992. Gougeon Laminating System.
PVC foam and balsa in some applications, Product Literature, Gougeon Brothers Inc.
Gougeon, M.A. and Bertelsen, WD. 1993. The
particularly to reduce structure-borne noise.
Gougeon Hydromat Test System: Special
Use of composite core will probably gain Methods and Equipment For Investigating the
wider acceptance. In addition, new structural Effects of Various Pressure Loadings on
foams are being developed that are not PVC- Sandwich Composite Panels. Froc. 48th Ann. SFI
based. Con! and Expo '93.
 References 929

Greene, E. 1990. Use of Fiberglass Reinforced Kelly, J.J. 1993. Thick Composites Technology
Plastics in the Marine Industry. Ship Structure Development Program.
Committee Report SSC 360. Kramers, S.D. Engineering for the Americas Cup.
Hepburn, LCDR RD., Magliulo, G. and Wright, Proc. 1993 Intern. SAMPE Conf, Philadelphia,
T.w. 1991. The US Navy's New Coastal PA, 25-28 Oct 1993.
Minehunter (MHC): Design, Material and Korczynski, I.E 1993. Development of Composite
Construction Facilities. Naval Engineers J., May Heat Exchangers For Seawater Applications.
1991. Proc. 1993 Intern. SAMPE Conf, Philadelphia,
Hitco 1986. Hitco Technical Report 20-04.16, Rev.C PA, 25-28 Oct 1993.
Hoffman, P. and Kelly, J.J. 1992. Manufacture of Leon, G.F. and Coffenberry, B. 1992. Proposed
Advanced Composite Submarine Structures Fabrication Process for Thick Composite
(MACSS). Proc. 37th Intern. SAMPE Symp. Submarine Structures. Proc. 37th Intern. SAMPE
Huss, J. Rowland. 1990. Structural Response of Symp.
Marine Sandwich Panels to Uniform Pressure Lewit, S. 1990. Use of Polyester Mat as a Bedding
Loading. Naval Engineers Thesis, Massachusetts Material for Cross-linked PVC Foam Cored
Institute of Technology, Cambridge, MA. Production Laminates. Proc. 3rd Intern. Conf on
Interlux 1993. Technical Bulletin 900B Marine Applications of Composite Materials,
Jacquemet, R and LaGrange, A. 1988. Aging of Florida Institute of Technology, Melborne, FI.
Laminated Polyester/Glass and Evolution of Little, Bl, Ray, R, Wagner, P., Lewandowski, Z.,
Their Mechanical Characteristics in Marine Lee, W.C, Characklis, w.G., Mansfield, F. 1991.
Media. Composites, 4:40-46. Biofouling, 3:45.
Johannsen, T.}., 1990. Correct Core Installation, Little, B.I. and Wagner, P. 1986. J. Adhesion 20:187.
Bonding Techniques and a Repair Case History Marino, R, Rockett, T., Rose, V. 1985. Blistering of
of a 56 ft Sailboat and a 47 ft Sportfishing Yacht. Glass Reinforced Plastic Marine Materials: A
Paper read at the Atlantic Marine Surveyors Review. NOAA/Seagrant Technical Report 88,
Conference on Newest Developments in Marine Advisory Service, Narragansett, RI.
Boatbuilding and Repairing, 9-10 October 1990, Miriyala, S.K., Tucker, W.C, Brown, R and Rockett,
Annapolis, MD. T.J. 1993. The Mechanism of Galvanic Blistering
Jones, J.M., Walsh, M., Mansfield, EB. 1991. in Carbon Fiber Composites. Proc. [CCM/9
Microbial Electrochemical Studies of Coated Madrid, Spain V.5 pp. 546-553.
Steel Exposed to Mixed Microbial NRaD 1992. Naval Command, Control and Ocean
Communities. Corrosion 91, Paper 108, NACE. Surveillance Center Technical Document 2348,
Jones, I.M., Vasanth, K.L., Conrad, T.K., Little, B., San Diego, CA 95152-5000.
Ray, R 1992a. Corrosion Resistance of Several NSWC, Carderock Division. 1993. Low Cost, High
Conductive Caulks and Sealants From Marine Quality Composite Ship Structures Technology
Field Tests and Laboratory Studies With Mixed Demonstrated. Research Release, David Taylor
Communities Containing Sulfate Reducing Model Basin, Bethesda, Maryland, 20084-5000,
Bacteria. Intern. Symp. on Microbiologically May 1993.
Influenced Corrosion (MIC) Testing ASTM. Owens Coming Fiberglas 1991. S-2 Glass Armour.
Jones, J., Walsh, M., Little, B., Ray, R, Mansfield, F. Case History. Landing Helicopter Deck (LHD)
1992b. ESEM/EDS Studies of Coated 4140 Steel Deckhouse Armor Application, Owens Coming
Exposed to Marine, Mixed Microbial Fiberglas Corp., Pub. No. 15-ASP-16157-A., Jan
Communities Including SRB. 8th Intern. Congr 1991.
on Marine Corrosion and Fouling, Taranto, Italy. Pfund, B. 1992. Light-Curing Resins. Professional
Juska, T., Mayes, J.S. and Seemann, W. 1993. Boatbuilder Magazine, No. 18, August/
Mechanical Properties and Impact Damage September 1992.
Resistance of Composites Fabricated by Low Pfund, B. 1993. Pre-preg Reinforcements.
Cost, Vacuum Assisted, Resin Transfer Molding. Professional Boatbuilder Magazine, No. 24,
Naval Surface Warfare Center, Carderock August/September 1993.
Division, SSM-64-93/04. Plunkett, J.D., Cohen, R, Kunz, B.P. and Kunz, J.
Juska, T. 1993. An Evaluation of Low Temperature 1992. Light-weight, Volume-Produced
Curing Prepregs. Naval Surface Warfare Center, Industrial Honeycomb Composites For Marine
Carderock Division, SSM-64-93/07. Structural Applications. MTS '92 Conf Proc.,
930 Marine applications

Marine Technology Society, Washington, D.C, Suitt, D. and Girona, F. 1993. Development of a
October 19-21. Standard Family of Composite Material
Reichard, RP. 1992. Pressure Testing of FRP Centrifugal Pumps for Naval Surface Ships.
Sandwich Panels. MACM '92 Con! Proc., Naval Engineers J., May 1993.
Composites Education Association, Inc., March Tucker, W.C, Brown, R and Russell, L. 1990.
22-24, 1992. Corrosion Between a Graphite/Polymer
Russell, M. 1992. Inspection and Repair Manual for Composite and Metals. J. Composite Materials
Fiber Reinforced Plastic Boats and Craft. 24:92.
NAVSEA Technical Manual T9008-B4-MAN- Tucker, W.C 1991. Degradation of
OlD. Graphite/Polymer Composites in Seawater. J.
Seemann, W. 1990. US Patent Number 4 902 215. Energy Resources Technol. Trans. ASME 113 4:264-
Sloan, F. and Nguyen, H. 1992. Applications of 268.
Extended-Chain Polyethylene Fibers in the Tucker, W.C, Little, B.J., Ray, R, Wagner, P. 1993.
Marine Composites Industry. MACM '92 Con! Microbial Degradation of Fiber Reinforced
Proc., Composites Education Association, Inc., Polymer Composites. Proc. ICCM/9, Madrid,
March 22-24, 1992. Spain, V.5, p. 554
Smith, CS. 1990. Design of Marine Structures in Wilhelmi, G.F. 1988. Composites for Ship
Composite Materials, London and New York: Machinery Applications. Presented at the
Elsevier Applied Science. SNAME Spring Meeting, lD June 1988,
Stachiw, J.D. and Frame, B. 1988. Pittsburgh, PA.
Graphite-Fiber-Reinforced Plastic Pressure WIlhelmi, G.F., Appleman, W.M. and Loo, T.C 1986.
Hull Mod 2 for the Advanced Unmanned Composite Shafting for Naval Propulsion
Search System Vehicle. Naval Ocean Systems Systems. Naval Engineers I., July 1988.
Center Technical Report 1245. Venus-Gusmer 1989. 6th Edition Catalog.
COMMERCIAL AND INDUSTRIAL 43
APPLICATIONS OF COMPOSITES
Stewart N. Loud

43.1 INTRODUCTION materials such as E-glass fiber reinforcements,

and polyester, some of the low-temperature
This chapter describes the use of composites
performing epoxies, or a variety of engineering
for industrial and commercial applications,
thermoplastic resins, usually referred to as
which perhaps is best described as
fiberglass-reinforced plastics or FRP (or in
'Miscellaneous' or 'All Other Applications'. It
Europe, GRP). Processes used have been
is ~n overview of the myriad number of appli-
mainly the hand lay-up, compression, injec-
~ations t~at cannot always be neatly grouped
tion, filament winding and pultrusion
mto major sectors as with transportation,
methods. Most unit volumes for advanced
marine, aerospace/defense, sports, and the
composites range from the hundreds of units
like. Briefly described are representative appli-
per year to a few thousand, unlike the tens of
cations, the benefits offered by composites,
thousands of units typical with automotive
highlights of the primary materials and
parts. Some of the most mature FRP applica-
processes, and some conclusions about the
tions (yet still with solid growth rates) include
a~proximate market size for this extremely
electrical insulating sheet and sheet molding
dIverse market sector. No attempt is made to
~ompound (SMC), bulk molding compound
define the technology of the materials and
msulators, glass fabric reinforced electronic
processes, as these are well covered in other
chapters. printed circuit boards, pultruded items such as
ladders, filament wound electrical tubing,
thermoset SMC and injection molded thermo-
43.2 HISTORICAL PERSPECTIVE plastic business equipment and appliance
housings, hand lay-up antennas, and others.
Advanced composites (typically produced
using high-performance fibers and/or resins) This base of FRP applications has been used as
historically have been dominant in the aero- a launching platform for a huge diversity of
new applications. Many now utilize the more
spac~ and defense sectors and in many
advanced fibers and resin systems or hybrids
sporting goods products. Hand lay-up with
of these material systems.
autoclave molding have been the major
processes used for part fabrication. Engineered
composites applications for several decades 43.3 PROGRESS
have been fabricated predominantly from
With the slowdown in worldwide defense
spending during the early 1990s, numerous
Handbook of Composites. Edited by S.T. Peters. Published pro~~cers have looked to new market oppor-
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 turuties for growth, especially for advanced
932 Commercial and industrial applications of composites

fiber and resin products, and increasingly to industry to be counter productive to even use
hybrids of several materials. As a result, the term 'advanced composites' anymore.
advanced and hybrid fiber reinforced prod-
ucts are now moving into the application
43.4 APPLICATIONS REVIEW
sectors formerly dominated by metals and
even glass fiber products. Also, the focus on Following is an alphabetical review of applica-
Idefense conversion' has resulted in retarget- tion segments of this market or of unique and
ing of some governmental budgets to support emerging material systems used broadly in
defense contractor conversion to civil applica- related industrial applications. Representative
tions. Many of these applications reside in the organizations are cited as examples of the
commercial and industrial market sectors, the types of companies and institutions active in
focus of this chapter. these market sectors so that the reader can
A vast diversification of composites has research further into the applications of inter-
occurred in this sector. From the 1960s and est. Applications believed at this time to offer
1970s with many electrical, electronic, appli- the opportunity for significant growth are pre-
ance and business equipment applications, sented in the most depth. Some applications
the business has grown into pressure vessels, covered also fit into other market sectors, but
laptop computers, centrifuges, flywheels, fuel their technology is unique or they are indus-
cells, railroad car components, drive shafts, trial components sold into the other markets,
fuel pumps, loom components, musical so they are described here. Throughout the
instruments, oil and gas riser pipe, power chapter, tables highlight many applications by
lighting and distribution poles, industrial major market sector, some long established
process rolls, advanced wind turbines and and many in the emerging growth stage, to
much more. The processes used here have stimulate the reader's thinking on other poten-
remained fairly constant: pultrusion, filament tial applications that might benefit from
winding, compression molding, injection applying composites. Those applications in
molding and hand lay-up, but the use of resin the tables that tend toward more usage of the
transfer molding has increased dramatically highest performance materials are listed with
for many of these applications. Materials an asterisk (*).
applications that were almost exclusively
fiberglass and polyester, thermoplastics, or
43.4.1 ANfENNAS
low-temperature curing epoxies have evolved
into much more vinyl ester usage and Telescoping composite tubes are used to form
migrated to the use of more high-performance a lightweight antenna support intended for
fibers such as Kevlar® and Twaron® aramids, military use and the ultimate in portability.
Spectra® high-molecular weight polyethylene The tubes must be exceptionally light, so the
(HMWPE), and many types and forms of the producer used woven Spectra UHMWPE fiber
carbon fibers. Ceramic matrix composites in epoxy resin. They are fabricated using a
(CMCs) and metal matrix composites (MMCs) conventional roll-wrap process similar to that
are of increasing interest and are sometimes used to produce composite fishing rods. The
competing with polymer-based fiber compos- total weight is only 1.8-2.2 kg (4-5 lb) includ-
ites (PMCs), for example in automotive drive ing 11 nested tubes that collapse to only 711
shafts. Even carbon fiber reinforced carbon mm (28 in) for portability but can easily be
matrix materials (C-C) are moving from the extended and locked to form a support pole
aerospace sector into commercial use. The use 6.7 mm (22 ft) long.
of 'advanced' materials has now matured Resin transfer molding is increasingly dis-
enough that it is considered by many in the placing hand lay-up for fiberglass/polyester
 Applications review 933

communications antenna dishes. When the economics and performance versus conven-
new technologies for direct-broadcast satel- tional technologies of steel and concrete.
lites become more commercial, small Other organizations are working on a vari-
composite dish antennas, probably produced ety of bridge programs where composites are
using compression molded SMC, could the central enabling technology for improved
become a booming market. civil engineering structures. For example, FRP
and CFRP laminated plates are
adhesively / mechanically bonded to the
43.4.2 BRIDGES
underside of steel or concrete bridge beams to
Mostly small foot bridges or the railings, lad- extend their structural in-service life. In
ders and decking of structures in highly Scotland, in 1992, an all-composites cable-
corrosive environments typically have used stayed footbridge was installed between two
FRP components over the last few decades. portions of the Linksleader Golf Club and
However, in the last several years there has across the Tay River (Fig. 43.1). Installation
been a great increase in interest in the use of was accomplished using mostly on-site hand
fiber composites for civil engineering and labor. Much of the structure was constructed
infrastructure applications. One segment of
particularly high interest is composite bridges
that offer great corrosion resistance and
weight savings. Composite tendons or rein-
forcing bars reinforced with fiberglass, carbon
fiber or aramid fiber are being studied aggres-
sively for use as rebar or for prestressing of
concrete bridge decks. Defense Advanced
Research Projects Agency (DARPA, Arlington,
VA) has funded a Technology Reinvestment
Project (TRP) at the University of California at
San Diego (UCSD) and with the Advanced
Composites Technology Transfer /Bridge
Renewal Consortium for three technology
demonstration areas. First is the development
of low-cost fiberglass or carbon fiber rein-
forced plastic (CFRP) filament wound jackets
for the repair and upgrading of freeway
bridge or parking structure columns to make
them more earthquake resistant. Second is a
composite or hybrid composite/concrete
bridge deck technology that would allow
rapid replacement of bridges on highways
and interstates. Last is a plan for a 130 m (450
ft) cable-stayed all-composite vehicular
bridge to be built at UCSD over the Interstate
freeway at La Jolla, CA, in the late 1990s. All
of the advanced fibers, along with E-glass
Fig. 43.1 All-composite cable-stayed bridge, assem-
fiber, are of high interest for these projects. bled from isopolyester composites. The bridge's
Also, high-speed, low-cost fabrication A-frame towers are 17.5 m (57 it) high. (Courtesy of
processes will help to achieve competitive Amoco Chemical Co.)
934 Commercial and industrial applications of composites

Table 43.1 Construction/ civil engineering/ infra- and internal components. For example,
structure Harbec Plastics (Ontario, NY) is a custom
injection molder for a compensating arm for a
Bridge girder upgrade plates" business machine printer housing using 30
Foot bridges"
Cable-stayed bridges"
v /0 fiberglass / polycarbonate compound.
Replacement bridge decks" They also mold a print head carriage using
Earthquake repair and upgrades/column wrap* carbon fiber fPC with Teflon added for wear
Bridge decking, panels, enclosures" resistance. Such parts are common to most
Pipeline rehabilitation liners" typewriters and printers. Carbon fiber pro-
Manhole covers vides electrostatic dissipation. Equally
Trench covers important is the dimensional stability they
Cables, tendons, rebar, dowel bars*
obtain to meet a specification for precision-
Acoustic wall panels"
Marine piling molded inserts with a 0.025 mm (0.001 in)
Piers" concentricity on two critical shaft bearings.
Computer room flooring" Growth in this segment comes mainly from
Curtain walls" further displacement of other materials as
Glulam wood beams* composites gain market share. The benefit of
Mine roof bolts high value here is increased design freedom
Earth anchor rods
offered by the light and strong FRP materials.
Residential ballistic wall panels
Coalescent grid/water treatment" This sector is dominated by the molding com-
Toxic waste disposal tanks" pounders and the custom molders such as
Chemical grating Premix (No. Kingsville, OH) and Rostone
Platforms (Lafayette, IN).
Walkways

Table 43.2 Appliances/business equipment

from pultruded shapes. In rnid-1994, in the Sanitaryware

UK, a 2720 kg (6000 Ib) replacement lift bridge Dishwashers
deck for a road across a canal was constructed Oven door handles
from pultruded FRP and mounted on steel Frying pan components
beams attached to the lift mechanism. Bridge Phone/fax EMl covers"
counterweights were reduced, offering cost Copier housings and components*
Computer cassette robotic system"
savings. The infrastructure business sector, Refrigerator liners
particularly roads and bridges, could become Business machine bases
one of the largest growth areas for composites Laptop computer cases"
to displace steel and concrete over the next Microphone booms"
decade or more. Development of appropriate Dishwashers
codes and standards along with strong Air conditioners
cost/performance demonstration projects are Office furniture*
Toasters
keys to success.

43.4.3 BUSINESS EQUIPMENT 43.4.4 CABLE CARS

This sector is one of the more mature and has Composite cabins were retrofitted to replace
been heavily dominated for many years by steel and aluminum at Chamonix, France to
compression molded or injection molded reduce the weight of the cable cars at a ski
reinforced thermoplastics for housings, bases, resort. Ingenex, a French engineering firm,
 Applications review 935

built new lightweight car bodies using sand- Industrial Technologies. Their Continuous
wich construction with Kevlar/epoxy face Fiber Ceramic Composites (CFCq Program
sheets and a Nomex:® honeycomb core. The Plan (lO-year plan) was an expansion of the
supporting arm that holds each cable car is fab-continuing support activity directed toward
ricated from carbon fiber/epoxy over a Nomex 'enabling technology' efforts since 1987. DOE
core. These replacement structures allow for defines a CFCC as a long fiber (ceramic, glass
many more passengers to be carried on the or carbon) embedded in a ceramic matrix.
existing system. This avoided having to scrap Concerns about CMCs raised by potential
the existing system that was becoming over- industrial users include material and process-
loaded, thus avoiding rebuilding the whole ing costs, long-term component fracture
system reportedly at a cost 10 times as much. toughness, general durability in a variety of
environments, the ability to join components
with different compositions and the availabil-
43.4.5 CARBON-CARBON COMPOSITES FOR
ity of near-net-shape fabrication methods with
NON-AEROSPACE MARKETS
3-D fiber architectures (preforms). A relatively
The vast majority of uses for C-C composites new type of ceramics, CFCCs offer the poten-
are in aerospace and defense. However, many tial to meet the demands for a variety of
potential industrial applications are emerg- industrial applications. Some primary applica-
ing for this unique high-temperature resistant tions envisioned for these materials include:
material system. DOE has studied C-C com- burners/combustors, chemical reactors and
posites for fusion energy systems. One process equipment, heat recovery systems,
project involved brazing C-C composites. refractories and related products, separa-
Another program studied development of tion/ filtration systems, stationary engines,
pitch-based carbon fiber (PBCF) reinforced waste incineration systems, tooling, structural
C-C composite tiles to line the inside walls of components, biomedical parts, wear parts in
future fusion devices such as the Compact machinery, and other specialty high-tempera-
Ignition Tokamak (CIT). Perhaps such tiles ture products.
later will find application in commercial heat Hexcel Corp. (Pleasanton, CA) offers high-
exchangers. Kureha Chemical Industry Co., performance ceramic materials for various
Ltd, and Osaka Gas Co., Ltd of Japan are military components such as missile fins,
active in the USA market. Applications tar- radomes, jet engine parts, and high-tempera-
geted include C-C structures for ture tooling, but much larger potential
high-temperature furnaces and fuel cells, wet markets might develop in the commercial sec-
and dry friction materials, plus brake disks tor, especially for less-expensive CFCCs using
and self-lubricating bearings for automotive E-glass fiber or other low-cost reinforcements.
and truck applications. C-C brakes and Examples cited by Hexcel include honeycomb
clutches are used worldwide in some racing heat exchangers used in high-pressure coal-
car classes such as Formula 1 and for the fired gas turbine generators and feed tubes
Indianapolis 500 race but are banned from (used to overcome corrosion and thermal
other classes due to the premium cost. shock problems) for a toxic waste incinerator
operating at 870°C (1600°F).
Successful conversion of the applications
43.4.6 CERAMIC-MATRIX COMPOSITES (CMCs)
noted above could generate savings of 1.1
These materials have received a significant quads (a quadrillion BTUs) of thermal energy
boost in funding because of continuing inter- per year, and reduce NOx emissions by 0.6 mil-
est by the USA Department of Energy Office of lion tons/year, according to the DOE.
936 Commercial and industrial applications of composites

43.4.7 CONDUCTIVE FIBER FOR SHIELDING

BUSINESS EQUIPMENT

Cytec Industries sells a conductive fiber con-

centrate or intermediate containing carbon
fiber for electrical conductivity. In most of
these products, Cytec starts with continuous,
standard 33 msi, 12K carbon fiber that is elec-
~oplated with nickel before being chopped
mto 6.35 mm (0.25 in) lengths. The fiber load-
ing is typically 60-75%, even to 90% if sold to
a compounder who adds resin to make a spec-
ified injection molding compound.

43.4.8 CORROSION-RESISTANT SECONDARY

STRUCTURES - GRATING, LADDERS, PIPING Fig. 43.2 Conductive FRP grating for static electric-
ity dissipation. (Courtesy of CI/SPI.)
Composite products are offered by many com-
panies including Amalga, Brunswick, 43.4.9 DRNE SHAFTS - PICKUP TRUCKS AND
Fibergrate, International Grating, Creative COOLING TOWERS
Pultrusions and others. Strongwell takes pul-
truded structural fiberglass shapes Strongwell continues to use substantial
(tradename: Extren), fabricates them into use- amounts of carbon fiber for the hybrid drive
ful structures, and distributes to various shafts they sell to Dana for shipment to
industrial customers. The material is available General Motors. At close to 226 800 kg (500 000
in over 100 standard sizes and shapes, and is lb) per year of carbon fiber usage and esti-
mated to climb soon to one million lb, this is
produced in three series: general-purpose, fire-
retardant and a premium grade that is both fire said to be the largest industrial application for
car~on fiber reinforcements. The carbon fibers,
retardant and highly corrosion resistant. This
gra.de, called Extren 625, uses a vinyl ester mainly from Hexcet are pultruded with resin
resm; the other two are made with polyester along an aluminum tube and overwrapped
formulations. Applications include handrails with fiberglass for damage tolerance.
stairways, ladders, cable and channel trays: Carbon fiber/epoxy is used for cooling
platforms and walkways, troughs, weir and tower drive shafts as well. SpyroTech (Lincoln,
N~) is a major supplier of these shafts along
baffle plates, racks, bridges, identification
signs, sump tanks, battery boxes, drain pipes, WIth Amalga Composites (W. Allis, WI),
etc. Besides the offshore drilling and produc- Addax, Inc. (Lincoln, NE) and ACPT
(Huntington Beach, CA). A circumferential
tion industry, Strongwell serves a variety of
industries, including aquaculture, chemical wrap of E-glass / epoxy is added for debulking
and impact protection. These shafts exhibit up
proc~ssing, food/beverage, marine, metal pro-
cessmg, waste treatment, petrochemical, to five times the durability at one-fifth the
power plants, pulp and paper mills, radar weight of steel shafts and are far more corro-
installations, transportation and many special- sion resistant. Cooling tower shafts (as with
ized applications such as raised computer vehicular shafts) produced with carbon fibers
ar~ stiffer and can be produced in lengths
floors. For static diSSipation, carbon fibers can
be located at the surface of grating (Fig. 43.2). n:Ice .as long as with steel without suffering
VIbratIon damage in use, eliminating the need
for intermediate bearings and supports.
 Applications review 937

43.4.10 ELECTRICAL AND ELECTRONICS insulating hardware, ladders and hotline tools
for line installers, lighting poles, connectors,
This sector of the market is composed of many
microchip encapsulation and many more.
applications that are maturing or have high
Fiberglass reinforcement predominates due to
penetration compared with other materials.
its insulating characteristics and low cost.
Applications include motor and generator
Sales are cyclical due to the high market share
components, switch gear, insulating sheet,
held by FRP, but there is still strong growth
molded insulators, subway third rail insulat-
seen from penetration gains versus phenolic,
ing cover board, FR-4 and G-10 /11 glass fabric
porcelain, and other materials.
reinforced epoxy printed circuit boards
(PCBs), utility transmission and distribution
43.4.11 FIREFIGHTER BREATIflNG APPARATUS

Table 43.3 Electrical/ electronics/ energy The key component of firefighter breathing
systems/ communications apparatus is the air storage bottle that comes
in different sizes, materials, and pressure rat-
Electric motors ings. Structural Composites Industries
Fuse tubes (Pomona, CA) continues to be well ahead of all
Lighting poles competitors making small filament-wound
Distribution poles composite pressure vessels. SCI supplies air
Cable tray bottles (all of their firefighter's bottles use 5-2
Power tool cases
Motor control centers Glass) to Mine Safety Appliances. They report-
Printed circuit boards edly have never had a field failure. Many air
Circuit breakers breathing systems distributed by MSA are
Guy wire insulators used to escape smoke or other noxious materi-
Utility transmission towers als in mines, chemical plants, refineries, hotels,
Electrical bus bars aircraft, etc. SCI also is a major supplier for the
Distribution transformer cores, cases escape-slide air bottles in use on transport air-
Shielding - RFI, ESD, EMI"
Battery boxes, casings craft built by Boeing and Airbus. Many
Electron beam accelerators" composite oxygen storage bottles are pro-
Electrical insulators duced for hospital and home care and for use
Electrified third rail cover board in highflying commercial and general-aviation
Electrical switchgear aircraft (for example, for escape slide inflation
Power hot-line tools but using Kevlar 49/ epoxy overwraps for
Electronic connectors decreased weight).
Outlet boxes
Aerial lift truck booms Other composite bottles are used to inflate
Electronics chassis* rafts, balloons and other flotation systems. EFI
Computer chip carriers* Corp. (San Jose, CA), a subsidiary of Racal
Electronic equipment racks* Electronics in the UK, and Luxfer USA Ltd
Fuel cells* (Riverside, CA) continue to be SCI's main
Flywheel mechanical batteries* competitors as suppliers of firefighter air bot-
Superconducting magnetic energy storage (SMES) tles. They sell air bottles to Scott, Survivair,
Wind turbine blades"
Telescoping portable antennas" North Safety and others in the USA and over-
Communications antenna dishes seas. About 250 fire departments in American
Radio masts" cities have bought hundreds of thousands of
Satellite TV dish antennas composite bottles to replace the much heavier
Microwave guides steel bottles. Lighter bottles with longer last-
Fiber optic cable tension members ing air supplies give firefighters more time on
938 Commercial and industrial applications of composites

station in the heat of a fire fight. The bottles Larger flywheels also show promise for
must store air for long periods at 31 MPa stationary energy storage and peak power
(4500 psig). shaving applications, and the business
opportunity for such systems could be enor-
mous. Federal technology centers, such as the
43.4.12 FLYWHEEL MECHANICAL BAITERY
Lawrence Livermore and Oak Ridge National
SYSTEMS
Laboratories, now are part of teaming
US Flywheel Systems, Inc. (Camarillo, CA), arrangements to enhance flywheel technol-
Calstart (Burbank, CA), American Flywheel ogy transfer from defense and other
Systems (Seattle, WA) and others all hope to government programs to the private sector.
supply kinetic energy storage flywheel sys- The rims of the rotating wheels in these sys-
tems (now often called mechanical storage tems are filament wound using advanced fiber
batteries) for future electrical vehicles in composites, including E glass, S-2 glass,
California and elsewhere (Fig. 43.3). Flywheel Kevlar and carbon fibers. At speeds attainable
batteries have an estimated 5-10 times the by composite wheels, metal flywheels disinte-
watt hours per pound storage capacity of con- grate. Use of this system is a good example of
ventional lead acid batteries. They are technology transfer from aerospace to the
impervious to temperature variations, may be civilian sector.
drained repeatedly of stored energy without An electric car with a flywheel storage sys-
cyclic degradation, can be recharged in as little tem on board (and perhaps with regenerative
as a few minutes, should last 30-40 years with- braking through the flywheels) could
out requiring replacement or maintenance, approach zero emissions and is expected to
and create no toxic waste problems. Mileage meet the much tighter California air quality
range in an electric car (such as GM's Impact) standards set for early next century. California,
could reach a projected 482--640 km (300-400 New York and Massachusetts, have mandated
miles) or more according to some developers, that 2% of cars sold in those states in 1998
as opposed to a practical range of only 95-128 must be zero emission vehicles; the percent-
km (6<H30 miles) for today's electric cars using ages increase to 5% in 2001 and 10% in 2003.
conventional battery systems. The challenge is how to offer an affordable car
with desirable range and performance that can
meet those goals.
Lawrence Livermore National Laboratory
reported during a SAMPE seminar that the use
of composite flywheels in an electromechani-
cal battery 'would be ideal for use in electric
powered vehicles'. They found that a high-
strength carbon fiber (such as Toray's T700 or
TlOOO) yielded the highest energy storage
capacity. Since then, Lawrence Livermore
National Laboratory and General Motors
Corp. have teamed on a three-year, $3 million
project to develop a new generation of more
efficient automobile\flywheel batteries. The
GM vehicle project is part of the 'Partnership
for New Generation Vehicles program'
Fig. 43.3 Flywheel battery components. (Courtesy launched by President Clinton in September
of American Flywheel Systems.) 1993.
 Applications review 939

43.4.13 FUEL TANKS FOR NATURAL GAS Chrysler and many city transit bus operators
VEHICLES (NGV) such as Tacoma, Sacramento, Salt Lake City,
Reading, Binghamton and other Northeast
Lincoln Composites (Lincoln, NE) and EDO and Canadian cities. Recent failures of CNG
Corp. (Canada) introduced certified 'all-com- tanks during refueling (reportedly caused by
posite' compressed natural gas (CNG) exposure to unexpected acid solutions) appear
pressure vessels under the 1993 US to have been solved with an installation
ANSI! AGA standard called NGV2 for ultra- redesign and protective coatings. Additional
light, nonmetallic natural gas vehicle fuel interest comes from utilities such as Southern
tanks. The availability of these certified plas- California Gas & Electric and Columbia Gas
tic-lined, composite-overwrapped fuel tanks is companies.
expected to accelerate the usage of advanced Carbon fiber priced today at about $5-6/kg
composites into this potentially large alter- ($11-13 /lb) for high-strength 12K fiber is
nate-fuel marketplace for composites. Besides about twice as expensive as S-2 glass fiber and
weight savings of up to 70% (compared with much higher than E glass but appears eco-
steel, aluminum, or metal-lined NGV tanks) nomically competitive for this market. One of
and greater durability, the new tanks offer the above supplier companies did a cost and
increased fuel capacity, extending driving weight analysis to decide how much of a pre-
range. mium would have to be paid if
The weight savings are achieved through higher-strength fibers were used rather than E
use of a hybrid composite, for example using a glass. Using E-glass/epoxy overwrap, the
30 msi modulus T650/35 carbon fiber from weight would be about 45 kg (100 lb). 5-2 glass
Amoco Performance Products (Alpharetta, reduces the weight to 36--38 kg (80-85Ib), and
GA) or a Toray T700S/12K carbon fiber. a 12K carbon fiber/epoxy overwrap reduces
Hybridized with the carbon fiber might be an the weight to 18-20 kg (40-43Ib). Using fiber
E-glass fiber roving from Owens-Corning, costs of $2.20,11-13.00 and 26.50/kg ($1, $5-6
PPG, or Vetrotex Certainteed to reinforce an and $12/lb) respectively, the estimated fiber
epoxy resin formulation. The composite is fila- cost per NGV pressure vessel is $64 for E glass,
ment wound over a thermoplastic liner of $300-315 for 5-2 glass, and $285-310 for car-
high-density polyethylene (HOPE). In the bon fiber. Combining the various cost
hybrid, carbon fiber provides an extremely components brings the heaviest (E glass) tank
high strength-to-weight ratio, with the ability cost to $380-400, 5-2 tank to $700-720, and the
to withstand many pressure cycles along with carbon fiber tank to $600-620. Thus, to save
superior fatigue characteristics, while the about 27.2 kg (60 lb) by switching from E glass
fiberglass enhances toughness. HOPE liners to carbon fiber reinforcement, the cost pre-
are lighter weight than steel or aluminum and mium is about $220, or $3.6--3.7 per pound
more resistant than both to the highly corro- saved. Auto designers in the past have sug-
sive elements sometimes found in natural gas gested that a $2 premium per pound saved
supplies around the USA. A minimum 15-year was acceptable, but they might be willing to
life expectancy is set for the NGV2 tanks; after pay more now in view of the environmental
that they are to be discarded. benefits gained by burning cleaner fuel.
Other pressure vessel programs are under- The combined production of mid-size
way at Comdyne I, Compositek (Kaiser), Fiber pickup trucks in the USA by Chrysler, Ford
Dynamics, Amalga and Structural and General Motors is reported to be more
Composites. Large customers for the older than 600 000 per year. If only 10% are built
fiberglass/epoxy overwrapped aluminum with natural gas as the fuel and there are two
NGV fuel tanks include General Motors and or three fuel tanks on each truck, there is a
940 Commercial and industrial applications of composites

very large potential market in the millions of American Laminators (Drain, OR) and the
pounds of materials required. Transit buses Wood Science & Technology Institute
may be an even larger market. (Corvalis, OR) have developed a superior line
of glued-laminated (glulam) wood beam
Table 43.4 Transportation system equipment products based on Douglas fir and compos-
ites. Because these beams are a hybrid
Seats/frames material system, they marry the light weight
Monorail cars and high strength of aramid/ carbon fiber
Mobile storage modules - CNG* composites with the already light weight,
CNG/NGV fuel tanks* high strength nature of wood to make a prod-
Rail freight car doors
uct that can compete favorably with steel,
Rail car knuckles
Hydrogen storage tanks* concrete, and other wood beams. Of special
Intermodal containers note is that these new composite/wood
Maglev train guideways* beams for buildings or bridges outperform
High-speed train brakes* conventional wood beam products and usu-
Concept/show car chassis* ally can be offered at a purchase price or first
Electric vehicle frames/chassis* cost of about 25-45% less than conventional
Racing car brakes and clutches* glulam beams!
Solar-powered racing vehicles*
Highway reflecting markers Dead weight design load is reduced by
Airport approach light towers 40-60% for large structures and long spans.
Airport ticket counter tops For structures where deflection is a major con-
Drive shafts - pickup trucks* sideration in design, the patented technology
Aircraft escape slide inflation bottles* takes advantage of the high modulus of Akzo
Ship propulsion shafts* Nobel Fortafil carbon fiber with 33 msi modu-
Cablecar cabins and support arms*
lus and 550 ksi tensile to more than double the
Shipboard crane components
design capacity in compression versus typical
unreinforced glulam products. Where
increased tensile performance is required,
43.4.14 GLULAM BEAMS
Akzo Nobel Fibers' Twaron or DuPont's
Composites are leading to revolutionary wood Kevlar 49 aramid yam is used. Some of these
beams for bridges; laminated beams benefit beams can be massive; e.g. typically 24 m long
dramatically from use of carbon and aramid x 1200 mm deep x 254 mm wide (80 ft x 4 ft x
fibers and cost less. Wood structural building 10 in) to much larger. The fibers reinforce pul-
materials have required only the highest qual- truded plates that are adhesively laminated
ity wood to achieve necessary performance during the standard beam forming and bond-
levels. Wood ('the original composite') has a ing.
high strength-to-weight ratio; however, highly Many structures (100 as of this writing) are
engineered components such as beams have already in place using the new composite glu-
had to allow for the inherent variability of lams, in California, Oregon, Wisconsin and
wood and wood types. The best woods have Hokkaido, Japan. More structures now are in
come from 'old growth' forests that today are the design stages (ca. 100) including large
in decline or are in designated wilderness domed structures with over 164 m (500 ft) clear
areas. Logging practices and environmental spans, for Japan, USA and European clients.
considerations (remember the spotted owl?) Estimates by various participants suggest that
have removed 'old growth' as a source and the global opportunity for this
have driven producers to consider new tech- composite/wood beam technology is in the
nologies. billions of dollars. In the USA, the conventional
 Applications review 941

glulam market is over $1.2 billion/year and is TGV car designer and builder, must decide
supplied by 147 beam manufacturers. Potential whether the weight savings, lack of fade and
requirements in the 45 million kg (100 million projected longer life (as with similar C-C air-
lb) range for just the high strength carbon and craft brakes systems) justify a change in the
aramid fibers are seen about the year 2000. One steel brakes now in service. With plans to
observer close to the project said that 'this has build similar high-speed rail systems in Texas
to make a big splash in composites and could and other parts of the USA, there is growing
be the biggest new composites application in interest in the results of the tests. Several com-
20 years!' panies in Japan have similar interest in using
C-C brakes on proposed advanced bullet
trains, which also might reach speeds of 350
43.4.15 HIGH-PRESSURE TUBING AND
km/h. Later magnetic levitation (maglev)
BATTERY CASINGS
trains which will approach 500 km/h may be
Amalga Composites mament winds carbon assessed for use of these ultra-high-perfor-
fiber/epoxy over battery casings for Johnson mance brakes.
Controls. The Inconel liner has a carbon
fiber / epoxy overwrap which is under 0.34
43.4.17 HYDROGEN FUEL STORAGE
MPa (50 psi) stress when discharged and 5.86
MPa (850 psi) when fully charged. These bat- Hydrogen storage tanks may offer potential
teries are designed to be buried in the desert to for filament-wound pressure vessels, but it
power instruments during the night and they will be several years before this can possibly
are recharged during the day. Amalga also occur. Syracuse University received a grant
makes vessels for high-pressure applications, from the Department of Energy to prove a
including a filament wound carbon novel approach to hydrogen storage. The
fiber / epoxy-overwrapped steel tank to with- hydrogen would be stored in a filament-
stand 3.41 MPa (500 psi) for use as an wound tank in which activated carbon
emergency backup pressure cell for the F-22 granules are closely packed. After being
fighter. Other interesting high-pressure prod- cooled in a liquid nitrogen heat exchanger to
ucts are carbon fiber / epoxy-reinforced the desired storage temperature, the hydro-
metallic cylinders that achieve 41-83 MPa gen is passed through the activated carbon
(6000-12 000 psi) of pressure capability. Lower where some is adsorbed. It is then stored at
pressure pneumatic tubes are being made by low temperature (150K) and relatively low
Amalga for service at 1.7 MPa (250 psi), and pressure 5.51 MPa (800 psig) before being
these usually are reinforced with a fiber- released via slight heating of the granular
glass/epoxy overwrap. carbon while opening a valve for controlled
release into the nearby hydrogen-oxygen fuel
cell. A design engineer predicts that a car
43.4.16 HIGH-SPEED TRAIN BRAKES -
powered by an electric motor energized with
CARBON-CARBON
a hydrogen-oxygen fuel cell will be able to
Carbone Industrie in France has been testing outperform a similar vehicle powered by a
carbon--carbon brakes on the French TCV gasoline engine. Since hydrogen as a fuel
Atlantique line. Two similar but slightly differ- requires a massive change to the fueling
ent applications are being tested; for the drive infrastructure, the timing for this market sec-
wheels and for the bogey. The C-C brakes tor is hard to predict. If such an infrastructure
must qualify on cars being stopped from is installed, there will likely emerge many
300-350 km/h (to 215 mph) before they can be more composites applications such as large
considered for regular service. Alsthom, the tanks and pipelines.
942 Commercial and industrial applications of composites

Table 43.5 Miscellaneous molded with nonwoven-fabric are 20-50%

lighter than older, thicker plastic cases and
"Jurassic Park" dinosaurs have a wall thickness of only 3.04 mm (0.12 in)
High-altitude balloon gondolas" without sacrificing ruggedness. The carbon
Aquaculture cages for fish farming
fiber/thermoplastic is supplied by Kobe Steel
Telescope tubes"
Helmets" Ltd. and Mitsubishi Rayon Co., Ltd of Japan.
Superconducting supercollider" In a related competitive development,
Firefighter air bottles" Advanced Logic Research, Inc. (ALR - Irvine,
Oxygen storage bottles" CA), introduced the world's first modular
Life raft inflation bottles" notebook PCs offering processor chip
Flag poles upgrades. The ALR notebooks are ergonomi-
Deep-submergence capsules"
cally designed with a thin carbon fiber
Riot control body armor / shields"
reinforced, lightweight case that fits into a
standard-size briefcase. Other notebook com-
43.4.18 LAPTOP COMPUTERS puter producers have composite development
programs underway.
Lightweight laptop computers with carbon
Based on market share estimates provided
fiber reinforced plastic cases were first intro-
by Stanford Resources, Inc. in PC Magazine
duced in the USA in the early 1990s at the giant
(August 1994), Compaq has the largest market
Comdex computer show in Las Vegas. Toshiba
share today at 14.1% of shipments in 1993,
America Information Systems (Irvine, CA)
Apple is second at 13.6%, Toshiba is third at
introduced three models, reporting that their
11.1%, IBM is fourth at 7.1%, followed by
light weight 2.5-3.6 kg (5.5-7.9 lb) and porta-
dozens of other producers.
bility were partially due to the use of a carbon
fiber molded case (Fig. 43.4). The new cases
43.4.19 LOOM COMPONENTS
DFI Pultruded Composites, Inc. (Erlanger, KY)
was one of the first to successfully pultrude car-
bon fiber/epoxy, once quick-curing epoxies of
the proper viscosity became available. One of
their better customers is in Europe for DFI's
pultruded lightweight, low inertia, carbon fiber
shuttle arms that rapidly reciprocate in the
older style but updated rapier looms still being
used by many industrial fabric weavers. The
shuttle arms are delivered in six lengths rang-
ing from approximately 1.3 to 2.1 m. DFI used
34 msi modulus carbon fibers as the reinforce-
ment. [Incidently, one of their unique products
is a pultruded rod using carbon fiber/epoxy
with a diameter as small as 0.25 mm (0.010 in).]
Polygon Company Div. of Plas/Steel Industries
(Walkerton, IN) pultrudes a carbon
Fig. 43.4 Lightweight computer case with nonwo- fiber / epoxy, high-speed loom frame compo-
ven fiber reinforcement. (Courtesy of Toshiba nent for a USA equipment producer, using
America Information Systems, Inc.) Akzo Nobel Fortafil's SOK filament fiber.
 Applications review 943

43.4.20 MAGLEV TRAIN GUIDEWAYS km ($10 million per mile) for a dual track
guideway. Their design includes a concrete
Magnetic levitation or maglev train system con-
box beam for the guideway with the bottom
cept definition contracts are funded by the
of the vehicle 'wrapped' around the beam for
OOT-Federal Railway Administration under
~ety. The guideway might use a hybrid of
the $151 billion Intermodal Surface
fIberglass and carbon fiber rebar rather than
Transportation Act which was passed in 1991
steel rebar reinforcement. Foster-Miller is
and included $750 million to be used specifi-
focused on high-performance, low-cost guide-
cally to complete the design and construction of
ways. Most of their current emphasis is on
a USA demonstration maglev line about 32 km
using high-strength FRP post-tensioning
(20 mile) long. The German Transrapid maglev
~embers (both transverse and longitudinal)
system design is planned for the 20 km (13
mile) demonstration system to be built in m the upper areas of the guideway. West
Virginia University is working on a DOT-FRA
Orlando, FL. USA and Canadian companies
contract titled: ' Assessment of Maglev
competing for future maglev system contracts
Guideway Systems'. Imbedded fiber optic
include Northrop Grumman, Foster-Miller,
systems may be used to monitor entire guide-
Magneplane, Boeing, Bombardier and
way systems and to detect potential trouble
Morrison-Knudsen (the latter three are propos-
spots such as joints. West Virginia University
ing the most extensive use of composites on the
data show that if FRP rebar or tendons are
'rolling' stock, perhaps including the outer
used, the load-carrying capacity is about
skins of the maglev cars plus the interiors).
twice that achieved with steel. Although the
On the maglev vehicles, General Electric is
initial cost might be five times greater, the
developing shielded superconducting mag-
nets and their coil design, shielding, motors, weight reduction can be 80%, so on a perfor-
mance-per-pound basis, FRP would be lower
levitation and carbon fiber/epoxy cryostats.
in installed cost than steel rebar reinforced
To employ the latest methods of producing
concrete. WVU also is looking at more
low-cost composite train vehicle structures
advanced composite materials, including car-
and to keep costs down, carbon fiber will be
used selectively, with high-performance fiber- bon fiber and hybrids. Another option is to
use an aramid fiber tendon, and candidates
glass the main reinforcement for sandwich
include Teijin's Technora, Akzo Nobel Fiber's
pane~, possible low-FST (fire, smoke toxicity)
matenals for the vehicle interiors. Twaron and Du Pont's Kevlar.
Regarding the guideway, they favor a U-
shaped configuration and either reinforced
43.4.21 MEDICAL, HEALTH AND SAFETY
co~crete throughout or with a composite
PRODUCTS (SEE CHAPI'ER 44)
gnlle at the center of each guideway beam. To
reinforce the concrete, they recommend using BioMechanical Materials Corp. (Camarillo,
post-tensioned FRP tendons in the magnetic CA) and Composiflex (Erie, PA) are among
field and closest to the path of the cars and those selling carbon fiber/acrylic and carbon
more conventional steel rebar prestressing in fiber / urethane sheet stock to orthotic equip-
the lower regions of the supporting columns, ment distributors. They in tum sell this
well away from the high magnetic field to material to hundreds of small local laboratories
a~~id raising the maglev drag coefficient sig- ~t fabricate components tailored by ortho-
nificantly. Magneplane International predicts tists for each patient, including such items as
that the guideway will be the dominant cost arch supports, knee braces and custom sports
component, so it has first design priority. braces. BioMechanical Materials among others
Their target is not to exceed $6.2 million per also sells large quantities of composite sheet
944 Commercial and industrial applications of composites

and fabricated parts to sports shoe manufac- older heavy tungsten studs to impact (rather
turers, such as Brooks and L.A. Gear, for than scrape) the road surface. A Duralcan
enhancing performance and the marketability MMC with 15 v / 0 alumina in 6061 aluminum
of the top-end shoes. Hexcel supplies a unique is being substituted to limit road damage
hybrid E-glass and carbon fiber fabric rein- because of its lower weight. MMC wire is
forced epoxy 'chip' for arch supports in Reebok being shipped at the rate of 'several tons per
sports shoes. For electrical hazard protection, month' for this application. Another interest-
Composiflex produces a hybrid composite of ing wire application is for flame spraying
fiberglass/ carbon fiber/epoxy to fabricate where the MMC adds to service life; alumina-
lightweight toe plates for safety boots and reinforced aluminum wire is sprayed onto
puncture-resistant insoles, replacing the tradi- steel as an antiskid surface in marine environ-
tional steel inserts. This company gets its name ments such as ships, piers and oil platforms.
from their product with multiple plies of ori- Previously, pure aluminum was flame sprayed
ented fiber composite which can control onto a steel surface before painting, serving as
bending to certain directions. a corrosion barrier.
Among the many other applications in
Table 43.6 Medical products work or which have been tested to date are the
following: extruded truck frame channels [C-
MRI scanner cryogenic tubes" channels of about 304 X 76 x 7620 mm (12 X 3 x
X-ray tables" 25 in)], motorcycle drive sprockets, a propor-
Underwater therapy treadmill
Wheelchairs"
tional valve for a truck lift gate, compressor
Orthotics - shoe arch, neck & knee braces" piston-ring inserts, sporting goods such as golf
Prostheses - artificial limbs" clubs and brake components for standard and
External fixaters - orthopaedics" high-speed rail cars. Other promising applica-
Rehabilitation baths tions are automobile brake rotors (initial
Medical centrifuges* orders are now being used for early produc-
Hip implants* tion at this time) to reduce weight by replacing
Heart valves"
Dental/prosthetic cements
cast iron. Further component applications
X-ray file cassettes* likely to emerge include pickup truck drive
Medical sensors* shafts, engine cylinder liners, automotive sus-
Walking canes" pension arms, aircraft camera mounts,
Artificialligaments* electrical housings, and pistons for automotive
engines, piston inserts that reduce the energy
required for high-speed reciprocating motion
43.4.22 METAL-MATRIX COMPOSITES (MMCs)
in large gas compressors, and tubing and forks
Duralcan USA (Novi, MI) markets low-cost for high-performance racing and mountain
MMCs reinforced with alumina or silicon car- bicycles.
bide particulate reinforcement. One of their Recyclability is a major market issue for
most successful applications is the use of rein- high-tonnage applications of many materials,
forced aluminum wire for tire studs. Finland especially now in Europe. Duralcan claims
and Sweden (and many USA states) banned that their MMCs have a potential advantage
the older steel-jacketed tungsten carbide studs over many other materials, including poly-
from all-weather/snow tires due to road wear. mer-matrix composites, since MMCs can be
Many people assumed that studs degrade recycled as a composite material or easily
highways prematurely because they wear reclaimed into high-quality aluminum by
away road surface. The real culprit is centrifu- melting and removing the reinforcing parti-
gal force acting on the studs that causes the cles. Whether recycled or reclaimed, Duralcan
 Applications review 945

USA can achieve virgin material properties in Composite trailers could haul more fuel or
new parts. For example, they have recycled would be allowed on roads with weight
pickup truck drive shaft extrusions up to four restrictions or service pads not capable of han-
times, while retaining original property per- dling the heavier steel trailers. Manufacturers
formance in subsequent parts. In the of the current steel tube trailers include Christi
aluminum casting process, 30-40% of the Park Industries in McKeesport (PA) and Trend
process metal is not in the part but in the gates Fuels in Austin (TX). Certification by the
and risers, and all of this material is now being Department of Transportation (DOT) is
recycled. required to allow use of large composite stor-
Duralcan USA's parent company is Alcan age tanks on highways if loaded with
Aluminium Ltd in Canada Uonquiere, compressed natural gas to 25 MPa (3600 psi) or
Quebec), and their plant has an ultimate higher. Projections for the composites poten-
capacity for these products of around 11-16 tial in this one application range from 36-45
million kg (25-36 million lb) per year. In some million kg (80-100 million lb).
applications (bikes and drive shafts, for exam-
ple) these MMCs could readily displace some
43.4.24 MOLDING COMPOUNDS FOR
growth opportunities now seen for polymer
INJECTION MOLDING
composites.
MMCs from other producers also show A steadily growing segment of the advanced
great promise in the field of electronics. Multi- composites market is a variety of injection
chip modules (MCMs) require high molded parts made with carbon fiber, or
dimensional stability, thermal conductivity hybrids with fiberglass or aramid reinforce-
and thermal shock resistance. Mounting them ments. LNP Engineering Plastics, Inc. (a unit
on MMC provides these benefits, especially of Kawasaki Steel based in Exton, PA) partici-
for high-density avionics applications on air- pates in many current applications in the
craft or for radar and other systems on board commercial and industrial sectors. (See
ships. For maximum thermal conductivity in Section 43.4.3.) There is a trend toward
electronic backplanes, Amoco Performance increased use of the higher-temperature ther-
Products's K-1100 pitch-based carbon fiber moplastics, and at LNP these include
might be used. polyethersulfone (PES), polyarylsulfone
(PAS), polyetheretherketone (PEEK),
polyphenylene sulfide (PPS) and polyetherim-
43.4.23 MOBILE STORAGE MODULES FOR CNG
ide (PEl, or Ultem). Their long fiber
Aerojet and Pacific Gas & Electric have teamed compounds called Verton offer performance at
to demonstrate a low-cost, lightweight com- the high end of the mechanical performance
posite Mobile Storage Module that is a large spectrum for compounds.
NGV fuel tank made of carbon fiber or fiber- RTP Company (Winona, MN), Comalloy
glass composite overwrapped aluminum used International (Nashville, TN) and
to transport CNG from the utility to the user. Compounding Technology, Inc. (Corona, CA)
These tubes are currently 381 mm (15 in) manufacture molding compounds made with
diameter by 6096 mm (20 ft) long, and several the high-performance thermoplastic resins
of these would be mounted in a truck trailer listed above plus large amounts of polycar-
that can be used to transport, store, and bonates and nylons/polyamides, much of
deliver natural gas fuel to fixed loading areas. these for electrical and electronic insulating
These modules would be moved between fleet components. A smaller part of the business is
refueling stations, thereby avoiding the neces- nickel-coated carbon fiber for added conduc-
sity of installing a costly pumping system. tivity, and they also use some stainless steel
946 Commercial and industrial applications of composites

fiber to make compounds suitable for various sizes for reinforcing the necks of gui-
EMI/RFI Shielding. DSM Engineering tars or banjos, plus for making bridge blanks,
Plastics, Inc. (Evansville, IN) and others report fingerboard underlays, stiffening braces and
a noticeable trend toward using higher-tem- other components requiring extra strength,
perature glass-filled thermoplastics for stiffness and/or creep resistance. Other instru-
components used in the electronics industry, ments now offered in carbon fiber and other
mainly for connectors. Currently, the liquid composite materials, all claiming better tone
crystal polymers (mainly Vectra from Hoechst and clarity, are violins and bows, bases, drums
Celanese and Xydar from Amoco) are widely and sticks, sousaphones/tubas (an old applica-
used for electronics applications. Having been tion in FRP), harps, pianos and more.
part of the Akzo organization for several years
(and before that, Fiberfil), they have worked
closely with Akzo Nobel Fortafil Fibers Table 43.7 Consumer products
(Rockwood, TN) to develop special grades of
Camera tripods"
chopped and milled carbon fibers with suit-
SCUBA tanks"
able finishes. Other suppliers include Toho Banjos"
Carbon Fibers and Hexel. Heavy-tow fiber Sousaphones/tubas
(50K filament count and greater) made from Harps"
commercial-grade PAN precursor is well Pianos
suited to making lower-cost chopped fiber. Umbrellas"
Usage of carbon fiber has been going up every Attache/brief cases"
Cassette tape/CD player cases"
year and has reportedly passed annual con-
Exercise equipment*
sumption of 2.25 million kg (5 million lb) per Art work bracing systems"
year that would translate into compound vol- Guitars"
umes five to ten times that. Violins and bows"
Bases"
Camera cases"
43.4.25 MUSICAL INSTRUMENTS Eyeglass frames"
Caskets
Kaman Aerospace (Bloomfield, CT) offers the
Binocular /monocular bodies"
Ovation Instruments guitar. The top of the Running shoe arches"
Ovation series is called the Adamas and its Audio earphones"
excellent tone is attributed to the use of a com- Drums and sticks"
posite sandwich construction for the top. The
outer and inner plies are single layers of 5-mil
unidirectional carbon fiber/epoxy tape, while
the core is 0.762 mm (0.030 in) thick birch
43.4.26 OIL AND GAS APPUCATIONS
veneer. Fiberglass composites are also used for
other Adamas components, including the Conoco, Inc. (Ponca City, OK), Shell (Houston,
bowl. These are mostly handmade guitars so TX) and others are involved in oil and gas
annual production is small. A producer of clas- exploration to production depths typically to
sical guitars is Kuau Technology (Maui, 619-928 m (2000-3000 ft) and to even more than
Hawaii) with carbon fiber/epoxy and the RTM 4 km (15 000 ft). Conoco is the first company
process is their preference. The Luthier's with one of the new tension-leg platforms
Mercantile (Healdsburg, CA) catalog for (TLP) in use. Shell is building such a platform
stringed musical instrument producers gener- and planning several more in partnership with
ates 99% of their business via mail order. They Exxon and British Petroleum. The oil industry
offer carbon fiber/epoxy bar and sheet stock in has a 'steel and concrete bias' which is the result
 Applications review 947

of many years of experience, relatively low ini- closely with IFP and Aerospatiale Space &
tial cost, ample supply, current design codes, Strategic Systems are Lincoln Composites
etc. However, new materials are the major tech- (Lincoln, NE and teamed with Aarding BV in
nology thrust needed in offshore structures The Netherlands) plus Coflexip & Services,
because the industry is forced to spend too Inc. (Paris). Coflexip is reportedly the leading
much money fighting corrosion of steel. producer of sub-sea pipelines and risers to
Current offshore structural applications of transport oil and gas products at offshore pro-
composites include low-pressure pipe, grat- duction facilities. The initial product of the
ings, handrails, equipment covers or Lincoln-French joint venture was a composite
enclosures and ladders. Future application TLP riser pipe that connects well-heads to
development is likely to focus on FRP for tethered floating platforms. Risers provide a
high-pressure piping, pressure vessels, accom- casing to protect the well bore and to carry the
modation modules and blast walls. FRP oil to the surface. Tethers will be 244.5 mm
products (mainly E-glass reinforced) provide (9.60 in) diameter tubes, with a wall thickness
installation savings ranging from 10-70%, of about 13.5 mm (0.53 in) and 12-17 m
improved life-cycle costs, enhanced safety, (40-55 ft) long.
30-70% weight savings and environmental
advantages, according to studies by Conoco. It
is expected that a variety of composite materi- Table 43.8 Oil and gas exploration
als will be used for offshore structures, with
Tension leg platform tethers*
carbon fiber, aramid and fiberglass reinforce-
Fire shields - drilling rigs
ments, thermosets and thermoplastic matrix Drilling tubes*
resins and some hybrids. Sucker rods/slides
The Institute Fran~ais du Petrole supported Oil well logging probes*
development of advanced composite tubes Tribological bushings, seals, '0' rings*
developed by Aerospatiale for risers on off- Drilling motor shaft
shore stationary drilling platforms and tension Drill casing
leg platforms. These tubes are hybrids of high-
strength fiberglass and carbon fiber and are
offered in diameters ranging from 76 to 300 mm
(3 to 12 in) and in lengths up to 24 m (80 ft). Longer-term commercial application growth
Internal working pressures are in the 34--103 in this sector for advanced composites should
MPa (5000-15 000 psi) range, tensile loads to include moorings and tendons, drilling and
90000 kg (100 tons) will be accommodated by production risers, drill pipe and tubing and
a special high-strength threaded steel alloy some pressure vessels. With platform construc-
coupling, and the design life can be as high as tion costs exceeding $1.5 billion per unit, there
20 years. Typical drilling risers will be up to 10 could be a substantial dollar potential for the
times the height of the Eiffel Tower [up to 3000 various composite applications. One producer
m (9800 ft)] and weigh over 2.2 x 1()6 kg (2500 estimates the current usage of steel to be as high
tons). Syntactic foams control buoyancy and as 1.5-1.8 x 1()9 kg (1.7-2.0 million tons) used for
reduce the effects of sea loads from waves, oil- and gas-well downhole tubulars. Much of
currents and storms. The case for composites that could be displaced by composites (at one
is clear, as weight savings could exceed fifth the weight) because of their advantages in
544000 kg (600 tons). A 900 kg (one ton) high specific strength, fatigue resistance, corro-
reduction in the offshore superstructure or the sion resistance, lighter weight, low coefficient
risers results in a savings of 272 kg (3 tons) of of thermal expansion and reduced installation
displacement load on the platform. Working cost (due to lighter weight)
948 Commercial and industrial applications of composites

43.4.27 PIPELINE REHABILITATION 43.4.28 POLES - POWER DISTRIBUTION AND

A high-volume application for Kevlar 49 and LIGHTING
other aramid fiber reinforcement has been Shakespeare (Newberry, SC) developed fila-
developed by Insituform of North America, ment wound distribution poles that were
Inc. (Memphis, TN) with the help of Du Pont. installed by Montana Power Co. These fiber-
A patented product called Insituform has been glass/polyester poles weigh half as much as
used extensively for years in repairing and wood; material efficiency is improved through
generally 'rehabilitating' old large-diameter, the use of stitched fabric reinforcements as
gravity-flow and pressure pipes. Insituform well. Weight savings enhance the ease of
can repair or replace a damaged pipe very installation in remote areas through use of
rapidly to minimize downtime. The process smaller helicopters than required for wood
was developed in the UK many years ago but poles (Fig. 43.5). The 355 mm (14 in) diameter
most of the recent progress has reportedly poles are generally 12-14 m (40-45 ft) tall.
been made in the USA. A special needled poly- Another major advantage of FRP poles is the
ester felt reconstruction tube (called elimination of chemical treatments required to
Insitutube) is coated on the outside (typically preserve wood poles. An 80-year service life is
with a thin layer of polyurethane), and is cus-
tom engineered and manufactured to fit the
damaged pipe exactly. It is impregnated with a
liquid thermosetting resin (polyester, vinyl
ester, or epoxy as appropriate for the chemical
environment to be experienced) and lowered
into a manhole through an inversion tube. One
end of the Insitutube is attached to the lower
end of the inversion tube elbow. The resin-
impregnated tube is turned inside out using
cool water so that the smooth coated felt exte-
rior then becomes the new interior surface of
the pipe. Circulation of hot water through the
tube then cures the resin-impregnated felt.
This repair process typically is performed
on an old water or sewage pipe that has been
buried for many years and which is starting to
leak. The whole procedure can be done with-
out digging up or otherwise disrupting the old
pipe. Kevlar 49 is used to reinforce the poly-
ester felt in the hoop direction. This reduces
the felt thickness and the total weight, increas-
ing the inside diameter of the repaired pipe,
and generally adds both structural strength
and rigidity. Also involved is Shell Chemical
(Houston, TX) where a special quick-curing
epoxy is being developed for this application.
Dow Plastics (Midland, MI) supplies resins to
another company in this business, In-Liner Fig. 43.5 Installing lightweight composite utility
Corp., with their cured-in-place-pipe (CIPP) poles by helicopter. (Courtesy of Amoco Chemical
process, among many users. Co.)
 Applications review 949

targeted. In a related development, Composite fiberglass knitted multiaxial fabrics from

Power Co. (Las Vegas, NY) fabricates FRP util- Brunswick Technologies, Inc. (Brunswick, ME)
ity poles using pultrusion. Prospects for both and vinyl ester resin from Dow Plastics
companies are enhanced because of the envi- (Midland, Ml). The process is called Seemann
ronmental concerns when using treated wood Composite Resin Infusion Molding System or
poles and the cost escalation of older products SCRIMP. Co-owners of the process technology
that bring them closer to the current premium that can infuse preforms of 6 m by 2 m (20 ft by
costs of composite poles. 7 ft) in only 35 mm (1.4 in) are Hardcore,
The California DOT (Sacramento, CA) is Seemann Fiberglass and TPI (also see rail item
potentially interested in the use of composites above). They license this process technology to
to build 53 m (175 ft) tall light standards for others. The USA Navy has been assessing this
highway interchanges. Caltrans currently uses technology for years in the development of
galvanized steel but these must be replaced major ship structures, including masts.
periodically and are difficult to erect and Composites sheet pile from companies such as
maintain. Among the firms that have built tall Creative Pultrusions and others are also likely
steel light standards are Ameron, Union Steel, to experience a major growth trend.
Valmont and Shakespeare. The hope is that
current composite technology for lighting and
43.4.30 RAILROAD ROLLING STOCK
power poles and for sailboat spars to 55 m (180
COMPONENTS
ft) height can be expanded into such highway
infrastructure applications. Burlington Northern Santa Fe Railroad (Ft.
Worth, TX) is a very innovative company and
plans use of composite overwrapped air bot-
43.4.29 PORTS AND HARBORS
tles for use on locomotives. They favor the
A potentially significant new application with lightest version that would use a plastic liner
major potential for tonnage of FRP is that of and a carbon fiber / epoxy overwrap to replace
dock and wharf pile fenders. Du Pont fabri- heavy steel tanks for improved corrosion resis-
cated composite fenders for a systems tance and easier/safer installation by work
company. Six developmental fenders were crews. Weight saved on railroad rolling stock
installed in mid-1994 in the Chesapeake Bay also increases the capacity for revenue genera-
area. They replaced multiple sets of fender pil- tion from hauling increased cargo. Lincoln
ings in 10-pile clusters that had steel fender Composites and Compositek fabricate such air
surfaces. The treated wood pilings only last bottles while Dresser Industries' Nil-Cor
about seven years on average, the steel plates Operations makes a variety of composite
are drastically deformed and rust relatively valves. Most use fiberglass/vinyl ester con-
quickly. struction; others are made of carbon
The FRP fenders with UHMW polyethylene fiber /vinyl ester, fiberglass/ polysulfone and
skinned impact surfaces to date have better fiberglass/polyphenylene sulfide composite.
withstood the impact of a 159 000 kg BNSF also seeks to introduce composites on
(350 000 lb) ferry boat (eight to ten times/day) existing or new freight cars. A prototype com-
than did the wood/ steel system, with no dam- posite knuckle is in the design stage for
age reported in over two months as of this extended field tests. This is likely to be a
writing. The lightweight composite fenders hybrid featuring carbon fiber/epoxy with
were installed in only 3.5 h with a cherry steel inserts to protect the high-wear areas.
picker while the steel and wood system Another industry program is development of
require a barge crane at much greater expense. an FRP cargo door to be produced using a
Composite fenders are fabricated using resin infusion process called SCRIMP offered
950 Commercial and industrial applications of composites

by Hardcore Composites (New Castle, DE), reportedly buy most of their carbon
Seemann Fiberglass (Hanrahan, LA), and TPI fiber / epoxy rolls from Lincoln Composites.
(Warren, RI). Shafts from about 2-10 m (6-33 ft) long are
Composite auto hauler composite cars with fabricated using hybrid carbon fiber and fiber-
all-pultruded components of fiberglass/poly- glass in epoxy which are about one-fifth the
ester have been developed by W. Brandt weight, operate at faster speeds and use less
Goldsworthy & Assocs. (Torrance, CA) along energy than the conventional steel rolls.
with Alcoa Composites and Stoughton Fiberite (Wilmington, DE), Epoch Industries
Composites (Brodhead, WI) for the Union (Garland, TX), Addax (Lincoln, NE), and
Pacific. Many other freight car development Quality Composite, Inc. (Sandy, UT) are
programs are under way at Trinity Industries among others that fabricate such products.
and elsewhere. Corrosion resistance and light Amalga Composites, Inc. (Milwaukee, WI) is a
weight are the two primary benefits of com- more recent entry into the filament winding of
posites that are sparking this interest. carbon fiber/epoxy tubes for the industrial
rolls used in paper and plastic film converting
machinery and reportedly is a major supplier
43.4.31 RAILWAYS AND TUNNELS
to American Roller (Bannockburn, IL). They
Besides pultruded and filament wound com- use Grafil's 33 msi carbon fiber but leave the
posite cableways, ducts and piping, there is a finishing and coating operations to American
vast array of associated molded parts for Roller. According to ARC, there are many
trackside, station equipment and furniture. potential applications for composite rolls.
Development of fiberglass reinforced pheno- These include replacing the steel rolls com-
lic-matrix composites, with their superior FST monly used in winder core, winder shaft, idler,
properties, has become an important contribu- layon and dancer/transducer positions to
tor to the increasing acceptance of composites manufacture film, flexible packaging, foils,
for interior components. European countries magnetic audio and video tape and photo-
now virtually mandate using phenolic resin and pressure-sensitive papers on printing
composites for underground railcar and sta- presses, laminators, coaters and blown film
tion components. Some of ICI's Modar lines. However, selecting the right composite
acrylic-modified polyester and also the less for each of these positions is not quite as sim-
flammable phenolic-based composites were ple as ordering an off-the-shelf product. ARC
used in the Channel Tunnel, mainly for electri- technical and marketing people make a special
cal cable tray. The market for composites in point of explaining why each composite roll
railways and tunnels could double during the should be customized to the end-user's spe-
next few years due mainly to the greater atten- cific application and operating parameters,
tion being paid to safety, including improved especially because of the many types of com-
fire, smoke, and toxicity performance (reduced posite materials and structural designs
FST). available on the market today.
Another filament wound tube fabricator for
this application is SpyroTech Corp. (Lincoln,
43.4.32 ROLLS AND AIR SHAFfS
NE). Also, TPI (Warren, RI, and formerly
Industrial rolls and expanding air shafts used called Tillotson Pearson) has been using fila-
inside cardboard core tubes on winders to take ment winding to make fiberglass/epoxy and
up paper, films and foils will be major growth carbon fiber/epoxy tubes for paper mill rolls.
applications in the industrial composites sec- It is noteworthy that Mitsubishi Rayon in
tor. Tidland Corp. is said to be the world's Japan is one of the largest suppliers of such
leading supplier of core shafts of all types and rolls in the world. These suppliers market the
 Applications review 951

major benefit of hybrid carbon fiber-fiberglass Table 43.9 Industrial equipment

and 120°C (250 0 P) cure epoxy composite rolls
and shafts compared to steel and aluminum Coordinate measurement/3-D gaging"
shafts, namely a 50% weight reduction. Magnet support rings"
Torque tubes"
Precision balancing allows higher operating Paper folding plates"
speeds than with metal rolls. Advanced com- Coating doctor blades"
posite rolls exhibit 20% less deflection under Textile bobbins"
loads. Also, composite rolls offer superior Instrument housings
fatigue resistance for increased service life and Pump impellers"
improved damage tolerance versus steel or Crane components
Backhoe components
aluminum. Forklift components
Polygon (Walkerton, IN) is a pultruder and Mixer drive shafts"
braider active in fabrication of such industrial Containers
rolls. Composite Development Corp. (West Chemical mixer paddles"
Wareham, MA) also supplies to the converting Liquid level measuring sticks
industry that wants to use thinner gage Surveying transit tripod
Industrial rolls - paper, film, foil, printing"
paper / film/ foils yet process at higher speeds, Water tanks
so the lightweight, low-inertia rolls produced Chemical filters/frames
using carbon fiber composites with the fabric Radio-controlled vehicles
roll-wrap process are rolls that offer improved Mirrors"
handling characteristics (Fig. 43.6). Heavier Cryogenic vessels"
rolls such as plate cylinder rolls (used to Core tube air shafts"
Generator housings
mount printing plates) offer excellent weight Welding robot arms"
savings that are important in a process with Laser resonator spacer tubes"
constant and many job changeovers. Some Pump casings
producers of rolls use braided reinforcement Speaker housings"
preforms from companies such as A&P Signage
Technology (Covington, KY). Textile looms - rapiers, heddles, shuttle cocks,
picker sticks"
Ball and butterfly valves
Conveyor belt support shafts
High-pressure hydraulic pipe"
Pneumatic piping"
Uranium enrichment centrifuges"
Cooling tower drive shafts"
Tote trays/Storage bins
Tooling"
Optical benches"
PCB conveyor components
Coil springs"
Underwater light housings
Light bulb sockets
Propellers
Fan blades
Ladders
Safety shoe toe guards
Industrial shaft couplings
Broom handles
Window cleaning squeegee handles
Fig. 43.6 Composite industrial rolls. (Courtesy of
Cherry picker lifts
Composite Development Corporation.)
952 Commercial and industrial applications of composites

The considerable weight reduction com- Institute of Standards and Technology (NIST)
pared with steel rolls means less strain and in Boulder, CO, and the Phillips Laboratory at
fatigue and reduced injuries for the operators. Edwards AFB, CA, assist by conducting the
For example, a steel roll that is slightly over low-temperature testing work. Competing
2489 mm (98 in) long, 152 mm (6 in) O.D., and technologies for SMES include pumped water
127 mm (5 in) LD. would weigh approximately storage, batteries, and potentially, flywheels.
188 kg (239 lb). A fiberglass-reinforced compos- Recent work has been focused on develop-
ites roll of the same dimensions would weigh ing and testing the prototype of a full-size
27 kg (61 lb), while a carbon fiber composites mockup of a SMES unit. E-glass fiber rein-
roll of this size weighs only 22 kg (49Ib). forced plastic curved, pultruded vinyl ester
resin beams were developed to simultane-
ously provide essential electrical insulation
43.4.33 SMC - TOUGH, HIGH-PERFORMANCE
(between conductor coils and between the
Premix, Inc. (No. Kingsville, OH) and their conductor and the ground) and the conductor
Quantum Composites division (Midland, MI), support structure.
focus on 'ultra-high strength and toughness' Pultruded FRP was chosen because it pro-
Lytex SMC that is offered with carbon fiber or vides reasonable structural performance and
fiberglass reinforced epoxy or vinyl ester the required electrical insulation capabilities at
resins. Present nonaerospace uses for this cryogenic temperatures for an affordable cost.
unique SMC include a prosthetic foot, an auto- Test articles were fabricated and then the
mobile brake component, an electrical mechanical, thermal, and electrical properties
connector and more. Projected applications were measured over the temperature range of
include automotive suspension and structural 4-295K. FRP beams, pultruded by Compositek
chassis components, recreational-vehicle high- (of Brea, CA and now part of Kaiser
impact parts and other applications Aerospace), met the goal of comparable perfor-
demanding toughness and excellent dynamic mance to currently used and more expensive
fatigue resistance (including aerospace). materials such as G-lO laminate produced with
fiberglass fabric and epoxy resin. The tensile
failure strains for the materials tested using liq-
43.4.34 SUPERCONDUCTING MAGNETIC
uid helium ranged from 1.2% at 295K to 1.9%
ENERGY STORAGE SYSTEMS (SMES)
at 4K. The pultrusion machine was built by
Composites play a key role in the insulation Pultrex Ltd (Clacton-on-Sea, UK). Johnson
and structural systems of the SMES Program Industries Composite Textile Reinforcements,
at General Dynamics Space Magnetics, now a Inc. (of Phenix City, AL) supplied the stitched
unit of Martin Marietta. Years of development fiberglass preform materials, which ranged
may now offer dramatic potential for carbon from 9 to 18 plies in thickness.
fiber and glass fiber composites in the future. SMES develops large radial loads when
The purpose of SMES is to economically store charging. This requires strong coil support
electrical energy from power production late straps, which are integrated by General
at night and then to retrieve it during peak Atomics (San Diego) and which were filament
demand periods during the day. wound by Lincoln Composites (Lincoln, NE).
Heading the team is Bechtel National, Inc. The short straps look like a fan belt about
as system integrator. General Atomics is 127 mm (5 in) wide and 762 mm (2.5 ft) long
responsible for the coil supports and was the and are wound with S-2 glass fiber prepreg.
designer and fabricator of the first SMES unit Larger straps, about 89-1200 mm (3.5-4 ft)
to be operated successfully (1983) by an elec- long and 100 mm (4 in) wide, are wound using
tric utility. Bonneville Power System. National carbon fiber prepreg.
 Applications review 953

The SMES Engineering Test Model (SMES- and is produced using 68 w / 0 AS4 carbon fiber
ETM) consists of a two layer, radial, helically filament wound in PEEK thermoplastic matrix.
wound, 96 m (315 ft) diameter solenoid coil They have reportedly produced well over 45
that has been designed to store a minimum of 360 kg (100000 lb) of parts with this material
20 MWh of electrical energy in its magnetic 'under production conditions, to aerospace
field. The coil stack requires 66 layers of PRP requirements'. Also offered are parts molded
curved beams for the support structure or 825 from their injection-molding grade. Polygon
24 m (80 ft) sections weighing 907 kg (2000 lb) offers a line of 'Poly Lube' composite bearing
each. A full-sized, 20 MWh ETM demonstra- materials, hybrids of fiberglass and Teflon. Bal-
tion unit will require about 725 000 kg (1.6 Seal Engineering Company (Santa Ana, CA)
million lb) of FRP composite. A full-scale test markets carbon fiber and graphite-filled Teflon
model of SMES will consist of a large, cylin- fluid seals and '0'-rings. The primary benefits
drical coil 91 m (300 ft) in diameter sunk into a are dimensional stability and a low coefficient
9 m (30 ft) deep trench in the ground. Stored of friction under elevated temperature and
energy can be retrieved slowly or almost pressure, with no 'extrusion' of the resin
instantaneously, and the technology is consid- matrix. Carbon fiber adds to the tribological
ered a promising method of storing electrical value of the rings through prevention of adhe-
energy for peak load demands. sion of the Teflon to the bearing surface.
Bechtel estimates that the potential SMES Tiodize (Huntington Beach, CA) sells self-
market in the USA alone could be as high as $3 lubricating composites based on carbon fiber in
billion/yr. The Electric Power Research a high-temperature resin used for bushings,
Institute in Palo Alto (CA) and the SMES thrust bearings, rod-end bearings and ball-
Utility Interest Group (SUIG, Dallas, TX) are bearing retainer rings. They emphasize the
promoting the commercial development of the advantage of using these materials compared
SMES technology because it looks promising to aluminum in retaining specific modulus at
for electrical load leveling, increasing electrical temperatures to 316°C (600 0 P).
system reliability, savings in fossil fuels,
reduced air pollution, emergency backup
43.4.36 URANIUM ENRICHMENT
power, virtually instantaneous response and
CENTRIFUGES
the ability to store energy from intermittent
sources such as solar and wind generation of Louisiana Energy Services is a joint effort of
electricity. EPRI estimates that by 2010, storage three USA electric utilities, Duke Power Co.,
could usefully provide as much as 10% of the Louisiana Power & Light Co. and Northern
nation's generating capacity. Using storage to States Power Co. Along with the internation-
save power generated at night when power ally known engineering firm of Fluor Daniel,
plants normally run at very low rates some Inc., Urenco will build and help operate a ura-
day could potentially save the cost of building nium fuel enrichment plant for these utilities.
many new generating plants at far higher cost Urenco is a European organization with close
than adding SMES storage capacity. to 15 years of experience operating commer-
cial uranium enrichment plants in Germany,
the UK and The Netherlands. Together, the
43.4.35 TRIBOLOGICAL APPLICATIONS - SEALS
three utilities own and operate 11 commercial
AND 'O'-RINGS
nuclear power plants. In the enrichment
EGC Corporation (Houston, TX) is a custom plant, high-speed ultracentrifuges will be
fabricator of molded components for oil field used to enrich the uranium and hopefully can
and refinery applications. Their 'wear-resistant supply nuclear plant fuel at a price lower than
composite' is called XC-2 tribological material is being projected by DOE, according to the
954 Commercial and industrial applications oj composites

program plarmers. Carbon fiber/epoxy fila- while other blade manufacturers had to fold
ment wound composite tubes are to be used when the market started to collapse several
for the high-speed rotors but most of the tech- years ago.
nology details remain proprietary to Urenco. Sandia National Laboratories (of
At this writing, Louisiana Energy Services Albuquerque, NM) developed advanced wind
continues with the approval process for the turbine technology using the vertical axis
required environmental impact report in sup- approach and signed a technology agreement
port of their license application. They hope to with Flo-Wind Corp. (Pleasanton, CA). Flo-
start construction soon, and have the new Wind has used extruded aluminum blades but
plant in full operation in the late 1990s. plans use of composite blades. Flo-Wind cur-
Similar composite centrifuge rotors are rently operates 'over 500' such turbines with
believed to be standard design for European aluminum blades currently in operation on
and Japanese uranium enrichment work. two California windfarms, likely candidates
for retrofit with composite blades. Glasforms
(San Jose, CA) is one of their prime suppliers
43.4.37 VALVES
for composite parts.
Dresser Valve & Controls Division, Nil-Cor Competition continues to develop overseas.
Operations, (Alliance, Ohio), sells several Japan's Tomen Corp. has joined with
types of composite valves to the chemical American and British firms to build several
industry. Over 100 000 Nil-Cor advanced com- large wind power farms in Wales and
posite valves are currently in service at over Cornwall, UK, operated by a company called
100 USA paper mills. Their current valve prod- British Wind Farms, based in London.
uct line includes fiberglass and carbon fiber Included in their plans was the construction of
reinforced PPS, polysulfone and vinyl ester in 300 windmills with a total generating capacity
several sizes and types. Fiberglass reinforce- of 77 000 kW. Many European countries are
ment and Dow Chemical's Derakane 470 vinyl making sizable commitments to increased use
ester resin (among others) are used in ball and of windpower, including Denmark, Germany,
butterfly valves for the chemical industry. Holland and Spain.
The National Renewable Energy
Laboratory, a unit of the Department of
43.4.38 WIND TURBINE BLADES
Energy (formerly called the Solar Energy
TPI, Inc. located their composite wind turbine Research Institute), is in Golden, CO. They are
blade production in Portsmouth, RI. TPI interested in all types of alternate energy sys-
makes fiberglass/vinyl ester blades regularly tems, including battery-operated vehicles,
for large operators of wind energy farms in the engines that can use different types of fuels,
USA. The size of the individual generating including hybrids and biofuels. NREL spon-
units has gradually increased from the 50 and sored an effort called the Advanced Wind
I

100 kW sizes with 8 m (27 ft) composite blades Turbine Next Generation Preliminary Design
that were most common several years ago to Project', with the goal of bringing wind power
the larger 300-kW units that are now in generation costs down from the current 7-8¢
demand. Each of the three new size blades is to below the 4-5¢ per kWh with coal genera-
17 m (56 ft) long and uses a sizable quantity of tion. Wind generators must be made of
material. Production rates have reached as materials likely to last at least 30 years, so the
high as 5000-6000 blades per year. Among the additional cost of using carbon or aramid
critical aspects of making large windmill fibers to reduce rotor blade weight and/or
blades is the need for proper balance. This is improve performance might be amortized
one reason that the TPI product has succeeded over a long period.
 Market data 955

Today there are over 15 000 windmills in at twice the growth rate of the Gross Domestic
California alone, with a total electrical generat- Product (GDP).
ing capacity of 1300 megawatts. To date, they Sectors that fall mainly into the commercial
have generated a total of over two billion kWh, and industrial applications arena, the focus of
replacing 3.5 million barrels of oil. However, this chapter, are noted next. The corrosion-
this is only 1.7% of the total power requirement resistant application sector for chemical
of the state of California during the same storage, process tanks, pipe systems, cooling
period. There are more windmills overseas tower components, wastewater treatment,
where about 70% of the total has been installed. pressure vessels, and pollution control equip-
ment grew 5.9% in 1993. This is because of
their ties to capital spending and substitution
43.5 MARKET DATA
for other materials, and another 3.7% gain is
A frame of reference on market size is useful in likely in 1994. Electrical and electronic applica-
assessing composites growth prospects. The tion usage grew at 5.7% in 1993 (plus 5.5%
Composites Institute of the Society of the more in 1994). Power poles and fiber optic
Plastics Industry issued a news release cable splice boxes are examples of growing
February 8, 1994 stating that USA shipments of areas here. The appliance, business equip-
fiber reinforced plastics (FRP) reached an esti- ment, and consumer product markets gained
mated 1.23 x 109 kg (2.73 billion lb) in 1993. This only 3.0% (plus 3.7% more in 1994) and 2.2%
was an increase of 6.9% over 1992 (this is almost (with another 4.2% in 1994) respectively,
entirely fiberglass-reinforced polyester, vinyl mainly from material substitution. Lastly, the
ester, and epoxy material but includes 'other' category grew at 7.1% in medical
advanced fibers composites). Projected 1994 equipment such as orthopaedic appliances
growth to 1.32 x 109 kg (2.90 billion lb) of com- and dental materials. The Composites Institute
posites represents an increase of another 6.2%. (CI) represents over 415 firms in the compos-
Overall growth in sales of composites continues ites industry and is the SPI's largest division.

Table 43.10 USA composites shipments: 1984-1994 (in million lb)

Markets 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
projected
(revised)
Aircraft/ aerospace/ 29 32 37 36 39 41 39 38.7 32.3 25.4 24.6
military
Appliance/business 123 133 137 141 150 151 153 135.2 143.2 147.5 153.0
equipment
Construction 430 445 456 506 495 470 468 420.0 483.0 530.0 575.0
Consumer products 143 142 149 167 169 158 165 148.7 162.2 165.7 171.2
Corrosion-resistant 310 295 291 329 349 335 350 355.0 332.3 352.0 365.0
equipment
Electrical! 189 191 201 214 230 229 241 231.1 260.0 274.9 290.0
electronic
Marine 309 335 340 413 452 405 375 275.0 304.4 319.3 332.8
Transportation 540 563 585 656 695 677 705 682.2 750.0 822.1 890.0
Other 80 82 83 75 80 76 79 73.8 83.4 89.3 93.9
Total 2153 2218 2279 2537 2659 2542 2575 2360 2551 2726 2895.5
Includes reinforced thermoset and thermoplastic resin composites, reinforcements, and fillers.
Source: SPI Composites Institute, 1994 Semi-annual statistical report, February 8,1994
956 Commercial and industrial applications of composites

43.6 PREDICTIONS Thus, one should look to these markets not

only for growth but also for technology and
Much of the long-term growth for composites is
new ideas.
likely to come from this miscellaneous sector of
the business. The most likely segments to expe-
rience large unit growth rates are CNG tanks, REFERENCES
flywheels, industrial rolls, distribution and
Most of the information provided in this
lighting poles, civil structures, consumer and
chapter was generated through attendance at
medical products and some other oil and gas or
trade shows, interviews with industry partici-
transportation components. Composites have a
pants, review of press releases, other
low market share in many industries, and com-
extensive reading of publications, and the
mercial and industrial applications should
experience of the author over 30 years.
continue to experience well above average
Therefore, since most paragraphs would have
growth compared with other composites, other
required almost a reference per sentence, we
material systems, and the total economic
elected to list below the best generic and reli-
growth rate on a worldwide basis. Such growth
able USA-based sources available to research
over the long term is believed likely to exceed
these applications and markets or to track
the 5-8% growth rate expected for 1994 over
their future activities, participants and
1993 shipments for these composites markets.
prospects.
Composites Institute of the Society of the Plastics
43.7 CONCLUSIONS Industry, New York, NY.
Because these sectors of business and technol- Composites News: Infrastructure and Advanced
Material newsletters. Loud, Steve, (ed.) Solana
ogy are highly fragmented many smaller Beach, CA: Composites Worldwide, Inc.
producers are likely to participate. The scale of Composites Industry Monthly and ACM Monthly
production is often not well suited to large newsletters. Burg, Martin, (ed.) San Diego, CA:
producers. Also, the cost for large producers to Composite Market Surveys.
pursue such fragmented businesses, pieces of Composites Technology and High-Performance
which may not be very large, could be higher ~omposites magazine. Hazen, Judith Ray, (pub-
per unit or unit weight than for their huge lisher), Denver, CO: Ray Publishing, Inc.
Performance Materials and Fleets & Fuels newsletters.
composites commodity businesses such as Piellisch, Richard, (ed.) San Francisco, CA:
automotive, marine and others. Much of the August Pacific Press.
innovation in material forms and process tech- piasticsBRIEF Reinforced Plastics Newsletter. Best,
nology also resides in this group of markets. James R., (ed.) Toledo, OH: Market Search, Inc.
COMPOSITE BIOMATERIALS 44
Shalaby W. Shalaby and Robert A. Latour

44.1 COMPOSITES AS A FAMILY OF phosphate-based fiber showed large increases

BIOMATERIALS in both strength and stiffness (Casper et al.,
1985; Lin, 1986). However, these families of
Although the use of high modulus metallic
composites lose a substantial fraction of their
devices for internal bone fixation has been suc-
strength while retaining greater proportion of
cessful, there is a call to develop nonabsorbable,
their stiffness after short term exposure to an
~olymeric, composite substitutes having elas-
aqu~ous physiological environment. Attempts
tic modulus approaching or slightly higher
to Improve the strength retention have
than those of bones. This is to alleviate the
included coating the composite material to
problem of stress-shielding and consequent
retard the transport of fluids to the
bone resorption as well as any concerns related
polymer/fiber interface (Kelley et al., 1988;
to toxicity of metallic ions as corrosion prod-
Andriano, Daniels and Heller, 1991) and sur-
ucts. (Gillett et al., 1984; Tonino and Folmer
face modification of the phosphate fibers with
1987). Furthermore, to avoid a second surger;
a siloxane film barrier (Andriano, Daniels, and
to remove a non-absorbable device and to
Heller, 1992). Growing interest in the inorganic
allow for gradual load transfer to healing
phosphate-based fillers led Andriano and
bones, completely absorbable composites were
coworkers to compare the biocompatibility of
proposed and evaluated by many investigators
several phosphate, fiber-reinforced polymers
(Casper et al., 1985; Daniels et al., 1990; Lin,
in a preliminary study (Andriano et al., 1993).
1986; TormaHi. et al., 1991; Vainionpaa, 1987).
Among the key types of totally absorbable Thus, phosphate fibers of calcium-sodium
metaphosphate (CSM), sodium-calcium-alu-
composites are those based on polyesters or
minum-polyphosphate (NCAP) and
polyorthoesters, reinforced with organic fibers
(Tormala et al., 1991; Vainionpaa et al., 1987) or potassium metaphosphate (PMS), with copoly-
mers of £-caprolactone and lactide or
calcium or sodium/calcium polymetaphos-
phates (Casper et al., 1985, Andriano et al., polyorthoesters have been used as organic
matrices. The CSM and NCAP fibers were
1993). A family of composites having organic
found to be acutely nontoxic in cellular tissue
polyester fibers as fillers (e.g. poly-L-Iactide
and whole animal evaluation.
reinforced with polyglycolide fibers, displayed
large increases in strength and modest
increases in stiffness (TormaUi et al., 1991; 44.2 COMPOSITE BIOMATERIALS: GENERAL
Vainionpaa et al., 1987). On the other hand,
polylactide reinforced with inorganic calcium Like the more traditional composites, those
classified as composite biomaterials contain
two or more distinct constituent materials or
Handbook of Composites. Edited by S.T. Peters. Published phases on a microscopic or macroscopic size
in 1998 by Chapman & Hall, London. ISBN 041254020 7 scale but not on the atomic level. Thus, fiber-
958 Composite biomaterials

glass-reinforced polymeric composites and include strength, creep stiffness, and fatigue
other reinforced resins are composites while characteristics. For example, for anchoring a
metal alloys are not. Interest in the former sys- metallic hip prosthesis a high modulus
tems was associated with the need to replace crosslinked, filled methacylate 'grouting' or
or augment biological tissues which are com- bone cement is usually sought. For cementless
posites in their own rights. For instance bone, hip prostheses, carbon-fiber reinforced high
skin and blood vessels are typical natural com- performance thermoplastic polymers, such as
posites. In both natural and man-made poly(ether-ether ketone), PEEK, or aromatic
composites (or composite biomaterials) the poly sulfones (PS) are being explored.
strength and modulus of the matrix are depen- Cartilage is a low-load bearing, natural
dent on the shape, stiffness and orientation of composite and its synthetic substitutes can be
the reinforcing materials as well as their adhe- made of reinforced elastomers. Elastomeric
sion to the matrix. High modulus fibers tissues such as skin and blood vessels are
uniaxially oriented in a low modulus matrix made primarily of collagen and other biopoly-
can produce stiff solid composites with maxi- mers which impart compliance and elasticity.
mum strength and modulus along the fiber Synthetic substitutes of these tissues are far
direction as in carbon fiber-reinforced epoxy from being perfected. However, a key consid-
resins. Complex high density natural compos- eration is the use of an elastomeric polymer or
ites include cortical bones, dentin, cartilage as design to provide inherent or engineering
well as wood. Should air be a distinct phase of elasticity, respectively.
the composite, a foam is produced, as in syn- Bioceramics and particularly hydroxyap-
thetic sponges or lung tissues and cancellous atite (HA) have been used or proposed for use
bone (Park and Lakes, 1992). in many orthopedic and dental applications
Early application of hard or dense synthetic (Heimke, 1989, 1990). However, the mechani-
composites as biomaterials includes the use of cal properties of HA were not sufficient for
reinforced acrylics as dental fillings. The com- demanding load bearing applications, such as
posite resins consist of a polymer matrix such fracture fixation or spinal fusion, and this led
as crosslinked methacrylate resin and a stiff, to its limited use in this area (Bostman et al.,
inorganic filler including barium glass or silica. 1989). To address this issue, Knowles and
The methacrylate resin can be based primarily Bonfield (1993) developed a glass-reinforced
on the glycidyl methacrylate derivations of HA with enhanced mechanical properties.
bisphenol A. The choice of such components is Utilizing glasses of the types xNa20-(1-x)
consistent with the repair site, i.e. dentin, P20 S and xCaO-(1-x) pps (where x = 0.2, 0.3
which in turn is a composite, made primarily and 0.5), a systematic study was conducted on
of collagen and microscopic inorganic crystals the effect of increasing network-modifying
of an apatite of calcium and phosphate. In gen- oxides in the glasses on the mechanical prop-
eral, the primary tissue of the teeth such as erties of composites containing 2.5 and 5 wt %
dentin is denoted as mineralized tissue where filler. The results indicate that HA reinforced
the primary function is load-bearing. A closely with CaO-P20 S glass at about 2.5 wt % is of
related tissue to dentin, both positionally and significant benefit to both the stabilization of
functionally is bone, which is a complex min- HA and production of high flexural bend
eralized collagen. The composition and strength composites. This is illustrated in Fig.
orientation of the bone components as load- 44.1. for three types of CaO-P20 S glasses
bearing composites vary in different bones. where that mole ratio CaO/PzOs was 20/80,
Thus, in designing synthetic devices for repair- 30/70 and 50/50 for composites designated as
ing bone, one has to take into account the CzP, C3P and CsP, respectively.
properties of the hard tissue in question. These
 Orthopedic implant applications 959

160 nents for joint arthroplasty. Each of these

--HA
140 ° 2.S%C2P applications has a unique set of material and
120
2.S%C3P
_-2.S%CSP I- mechanical requirements which are addressed
100 £ in this section.
FRP composite materials can be consid-
'iii
a.
!. 80 --/V ered to be composed of at least three, and
8!IL 60
/ // possibly four, distinct components: the fiber
40 ~/ which reinforces the matrix; the polymer
i! matrix which provides three-dimensional
20 support to the fiber; and the fiber/matrix
o interface which serves as the agent of load
1150 1200 1250 1300 1350 1400 transfer between the fiber and polymer
Firing Temperature (e) matrix. In certain formulations, a fourth com-
ponent, the fiber/matrix interphase, must
Fig. 44.1 Effect of glass composition on flexural
bend strength for three different glasses at 2.5% also be considered. The interphase represents
wt % additions (from Knowles and Bonefield, the volume of matrix immediately adjacent to
1993). each fiber which is influenced by the fiber,
causing the matrix in this area to have mor-
phological differences (Le. crystallinity)
44.3 ORTHOPEDIC IMPLANT APPLICATIONS
and/ or compositional differences (Le. mix-
The incentive for use of fiber reinforced poly- ture of matrix and sizing agent from fiber)
mer (FRP) composite materials in most compared to the bulk matrix material. The
nonmedical applications is provided by their site-specific biologic environment can poten-
superior specific mechanical properties (i.e. tially influence each of these components of
strength/weight, stiffness/weight ratios). the composite differently. Fortunately, experi-
However, this offers very little benefit for ence has shown that the in vivo environment
orthopedic implant applications. Implants are can often be suitably represented by simple
typically sufficiently small such that weight is physiologic saline solutions at 37°C (99°F)
not an important design requirement. and pH = 7.4 for in vitro biomaterials perfor-
However, there are other very important med- mance investigations. However, it must be
ically related incentives for the development recognized that exceptions to this are not
of these materials for the treatment of muscu- uncommon, and site-specific in vivo testing is
loskeletal problems. essential prior to clinical evaluation of new
While there are numerous potential ortho- composite material formulations; not only for
pedic applications for FPR composite implant biocompatibility assessment, but also to
materials, most work has been concentrated in ensure materials response to the biologic
four specific areas. FRP composite materials environment has been properly understood.
are being investigated for the design of FRP composites are uniquely different com-
femoral components for total hip arthroplasty pared to metals in that they are permeable to
and for pins, plates, screws, and nails for frac- moisture and salt ions. Therefore, if environ-
ture fixation as alternatives to metal alloys. mental durability is to be properly
Fiber reinforcement is also being investigated investigated, test samples ideally should be
as a means of improving the fatigue resistance fully saturated in their test environment prior
of polymethylmethacrylate (PMMA) bone to testing. The American Society for Testing
cement and the wear, creep, and fatigue resis- and Materials (ASTM) F04 Subcommittee on
tance of ultrahigh molecular weight Composite Materials is currently developing
polyethylene (UHMWPE) articulation compo- standards for environmental conditioning
960 Composite biomaterials

prior to materials performance evaluations to severely complicate revision surgery (Engh

address this issue. and Bobyn, 1988).
The biologic response to an implant mater- Several theoretical and experimental
ial is as important as the material response to investigations have indicated that more com-
the environment. Biologic response to an pliant femoral components result in higher
implant material has been found to be strongly stress levels in the calcar area of the femur,
dependent upon whether the material in ques- thereby reducing stress shielding and main-
tion can be internalized (phagocytosed) by the taining greater levels of bone stock mass and
cells of the body which regulate the body's for- quality (Bobyn et al., 1992; Huiskes, 1992).
eign body response. Thus, a material which is While this is true, a widely reported miscon-
very well tolerated in bulk form when ception concerning this point is that an
implanted in the body (i.e. too large to be optimal condition would be reached if a
ingested by cells) may elicit a strong inflam- femoral component could be designed to
matory response when in particulate form if match the longitudinal elastic modulus or
stiffness of cortical bone. This idea is com-
the material particles are sufficiently small (i.e.
< 10 mm (0.4 in» to be phagocytosed by cells pletely erroneous and represents a failure to
which mediate inflammatory response understand the mechanics of femoral compo-
(Goldring, Clark and Wright, 1993; Black, nent/ femur load transfer. As long as a
1992). Thus, in the development of composite femoral component relies on intramedullary
biomaterials, it is important to not only assess fixation, non-physiologiC load transfer will
the biocompatibility of the device in question, occur, leading to some degree of stress shield-
but to also assess biologic response to wear ing in the proximal femur. However, as a
debris and degradation products of the fiber general concept, the degree of stress shield-
and polymer matrix materials. ing should decrease proportionately to stem
bending compliance.
Three mechanical factors must be consid-
44.3.1 FEMORAL COMPONENTS FOR TOTAL
ered for the successful design of FRP
HIP AR1HROPLASTY
composite femoral components. These are: (1)
Femoral components are permanent implants increased stem compliance to reduce calcar
ideally intended to last the entire life time of stress shielding; (2) fixation to bone and main-
the patient. Current day metallic devices for tenance of acceptably low stem/bone
this application have reported life expectan- interfacial stresses to prevent loosening; and
cies of 90% survival out to 10-15 years in the (3) maintenance of acceptable stress levels
elderly patient population, lower survival within the femoral component to prevent
rates in younger patients (Callaghan, 1990). fatigue failure. These factors are not indepen-
While this is considered good performance for dent; both interfacial stress level and
an orthopaedic implant, certainly 90% sur- component fatigue strength will be influenced
vival at 30 years would be much preferred. by implant compliance. An optimal design
Metallic components have the draw-back of may exist which can satisfy all three mechani-
being so stiff that much of the applied joint cal requirements and provide a superior hip
load is bypassed around the normally highly joint replacement over current day metallic
stressed proximal-medial region of the femur devices. Because of low bending stiffness, FRP
(calcar). This condition has been documented composite stems cannot be properly evaluated
to lead to stress-shielding induced bone by test methods developed for stiff metallic
resorption (Sumner et al., 1992). In severe stems. The ASTM is currently developing a
cases, this may not only contribute to compo- standard practice for fatigue testing of compli-
nent failure through loosening, but may also ant FRP composite stems.
 Orthopedic implant applications 961

Two FRP material systems are primarily Two classes of FRP composite fracture fixa-
being considered for femoral component tion devices have been primarily considered:
development: carbon fiber-reinforced polysul- nonabsorbable and fully absorbable. Compliant
fone (CF /PSF), and carbon fiber-reinforced nonabsorbable plates and nails have been
polyether etherketones (CF /PEEK) (Davidson, investigated as a means of aVOiding stress
1987; Skinner, 1988). Several studies have shielding associated with the use of stiff metal-
addressed the effect of the biologic environ- lic components (Woo et al., 1976).
ment upon the mechanical material properties Unfortunately, clinical results have indicated
of these two composite materials with mixed that, in high load bearing applications such as
results. CF /PEEK has been demonstrated to be the tibia, compliant fracture fixation plates
very durable in physiologic saline environ- allow excessive motion at the fracture site caus-
ments when PEEK was APC2 grade (ICI, ing unacceptable levels of pain upon weight
Tempe, AZ) (Strait et al., 1991; 0'Ariano et al., bearing (Tayton et al., 1982; Tayton and Bradley
1994; Zhang et al., 1994), while 380 grade PEEK 1983). This presents a 'Catch-22' situation in
(ICI Films Inc., Wilmington, DE) has been which compliant bone plates may only be able
found to be sensitive to hydrolytic degradation to be utilized in very low load bearing applica-
of the fiber / matrix interface (Meyer and tions, however, without high load bearing,
Latour, 1993). CF /PSF strength has been stress shielding is not a serious concern and
reported in separate studies to be both very compliant plates are then no longer needed.
stable (Overland et al., 1993) and significantly While the advantages of nonabsorbable com-
degraded by exposure to physiologic saline posite fracture fixation devices are therefore
and exudate environments (Strait et al., 1991; questionable, fully absorbable composite frac-
Latour and Black, 1992, 1993). Biocompatibility ture fixation devices offer significant potential
studies of both CF /PSF and CF /PEEK com- advantages over currently used metallic
posite materials have suggested that, in bulk devices. These materials have the potential to
form, these materials should provide accept- be developed into components with sufficient
able biocompatibility for use in femoral initial stiffness and strength for load bearing
component applications (Wenz et al., 1990). fixation, which are then slowly degraded and
However, the generation of wear debris from absorbed by the body following healing such
implant/bone abrasion, and subsequent that implant retrieval is unnecessary.
potential inflammatory reactions, is an impor- Two types of fully bioabsorbable compos-
tant concern. ites have been investigated: (1) polymer fiber
reinforced polymer, and (2) ceramic reinforced
polymer. Bioabsorbable polymeric materials
44.3.2 FRACTURE FIXATION DEVICES
for both fiber and/or matrix which have been
FRP composite material devices are being widely investigated are polyglycolic acid
developed for the replacement of metallic (PGA), polylactic acid (PLA), polydioxanone
plates, screws, pins, and nails for fracture fixa- (PDS) (T6rmaHi et al., 1991; B6stman et al.,
tion. In contrast to joint replacement, these 1991), polY-E-caprolactone (PCL) (in't Veld,
applications require only temporary implants. 1993), and polyhydroxybutyrate (PHB)
The ideal fracture fixation device would be (Knowles et al., 1992). More recently, bioab-
sufficiently rigid initially to provide adequate sorbable polycarbonates and 'pseudo'-
fracture stabilization, then gradually decrease polyamino acids have also been developed as
in stiffness over time to transfer stress to bone well (Pulapura and Kohn, 1992). Types of
after fracture union to avoid stress shielding, ceramic fibers which have been investigated
and then eventually disappear to eliminate the are phosphate glasses and calcium phosphate
need for retrieval surgery. based ceramics (Andriano, Daniels and Heller,
962 Composite biomaterials

1992). Implant designs have utilized both con- devices typically occurs by loosening via com-
tinuous and discontinuous fiber reinforcement bined fatigue fracture of the cement and the
of the polymer. bone/cement or cement/implant interfaces.
The main obstacle to widespread applica- Fiber reinforcement is a potential means of
tion in these materials is their rapid loss of improving the fatigue strength of bone
strength and stiffness in vivo. This currently cement. Several investigations have
restricts their use to only low load bearing addressed this issue with the use of short
applications. This behavior has primarily been fiber reinforcements of carbon (Pilliar et al.,
attributed to rapid hydrolysis of the 1976), polyaramid (Wright and Trent, 1979),
fiber/matrix interface (Andriano, Daniels and ultrahigh molecular weight polyethylene
Heller, 1992). The development of more (Wagner and Cohn, 1989), titanium
durable fiber/matrix interfacial bonding in (Topoleski et al., 1992), stainless steel
absorbable composites is essential if these (Fishbane and Pond, 1977), and PMMA
materials are to be successfully developed for (Buckley, 1991), to name a few. In general,
load bearing applications. these studies have demonstrated the
The biocompatibility of bioabsorbable poly- expected result that short fiber reinforcement
mers and fibers being developed for can increase both strength and toughness of
orthopaedic applications is considered to be bone cement. Problems with fiber wetting,
good, however, clinical use of self-reinforced fiber distribution, void content, and increased
PGA pins has demonstrated an 8% rate of viscosity are cited as the major problems
aseptic sinus tract development in patients which have prevented clinical implementa-
(Hofmann, 1992). While not compromising tion. Bulk bone cement is actually a
fracture union, this may require surgical inter- self-reinforced particulate filled composite
vention. Experience suggests this may occur materiaL The composite formulation with
when implant degradation product generation greatest potential may therefore be the
exceeds the local tissue clearance capability. replacement of the PMMA micro spheres with
The balance between degradation product similar quantities of PMMA fibers. This may
release and tissue clearance ability raises a enable mechanical properties to be improved
concern for the development of larger while not causing significant increases in
implants of these materials. This problem may cement viscosity during cure.
potentially be overcome with the development
of more slowly degrading implants.
44.3.4 ARTICULATION COMPONENTS
Ultrahigh molecular weight polyethylene
44.3.3 BONE CEMENT
(UHMWPE) is extensively utilized in total
Polymethylmethacrylate (PMMA) is utilized joint replacement prostheses to provide a low
extensively in orthopedic surgery as a friction surface for articulation against a
method of prosthesis fixation in joint replace- matched metallic or ceramic component. Wear,
ment. Bone cement is prepared in the creep, and fatigue resistance are the major
operating room by the surgical team by mix- problems associated with the use of plain
ing methylmethacrylate monomer with UHMWPE in these application, especially for
polymethylmethacrylate microbeads. The knee joint prostheses (Connelly et al., 1984).
mixture is first allowed to partially polymer- Fiber reinforcement offers a mechanism of
ize, and then is placed in a prepared surgical potentially improving these properties.
site within a bone cavity. PMMA is a brittle Reinforcement with carbon fibers was ini-
polymer and can be considered the weak link tially considered for this application, and even
in joint replacement. The failure of cemented utilized clinically for both knee joint and hip
 Composites for soft tissues 963

joint replacement (Wright et al., 1988). This, 44.4 COMPOSITES FOR SOFT TISSUES
however, proved to be a very poor choice of
Composite biomaterials have been mostly
reinforcement. Wear resistance studies with
associated with their use in conjunction with
this material provided mixed reports ranging
hard tissues. However, composites for repair-
from significantly decreased to significantly
ing or replacing soft tissues can potenti~lly
increased wear rates (McKellop et al., 1981),
become quite important due to the growmg
while fatigue resistance was found to be actu-
interest in vascular and skin grafts, as well as
ally decreased by an order of magnitude
bioartifical organs. Simple devices such as
compared to the unreinforced UHMWPE
sutures for soft tissue repair have been
(Connelly et al., 1984). This behavior can be
patented in composite forms. For instance,
explained by the very brittle nature of carbon
composite silk sutures with compliant copoly-
fiber and the very low interfacial bond
ester matrix have been described as having
strength between the fiber and matrix (Meyer
lower tissue reactivity, and higher strength
and Latour, 1991). This combination leads to
retention in the biologic environment as com-
rapid crack initiation and. prop~gati~n ~d pared with wax-coated silk sutures (Shalaby,
third body wear during articulation VIa fiber
Stephenson and Schaap, 1984). Composite,
fragment release. This incident in the history
woven vascular prosthesis with absorbable
of implant design provides a clear example of
[10/90 poly(L-lactide-coglycolide)] and non-
the complexities of composite materials
absorbable (polyethylene terephthalate)
behavior, and demonstrates the fact that fiber
segments, were made and used as model sys-
reinforcement of a polymer does not necessar-
tems for studying the derivation of neointima
ily improve mechanical performance.
in vascular grafts (Greisler et al., 1988).
Although carbon fiber was not a good
Composite artificial blood vessels wer~ pre-
choice for the reinforcement of UHMWPE, as
pared by injecting water-soluble chIt os an
the old saying goes, 'the baby should not be
derivatives (e.g. hydroxypropykhitosan) and
thrown out with the bath water'. Potential
heparin into a microporous polytetrafluo-
improvement of wear, creep, and fatigue resis-
roethylene tube followed by freeze-drying
tance of UHMWPE may still be achieved using
(Yamamura et al., 1992). What may be consid-
other types of reinforcement. In particular~ a
ered as a bioartificial vascular graft is the one
reinforcement is required which has properties
prepared by endothelial .cell .seed~g of a
of high strength, high strain to failure, high
woven synthetic graft usmg flltratlOn tech-
bond strength to UHMWPE, high wear resis-
nique (Idezuki, 1993). Thus, canine venous
tance, and which is as biocompatible as
endothelial cells which were seeded onto a
UHMWPE in both bulk and particulate form.
low porosity vascular prosthesis were
While UHMWPE is the most commonly
allowed to grow around the fibers in early
used polymer for articulation, other composite
periods. They then formed a monolaye~ on
systems have been investigated such as carbon
the internal surface of the tube at later penods
fiber reinforced triacine resin and polymer
in vitro. Properties of polyester fiber blends
fiber reinforced elastomeric composites
were evaluated as totally (Greisler, et al.,
(Harms, 1984; Sutphin et al., 1993). Despite the
1988a) or partially (Yu and Chu, 1993)
failure of CF /UHMWPE articulation compo-
absorbable vascular grafts.
nents, the reinforcement of polymeric
Melt-blended absorbable polymers made of
articulation surfaces is still an active area of
lactide and glycolide having the proper chem-
research for the development of improved
ical composition to provide controlled
prostheses.
miscibility in the liquid state were molded into
components of surgical staples having the
964 Composite biomaterials

desired in vivo strength retention profile Casper, RA, Kelley, B.S., Dunn, RL., Potter, AG.
(Smith et al., 1988; Jamiolkowski et al., 1989). In and Ellis, D.N. 1985. Fiber-reinforced
the solid state the molded articles exhibited a absorbable composites for orthopedic surgery.
Polymer Mater. Sci. Eng. 53: 497-501.
two-phase morphology. The texture of the dis-
Connelly, G.M., Rimnac, C.M., Wright, T.M.,
persed phase may allow one to denote these Hertszberg, RW. and Manson, J.A. 1984.
systems as microcomposites. Fatigue crack propagation behavior of ultrahigh
molecular weight polyethylene. J. Orthop. Res. 2:
119-125.
REFERENCES
Daniels, AU., Chang, M.K.O. and Adriano, K.P.
Andriano, K.P., Daniels, AU. and Heller, J. 1992. 1990. Mechanical properties of biodegradable
Biocompatibility and mechanical properties of a polymers and composites proposed for internal
totally absorbable composite materials for fixation of bone. J. Appl. Biomater. 1: 57-78.
orthopaedic fixation devices. J. Appl. Biomat. 3: D'Ariano, M.D., Latour, RA Jr., Kennedy, J.M.,
197-206. Schutte, H.D. Jr. and Friedman, RJ. 1994 Long
Andriano, K.P., Daniels, AU. and Heller, J. 1992. term shear strength durability of CF /PEEK
Mechanical Properties of Composites Reinforced composite in physiologic saline. Trans. Soc.
with Surface-Modified Absorbable Calcium-Sodium- Biomat., 17, 184.
Methaphosphate Microfibers. Proc. 4th World Davidson, J.A.. 1987. The challenge and opportu-
Biomat. Congr., Berlin, Germany. nity for composites in structural orthopaedic
Andriano, K.P., Daniels, AU. and Heller, J. 1991. applications. J. Composites Techn. Res., 9(4):
Biocompatability and mechanical properties of 151-161.
totally absorbable composite material for Engh, C.A and Bobyn, J.D. 1988. The influence of
implant use. Trans. Soc. Biomater. 14: 10. stem size and extent of porous coatings on
Andriano, K.P., Daniels, AU., Smutz, w.P. and femoral bone resorption after primary cement-
Wyatt, RW.B. 1993. Preliminary biocompatibil- less hip arthroplasty. Clin. Orthop. 231: 7-28.
ity screening of several biodegradable Fishbane, B.M. and Pond Sr., RB. 1977. Stainless
phosphate fiber reinforced polymer. J. Appl. steel fiber reinforcement of poly-
Biomater. 4(1): 1-12. methymethacrylate. Clin. Orthop. ReI. Res. 128:
Black, Jonathan. 1992. Biological Performance of 194-199.
Materials. Fundamentals of Biocompatibility. New Gillett, N., Brown, S.A, Dubleton, J.H. and Pool,
York: Marcel Dekker Inc. RP. 1985. Biomaterials 6: 113.
Bobyn, J.D., Mortimer, KS., Glassman, AH., Engh, Goldring, S.R., Clark, c.R. and Wright, T.M. 1993.
C.A., Miller, J.E. and Brooks, C.K 1992. The problem in total joint arthroplasty: Aseptic
Producing and avoiding stress shielding. loosening. J. Bone Joint Surg. 75A: 799-801.
Laboratory and clinical observations of nonce- Greisler, HP., Dennis, J.W., Endean, KD., Ellinger,
mented total hip arthroplasty. Clin. Orthop. 274: J., Buttle, K.F. and Kim, D.U. 1988 (a).
79-96. Derivation of neointima in vascular grafts:
Bostman, O.M. 1991. Current concepts review. Circulation 78 (3 pt. 2) 1+12.
Absorbable implant for the fixation of fractures. Greisler, HP., Endean, E.D., Klosak, J.J., Ellinger, J.,
J. Bone Joint Surg. 73-A: 148-153. Dennis, J.w., Buttle, K. and Kim, D.U., 1988 (b)
Bostman, 0., Hirvensalo, E., Vainionpaa, H., Polyglactin 91O/polydioxanone bicomponent
Makela, A, Vihtonen, K., Tormala, P. and totally absorbable vascular prosthes. J. Vasco
Rokkanen, P. 1989. Ankle fractures treated using Surg. 7: 697-705.
biodegradable internal fixation. Clin. Orthop. Harms J., Mittelmeier H, and Mausle, E. 1984.
238: 195-203. Results of animal studies on the use of carbon
Buckley, C.A, Lautenschlager E.P. and Gilbert J.L. fiber-reinforced plastic protheses. In The
1991. High strength PMMA fibers for use in a Cementless Fixation of Hip Endoprostheses. (ed. E.
self-reinforced acrylic cement: Fiber tensile Morscher), pp. 249-251. New York: Springer-
properties and composite toughness. Trans. Soc. Verlag.
Biomat. 14: 45. Heimke, G. 1989. (Ed.) Bioceramics Vol. 1, Ishiyaku
Callaghan, J.J. 1990. Total hip arthroplasty. Clinical Euro-America, Japan.
perspective. Clin. Orthop. 276: 33-40. Heimke, G. 1990. (Ed.) Bioceramics, Vol. 2, Cologne,
 References 965

Germany, Deutsche Keramische Geselleschaft. term durability of interfacial bonding in carbon

Hofmann, G.O. 1992. Biodegradable implants in fiber / polytheretherketone and carbon
orthopaedic surgery - A review on the state-of- fiber/polysulfone composites following expo-
the-art. Clinical Materials. 10: 75-80. sure to simulated physiologic saline. Trans. Soc.
Huiskes, R 1992. The relationship between stress Biomater. 16: 15.
shielding and bone resorption around total hip Overland, M.K, Clayden, N.J., Everall, N.J.,
stems and the effects of flexible materials. Clin. Koeneman, J.B. and Magee, F.P. 1994. Effect of
Orthop. 274: 124-134. long-term in-vivo / in-vitro environmental expo-
Idezuki, Y., Shindo, S. Shirakawa, M. and Egami, J. sure on the shear strength of
1993. New Functional Materials, Vol. B. (Tsuruta, polysulfone/ carbon fiber composites. Trans.
T. ed.), pp. 333, Elsevier, Amsterdam. Soc. Biomater, 17, 159.
in't Veld, P.J.A., Dijkstra, P.J., and Feijen, J. 1993. In Park, J.B. and Lakes, F.S. 1992. Biomaterials - An
vitro degradation of polyesteramides containing Introduction. 2nd ed., New York, Plenum Press.
poly-caprolactone blocks. Clinical Materials. 13: Pilliar, R.M., Blackwell, R., Macnab, I. and
143-147. Cameron, H.u. 1976. Carbon fiber-reinforced
Jamiolkowski, D.O., Gaterud, M.T., Newman, Jr., bone cement in orthopaedic surgery. J. Biomed.
H.O. and Shalaby, S.w. 1989. Surgical Fastener Mater. Res. 10: 893-906.
Made from Glycolide-Rich Polymer Blends. US Pulapura, S.and Kohn, J. 1992. Trends in the devel-
Patent (to Ethicon Inc.) 4 889 119 opment of bioresorbable polymers for medical
Kelley, B.s., Dunn, RL., Jackson, T.E., Potter, AG. applications. J. Biomaterials Applications. 6:
and Ellis, D.N. 1988. Assessment of Strength Loss 216-250.
in Biodegradable Composite. Proc. 3rd World Shalaby, S.w., Stephenson, J. and Schaap, L. 1984.
Biomater. Congr., Kyoto, Japan, 471. Composite Sutures of Silk and Hydrophobic
Knowles, J.e. and Bonfield. 1993. Development of a Thermoplastic Elastomers. US Patent (to Ethicon
glass reinforced hydroxyapatite with enhanced Inc.) 4 461 298.
mechanical properties. The effect of glass com- Skinner, H.B. 1988. Composite technology for total
position on mechanical properties and its hip arthroplasty. Clin. Orthop. 235: 224-236.
relationship to phase changes. J. Biomed. Mater. Smith, e., Gaterud, M., Jamiolkowski, D.O.,
Res. 27(12): 1591-1598. Shalaby, S.w. and Newman Jr., H.D. 1988. High
Knowles, J.e., Hastings, G.W., Ohta, H., Niwa, S. Glycolide Blends for Absorbable Staples. US Patent
and Boeree, N. 1992. Development of a degrad- (to Ethicon Inc.) 4 741 337.
able composite for orthopaedic use: in vivo Strait L.H., Jamison RD., and Gavens, A 1991.
biomechanical and histological evaluation of Effect of environment on the flexural and com-
two bioactive degradable composites based on pressive strength of carboni polysulfone and
the polyhydroxybutyrate polymer. Biomaterials. carbon/polyetheretherketone composites.
13: 491-496. Trans. Soc. Biomater., 14: 286.
Latour, RA Jr. and Black, J. 1993. Development of Sumner, O.R., Turner, T.M., Urban, R.M. and
FRP composite structural biomaterials: Fatigue Galante, J.O. 1992. Experimental studies of bone
strength of the fiber/matrix interfacial bond in remodeling in total hip arthroplasty. Clin.
simulated in vivo environments. J. Biomed. Orthop. 276: 83-90.
Mater. Res. 27: 1281-1291. Sutphin, e.M., LaBerge, M., Drews, M.J., Lickfield,
Lin, T.e. 1986. Totally absorbable fiber-reinforced G.e. and Black, J. 1993. Design of an elastomeric
composite for internal fixation devices. Trans. composite for orthopaedic applications. Trans.
Soc. Biomater. 9: 166. Soc. Biomater. 16: 147.
McKellop, H., Clarke, I., Markolf, K and Amstutz Tayton, K and Bradley, J. 1983. How stiff should
H. 1981. Friction and wear properties of poly- semi-rigid fixation of the human tibia be? A clue
mer, metal, and ceramic prosthetic joint to the answer. J. Bone Joint Surg. 64-B: 105-111;
materials evaluated on a multichannel screen- 65-B: 312-315.
ing device. J. Biomed. Mater. Res. 15: 619-653. Tayton, K, Johnson-Nurse, e., McKibbin, B.,
Meyer, M.R. and Latour, RA Jr. 1991. Fiber rein- Bradley, J. and Hastings, G., 1984. The use of
forcement of ultrahigh molecular weight semi-rigid carbon-fiber-reinforced plastic plates
polyethylene. Trans. Soc. Biomater. 14: 285. for fixation of human fractures. Results of pre-
Meyer, M.R. and Latour, R.A Jr. 1993. The long- liminary trials. J. Bone Joint Surg. 65-B: 312-315.
966 Composite biomaterials

Topoleski, L.D.T., Ducheyne, P. and Cuckler, J.M. Woo, S. L.-Y, Akeson, W.H., Coutts, RD.,
1992. The fracture toughness of titanium-fiber- Rutherford, L., Doty, D., Jemmott, G.P' and
reinforced bone cement. J. Biomed. Mater. Res. 26: Amiel, D. 1976. A comparison of cortical bone
1599-1617. atrophy secondary to fixation with plates with
Tonio, AJ. and Folmer, RCH. 1987. The clinical use large differences in bending stiffness. J. Bone
of plastic plates for osteosynthesis in human Joint Surg. 5~A: 190-195.
fractures. Clin. Mater. 2: 27~279. Wright, T.M. and Trent, P.S. 1979. Mechanical prop-
Tormiila, P., Vasenius, J., Laibo, J., Pohjonen, T. and erties of aramid fibre-reinforced acrylic bone
Rokkanen, P. 1991. Ultra-high-strength cement, J. Mater Sci. Letts. 14: 503-505.
absorbable self-reinforced polyglycolide (SR- Wright, T.M., Rimnac, CM., Faris, P.M. and Bansal,
PGA) composite rods for internal fixation of M .. 1988. Analysis of surface damage in
bone fractures: In vitro and in vivo study. J. retrieved carbon fiber-reinforced and plain
Biomed. Mater. Res. 25: 1-22. polyethylene tibial components from posterior
Vainionpiiii, S., Kilpikar, J., Laibo, J. Helevitro, P., stabilized total knee replacements. J. Bone Joint
Rokkanen, P. and Tormiilii, P. 1987. Strength and Surg. 70-A: 1312-1319.
strength retention in-vivo of absorbable, self- Yamamura, K., Sakurai, T., Kizawa, H. and Harada,
reinforced polyglycolide (PGA). Biomaterials. 8: H. 1992. Jap. Pat. Appl. 92-58845; Chem. Abstr.
46-47. 119(24): 256585Z.
Wagner, H.D. and Cohn, D. 1989. Use of high-per- Yu, T.J. and Chu, CC 1993. Bicomponent vascular
formance polyethylene fibres as a reinforcing grafts consisting of synthetic absorbable fibers.
phase in poly(methylmethacrylate) bone J. Biomed. Mater. Res. 27: 1329-1339.
cement. Biomaterials. 10: 139-141. Zhang, G., Latour, RA Jr., Kennedy, J.M., Schutte,
Wenz, L.M., Merritt, K., Brown, S.A, Moet, A and H.D. Jr. and Friedman, RJ. 1994. Long term
Steffee, AD. 1990. In vitro biocompatibility of compressive strength durability of carbon fiber
polyetheretherketone and polysulfone compos- reinforced PEEK composite in phYsiologic
ites. J. Biomed. Mater. Res. 24: 207-215. saline. Trans. Soc. Biomaterials, 17: 160.
SCIENTIFIC APPLICATIONS OF COMPOSITES 45
Vicki P. McConnell

45.1 INTRODUCTION research possible. And while national technol-

Science may be seen as a journey, encompass- ogy agendas are largely determined by the
ing both inner and outer space on the quest for presiding political administration, there is
greater understanding of the universe. This nothing temporary or partisan about the expo-
can take literal form, such as the journey nential growth of global competition.
toward ever farther, faster flight regimes of Government agencies in the USA have
experimental aircraft like the X-3D, or here on worked over the years to form a long-term
Earth aboard a high-speed train that levitates national technology strategy that integrates
on magnetic rails. Or the journey can take the- materials science. These have included the
oretical form in scientific instruments that National Critical Materials Council (estab-
measure the smallest particles of matter on lished by Congress in 1984), the White House
Earth and peer into the black holesc-of space. Office of Science and Technology (which coor-
From the human eye to 'eye in the sky' space dinated President Bush's National Materials
telescopes, science continually extends our Initiative and the $1.8 billion Advanced
vision of the universe. Materials and Processing Program), the
Science and composites have always inter- National Center for Advanced Technologies
sected in the R&D laboratory with new (NCAT) formed in 1989 by the Aerospace
materials and process discoveries, and in the Industries Association, the Department of
fabrication of actual structures used in scien- Commerce's Advanced Technology Program
tific applications. The material selected for the administered by the National Institute of
Hubble Space Telescope metering truss, for Standards and Technology (NIST), and the
example, and the process used to build the Defense Advanced Research Projects Agency
components of that structure are directly (DARPA) within the Department of Defense.
related to its ability to accomplish a unique sci- Grants totalling nearly $500 million were
entific mission. awarded under President Clinton's
As one might expect, pure science is rarely Technology Reinvestment Project (TRP) in this
accomplished in a vacuum; helpful data are program's first year, 1993. TRP channels
drawn from commercial, industrial and mili- defense department funds to projects that
tary programs and may return to those venues emphasize dual-use technology for both com-
if the scientific application involves technol- mercial and military applications, with the
ogy transfer. Government and private sector ultimate goal of enhancing cost and informa-
sources may team up to make the scientific tion sharing and ensuring that key USA
industries stay on the technological leading
edge (Lynch, 1993). The variety of composite-
Handbook of Composites. Edited by S.T. Peters. Published related TRP projects have ranged from rapid
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 densification of carbon-carbon components,
968 Scientific applications of composites

infusion-molded infrastructure products, were designed for compression load capabili-

marine craft and dockside structure develop- ties from 860-6000 kg (1900-13 230 lb). The
ment to electric vehicle body and propulsion same thermoset prepreg used for Magellan
systems. parts worked well in the science and
radioisotrope thermal-generator booms and
low-gain antenna struts and mast on the
45.2 APPLICATIONS
Galileo Jupiter observer.
For the scientific applications described in Originally envisioned as one of four 'great
this chapter, carbon fiber-reinforced ther- observatories' in the sky, NASA's Hubble
mosets and thermoplastics offer a superior Space Telescope (HST) took its place in 1990 as
strength-to-weight ratio over metal and other the largest optical system in orbit - the size of
materials, as well as extremely close fabrica- a railroad boxcar weighing 11 567 kg (25 500
tion tolerances, complex contours and load lb) (McConnell, 1989). Capable of studying
tailorability, part consolidation through inte- wave lengths from far UV to far IR, the focal
gral design, and resistance to multiple length of this 'eye in the sky' is 607 mm (189
degradation factors. In some cases, parts and ft). Pointing accuracy and stability are critical
structure simply could not be made to meet - the telescope may hold direction for as long
performance parameters without the use of as 24 h at a time depending on the object of
advanced composite materials. focus (Fig. 45.1).
Supporting the secondary mirror in precise
alignment with the primary mirror is a carbon
45.2.1 SCIENCE-MISSION SPACECRAFT
fiber / epoxy metering truss, engineered to
For composites, the hard vacuum of space pre- maintain stability of the nominal 5080 mm
sents a rigorous environment which may (200 in) spacing between the mirrors within
include atomic radiation, thermal cycling, UV ±15 x 10-4 mm (±60 microinches) in the pres-
ionizing radiation and micrometeroid impacts. ence of ±11 °C (±20°F) excursions at mean
In answering the challenges of these synergis- temperature of -73°C (-100°F). The truss was
tic effects on spacecraft, advanced composites fabricated by Boeing Defense & Space Group
provide crucial payload weight savings, near- (Huntsville, AL), along with a carbon
zero CTE, and good dimensional stability fiber / epoxy focal-plane structure that sup-
when material formulation or design takes ports five primary science instruments and
hygroscopy and outgassing into effect. three fine-guidance sensors.
Composites are commonly used in optical Additional composite HST parts include a
benches, precision mounts (including struc- support-systems module equipment shelf that
tural trusses), solar arrays and booms (Lubin holds three positioning gyros, and the fine-
and Dastin, 1982). guidance keel. Discrete internal ribs direct load
paths out the main core members and prevent
curvature in these parts. Shelf and keel were
Spacecraft truss structure
designed by Perkin-Elmer (Danbury, CT) and
Carbon fiber / epoxy tubes joined into truss fabricated at Composite Optics Inc. (San Diego,
structures are commonly used for supporting CA). COl has developed a proprietary mois-
instruments on science spacecraft, such as the ture-barrier plating technology to seal epoxies
propulsion subsystems and upper-stage from the effects of moisture prior to launch.
assembly for the radar-mapping Venus probe Both COl and Jet Propulsion Laboratory
Magellan. Some 60 tubes in the truss range in (Pasadena, CA) have built spacecraft test arti-
diameter from 25-100 mm (1-4 in) and in cles using a carbon fiber/cyanate ester
length from 180-1830 mm (7 in-6 ft). They (McConnell, 1992b) composite that shows less
 Applications 969

Fig. 45.1 Composite structural truss (a) aboard Hubble Space Telescope (b) aligns primary and secondary
mirrors. (Sources: NASA, Perkin Elmer.)

moisture absorption than epoxies and greater Coronograph Spectrometer (UVCS), designed
toughness than either epoxies or BMls. Lastly, as an occulted telescope to focus on solar
HST's high-resolution spectrograph optical winds (McConnell, 1993c). As with HST, opti-
bench was manufactured by Hercules cal mechanisms require accurate and
Aerospace Co. (Magna, UT) to specifications repeatable positioning, so UVCS's design cen-
from Ball Aerospace Systems Group (Boulder, tralized upon a three-segment carbon
CO). Carbon fiber/epoxy offers CTE equal to fiber / epoxy structure. Ball Aerospace Systems
quartz and maintains 2 m (6.5 ft) of optical sur- Group and COl built the seven-sided truss of
faces to within 0.0254 mm (0.001 in). flat panels that result in cylindrical shape.
Launched in December 1995, the Solar and Some 80 separate stability requirements had to
Heliospheric Observatory (SOHO) was be met, along with strict mass requirements
designed to be pointed directly at the Sun to (the structure weighs 21.7 kg (48Ib) but sup-
study the interactions between the Sun and ports 91.6 kg (202Ib) of instrumentation), and
Earth's environment. The largest moving the ability to survive launch loads of 18 G.
instrument aboard SOHO is the Ultraviolet High-modulus pitch-based carbon fiber/epoxy
970 Scientific applications of composites

prepreg met weight, mass and optical stabil- to the central cylinder, large curved and coni-
ity parameters with no outgassing or cal shell structure, mirror support sleeves, and
microcracking. Especially thin prepreg solar array panels. Eastman Kodak built the
(0.063 mm (0.0025 in) cured ply thickness) 26-ft long optical bench that forms the main
also enhanced the thermal conductivity of the structural element, with carbon fiber skins
fiber, an added stability factor (Kilpatrick, over honeycomb core.
1992).
LDEF composite specimen testbed
AXAF orbiting observatory
Proof positive of the hostile effects of space
Slated for launch from the Space Shuttle in startled cosmonauts on the Mir space station
August 1998, the Advanced X-Ray in 1992 when they found not a shred
Astrophysics Facility spacecraft will bridge remained of the Soviet flag after a year in
the gap between the two Great Observatories orbit on the station's exterior. Further proof
currently in orbit: Hubble Space Telescope and comes in data retrieved from NASA's Long
Compton Gamma Ray Observatory. Duration Exposure Facility (or LDEF), a 12-
TRW (Redondo Beach, CA) is prime con- sided circular spacecraft launched into low
tractor for AXAF and the spacecraft will Earth orbit (LEO) from the Challenger shuttle
contain four sets of unique cylindrical, grazing in April 1984 and retrieved in 1990 (Fig. 45.2).
incidence mirrors. These are mounted concen- By 1993, postflight analysis of the 86 experi-
trically in a nested array. Carbon fiber/epoxy ment trays containing more than 10 000
is used in equipment compartments attached different material specimens has been carried

Fig. 45.2 End panel of LDEF

spacecraft shows effects of space
exposure on materials speci-
mens. (Source: NASA.)
 Applications 971

out by a team of international scientists. (Fig. 45.3). Atomic oxygen is identified as the
Research on LDEF's composites specimens most detrimental factor to polymer-matrix
has provided a benchmark for next-genera- composites, especially on uncoated surfaces
tion space materials, such as those that make (George, 1992). Erosion has been observed to
their way onto the international space station 0.127 mm (0.005 in) (about one ply of laminate)
(Stein, 1993). on leading edge components made from car-
The six-year duration of LDEF's exposure bon fiber with epoxy, polyimide and
to the LEO space environment provided the polysulfone matrices.
longest term flight data ever retrieved for Micrometeroid damage was not cata-
analysis at that time. Post-retrieval sympo- strophic on any LDEF specimens, but can
siums have shared analytical conclusions, compromise surfaces and lead to substrate
and NASA selected M/Vision software from erosion (Blair, 1992) (Fig. 45.4). Another con-
PDA Engineering (Costa Mesa, CA) as one clusion drawn based on post-flight analysis
format for the sizeable LDEF Materials was that predicting erosion rates and formu-
Database which contained research informa- lating material with enhanced AO resistance
tion on about 35% of the LDEF materials by should key on fiber resistance as the dominant
the end of 1994. factor. A prediction model for AO erosion
An important conclusion drawn from LDEF yield in polymer materials has been developed
specimen analysis is that environmental (Tennyson, 1993), based on the repeatable
effects are dependent upon the location of atomic composition of the polymer.
composite hardware on a spacecraft - leading An LDEF follow-on spacecraft, the
or trailing edge, shielded or unshielded from European Retrievable Carrier (EURECA) was
degradation factors (only metal matrix com- retrieved by the shuttle Endeavor in June, 1993
posites with aluminum matrices showed no after 11 months in a 370 km (230 mile) orbit.
location-dependent degradation; magnesium- EURECA's mission focused on materials and
matrix composites oxidized on sample edges) fluids in low microgravity and LEO exposure

LDEF Orbital
• Gravity Gradient
SlibUIzod AWtude
Flight Orientation

Fig. 45.3 LDEF orientation in orbit. (Reprinted with permission of Bland A. Stein, NASA Langley Research
Center.)
972 Scientific applications of composites

45.2.2 THE X-30 NATIONAL AEROSPACE

PLANE (NASP)
This NASAl Air Force single-stage-to-orbit
technology demonstrator program begun in
1986 generated significant advanced research
into aerothermodynamics, propulsion, and
new enabling materials (Fig. 45.5). These
accomplishments were due in great part to
unprecedented teaming among competitive
aerospace companies contracted for the pro-
ject. NASP was designed as a piloted vehicle
capable of horizontal takeoff and landing on
conventional runways (with an airbreathing,
hydrogen slush-fueled engine), hypersonic
cruise (speed range to Mach 25), low Earth
orbit, and operation at temperature extremes

Fig. 45.4 Two LDEF composite samples show

micrometeroid damage in center. AO erosion
affected only epoxy in E-glass/epoxy sample (a)
but uniformly degraded matrix and fiber in carbon
fiber/epoxy sample (b). (Source: Christopher
Bland, Lockheed Missiles & Space Co.)

on solar array components. Shorter duration

experiments (from 11 to 40 h of exposure) AlB Met - 160
flown aboard the shuttle Atlantis in July 1992 Drive Actuator Arm
focused on further analysis of atomic oxygen
effects (particularly flux reaction and erosion
rates) upon nearly 1500 material and coating
samples. These experiments included NASA's Fig. 45.5 Materials science results in technology
EOIM-3, and payloads sponsored by the transfer from NASP demonstrator vehicle research
University of Alabama and Case-Western (a) to a computer part that speeds up data access
Reserve University (Friebele, 1992). (b). (Source: NASA/Air Force.)
Table 45.1 Innovative materials for hypersonic vehicle studied under NASP program

Material type Component Temp Key benefits Further properties

applications regime, development required
SF
Titanium-based Skins, internal 1100 Low density; oxidation and high- Improve alloy ductility, toughness, and
materials: structure, to temperature resistance; economical processing of aluminides; develop
• Rapidly solidified engine parts 1600 rapid-solidification hydrogen-resistant barrier coating
TlAI alloys (Alpha 2, processes now available
Beta 21S, Superalpha 2)
• TW MMCs (with SiC Structural panels, 1400 Stiffness and strength, thin-gauge form- Solve problems in fiber/matrix interface
tidiboride, ticarbide, rocketnozzle, to ability; plasma-spray techniques now reactivity and CTE mismatch potential;
tisilicate, or honeycomb 1800 available; room-temperature ductility fiber coating required for thermal-shock
alumina fibers) face sheet resistance in processing
• XD composites 'Ifot' structural 1400 Strength; high-temperature resistance Compile further database and testing on
(TiAl with tidiboride components to composite engineering of particle
or ticarbide particles) 1800 dispersion and shaping
Carbon-carbon Thermally protected 2500 Most mature enabling materials Oxidation-protection coating required
composites structure - nose cap, to system; light weight; large, complex, with thermal-cycling provability
wing leading edges 3500 curved shapes can be fabricated
Ceramic-matrix Ifigh-temperature, 1500 Inherent oxidation resistance; easy Need fiber maturity for higher;
composites (glass- lightweight surfaces to formability to complex shapes; high temperatures improve resistance to
ceramic, SiC/SiC) adjacent to nose cap, 2500 acoustic resistance; low thermal matrix-crack propagation during operation
wing leading edge conductivity; hydrogen resistant
(honey-comb-core
panels truss-core panels)
Beryllium-based Ifeat-exchanger 1000 Light weight; high elastic modulus; Improve fracture toughness at cryogenic
alloys (BeAl) panels, engine thermal conductivity comparable to temperatures and brittleness encountered
components aluminum; nonreactive to hydrogen; in high-temp engine applications;
considerable fabrication data base consider toxicity potential in handling
~
~
Copper-based Actively cooled 1200 lligh thermal conductivity; improved Continue processing and fiber-coating ~
composites (copper- structure/heat stiffness; inert to hydrogen technology, including discontinuous <=;.
;::.
carbon fiber) exchangers copper composites alloyed with niobium, ....
<:)
which can strengthen material with no ;:i

sacrifice in matrix conductivity '"

\C
(Reprinted from Advanced Composites magazine, Nov lDec, 1991, by permission of Advanstar Communications, Eugene, OR, USA.) 2j
974 Scientific applications of composites

(cryogenic to 2760°C (5000°F». R&D efforts necting rods and other engine parts. Thin-
resulted in innovative materials science gauge titanium aluminide foil has been tested
among five composite categories (McConnell, in heart valve assemblies and pacemaker cases.
1990), with more than 2000 airframe and Adaptation of computational fluid dynamics
scramjet propulsion components, structure, (CFD) - the study of high-speed laminar air
and systems built (Table 45.1). flow developed to maximize NASP's body con-
Until October 1994, NASP represented the figuration - has proven useful in analyzing
hypersonic flagship program of the USA. At aircraft and automotive aerodynamics, as well
that time, the initial program goals transitioned as blood flow in artificial hearts. Dr. Steve
into the hypersonic system technology pro- Charles of the Center for Retina Vitreous
gram (HYSTP), funded jointly by the US Air Surgery (Memphis, TN) began work with the
Force and NASA and which will focus on Rockwell Science Center (Thousand Oaks, CA)
development of hypersonic scramjet propul- in February 1992 to apply NASP-developed
sion. Yet resulting scientific R&D from the CFD to modeling fluid flow inside the eye. CFD
NASA program generated considerable tech- will help analyze the role played by rapid eye
nology transfer to a broad range of commercial motion (up to 600°/ s) in retinal detachment,
applications, including oilfield pipe, auto and possibly enhance the precision of surgical
engines and hip joint implants. One of the tita- procedures. Dr. Charles performs some 15 000
nium MMCs developed did not exist before the vitreous retinal surgeries annually and ulti-
NASP program and is 100 times more resistant mately expects to use CFD to design a new
to corrosion than standard aircraft titanium. suite of surgical tools, 'applying hard science
An aluminum/beryllium alloy tested in NASP rather than guesswork to the entire tool set'.
heat exchanger prototypes has since found
application in computer actuator arms, provid-
45.2.3 DELTA CLIPPER EXPERIMENTAL
ing a 20% increase in data access speed. Big
LAUNCH VEHICLE (DC-X SSTO)
Three automakers have studied NASP-devel-
oped titanium MMCs and alloys to cut weight Under a $60 million contract from the Ballistic
by 50% in valves, piston rings, cam shafts, con- Missile Defense Organization, McDonnell

Fig. 45.6 Composite aeroshell

and nose cone enabled DC-X
single-stage-to-orbit prototype
launch vehicle to fly nose-first
takeoff and tail-sitter landing.
(Source: McDonnell Douglas
Aerospace.)
 Applications 975

Douglas Aerospace (Huntington Beach, CA) -73°C (-100°F), the aeroshell incorporated
developed its Delta Clipper experimental pro- built-in longitudinal longerons to handle
totype (DC-X) to prove the practicality, engine/truss, parachute deployment and
reliability, operatability and cost efficiency of a landing gear loads.
reusable single-stage-to-orbit (SSTO) flight Glass fiber I epoxy was used in the 762 mm
vehicle. Comparative figures suggest that (2.5 ft) tall nose cone since it required signal
reusable rockets like a full-scale DC-X could transparency for antennae. McDonnell
reduce commercial/military payload delivery Douglas decided that carbon fiber I epoxy
costs to under $500 lIb - an attractive alterna- skins over syntactic foam in the DC-X base
tive to the $10 OOO/lb for the Space Shuttle heat shield provided the best material system
(1993 figures). for meeting weight, price, and scheduling
The one-third scale DC-X prototype had demands. The shield (about 396 mm (13 ft) on
completed five suborbital flight tests before a side) featured a unique 'super circle' sec-
being grounded by an on-board explosion that tional design, joined with titanium fasteners.
damaged the composite aeroshell. Despite this Both the prime and subcontractor worked to
accident, DC-X landed safely and the shell the philosophy of building real structures to
was repairable. Data collected in the first flight collect and prove data before moving on to
test series formed the advanced technology more expensive program goals, something of a
backbone for an upgraded DC-XA vehicle, first in such R&D efforts.
developed jointly by McDonnell Douglas The durability of composite aeroshell mate-
(now the Boeing Co.) and NASA (with some rials and overall system design was proven in
$43 million in NASA funding) in 1995. The the June 1994 explosion; according to the pro-
DCX-A was renamed the Clipper Graham in gram director, the 1220 by 4570 rom (4 by 15 ft)
1996, and completed several milestone flights vertical tear would probably have demolished
that year: altitude of 10 300 ft with 550 ft lat- other launch vehicles. McDonnell Douglas
eral movement, and a second flight 26 hours developed and fabricated a robust carbon
after the first to demonstrate rapid turn- fiber/epoxy liquid hydrogen tank for DCX-A
around. in 1996, 8 ft in diameter and 16 ft tall, but 33%
Rapid prototyping played an essential role lighter than previous aluminium alloy tanks.
in the cost effectiveness of building the DC-X
prototype, including construction of compos-
45.2.4 GROUND-BASED SCIENTIFIC
ite aeroshell, nose cone, and base heat shield
INSTRUMENI'S
components (McConnell, 1993c). The aeroshell
and nose cones were fabricated by Scaled
Wind tunnel blade
Composites (Mojave, CA); five aerodynamic
composite flaps on the carbon fiber/epoxy While CFD places air-flow analysis within the
aeroshell provided the ability to ascend nose domain of computer mathematics, there is
first and return to Earth for a 'tail sitter' land- nothing quite like proving an aerodynamic
ing. Composites accommodated the program's concept in a wind tunnel. At NASA's Ames
ambitious scheduling with rapid aeroshell Research Center (Moffett Field, CA), different-
construction (nine months) and by allowing sized wind tunnels can accommodate small
continuing modifications without huge cost models of experimental structures and shapes
overruns. Lay-up of woven fabric composite as well as full-scale vehicles (such as tractor
skins over foam core resulted in aeroshell wall trailers and an F/ A-18 aircraft). Compressors
thickness range of 12.7-38.1 mm (0.5-1.5 in). generate high-speed air flow with rotating
Besides enduring external temperatures to blades, usually constructed of aluminum or
149°C (300°F) and internal temperatures to laminated wood. In the Ames tunnel with a
976 Scientific applications of composites

3352 mm (11 ft) test section, composite design sandwich structure of syntactic foam core
and proto typing work supported by Dr. Clem covered with hybrid glass and carbon
Hiel, a former NASA senior design engineer, fiber/epoxy skins laid up at 0/+30/-30 ori-
and his associates has examined the potential entation. Though syntactic foam can typically
of using composite blades in the compressor, be four to eight times higher in weight than
which is capable of speeds to 711 rev/min. traditional foams, mechanical properties are
These blades would offer the advantages of several orders of magnitude higher. This is
improved damage tolerance, longer fatigue due to the energy-absorbing capacity of glass
life, higher damping and improved safety in microspheres inside the foam core, which
the event of catastrophic failure. also localize damage when they are crushed.
Design of a prototype composite blade in In addition, local regions of skin failure are
1995 (Fig. 45.7) (which is 1067 mm (42 in) clearly visible to the unaided eye by an exter-
long, 406 mm (16 in) wide at the root, taper- nal imprint. Carbon fiber skins served as the
ing to 241 mm (9.5 in) at the tip) featured a structural backbone by providing high spe-
cific strength and stiffness, while the foam
core supported skins against impact. Glass
fiber skins acted as sacrificial protective coat-
ing and visual enhancement of impact
damage during residual strength assessment.
Dr. Hiel reports that the prototype composite
blade tested to seven times the expected com-
bination of centrifugal and air loads, and was
clearly capable of carrying higher loads when
the test fixture failed. Manufacturing the
blade through resin transfer molding (RTM)
could result in a blade that would compete
with aluminum in acquisition costs while
reducing weight significantly (Hiel, 1993).

Particle accelerators
To study the exotic mixture of protons, neu-
trons, nuclei and the smallest particles of
matter - quarks - inside the atom, physicists
have turned to complex scientific instruments
that generate enough voltage and acceleration
to create high-energy particle collision.
Analysis of the collisions in this' quark soup'
could reveal mysteries among the building
blocks of matter.
At Newport News, VA, the Southeastern
University Research Association oversees
operation of the Continuous Electron Beam
Fig. 45.7 Wind tunnel blade made of carbon
Accelerator Facility (CEBAF). Here a continu-
fiber / epoxy skins over syntactic foam core (left) ous electron beam with up to 4 billion eV of
weighs 50% less than aluminum counterpart. energy is steered by superconducting magnets
(Source: NASA Ames Research Center.) through up to four orbits of an underground
 Applications 977

Fig. 45.8 Carbon fiber / epoxy skins over honeycomb core in particle detector shell provide structural rigid-
ity and low interference for sensing wires. (Source: Tolo Inc.)

racetrack. Upon particle collision, detectors cal. To meet these requirements and provide
distinguish electrons from target particles so additional rigidity, builder Tolo Inc. (Santa
experimenters can ultimately 'fingerprint' Ana, CA) selected a carbon fiber with 43 mil-
individual nuclear particles (Peterson, 1992). lion tensile modulus and a 121°C (250°F) cure
In the outer structure of CEBAF's Region III epoxy resin. Shear strength requirements
drift chamber, six shells utilize composite called for a 186 kPa (27 psi) minimum and
skins (two-ply carbon fiber/epoxy unidirec- compressive load of 2300 kg/m (130 lb/in)
tional fabric over Nomex® honeycomb core) when loaded parallel to the skin surface.
(Fig. 45.8). Dimensionally, each shell is about Syntactic foam filled in honeycomb edge cells,
9144 mm (30 ft) long, 3353 mm (11 ft) wide, and full bonding occurred within a curved
and 38 mm (1.5 in) thick. Thousands of electri- bond fixture to achieve the shell's 3962 mm
cally-charged sensing wires are mounted on (156 in) radius.
polyurethane-foam/ stainless-steel end plates The Superconducting SuperCollider - the
positioned between the shell segments, so largest atom smasher ever designed - was to
low-density low-interference material is criti- have been housed in an 85 km (53 mi) under-
978 Scientific applications of composites

ground tunnel in Ellis County, TX. The SSC posite prototype support post tubes were fab-
design was based on a dual ring of 10 000 elec- ricated to test various designs to balance
tromagnets accelerating needle-thin proton multiple performance parameters with cost
beams at 20 trillion eV cryogenic temperature constraints (Sondericker 1991; Nicol 1992,
and ultrahigh vacuum. After a decade of work 1993; Hiller, 1991, 1992, 1993). Material and
and investment of $2 billion, however, process systems examined included continu-
Congress cancelled the program. While much ous carbon fiber I epoxy laminates,
was made in the national press about the injection-molded chopped glass fiber IPEI,
wasted effort (and the huge hole in the and RTM'd glass fiber Ibismaleimide.
ground), SSC composites research has yielded Machine-woven near-net shape preforms were
unprecedented results that were directly also considered, along with RTM'd phenolic
applicable to other ongoing high-energy triazine (McConnell, 1991a) (Fig. 45.9). In other
physics applications (such as the Relativistic components, such as coil spacers and end
Heavy Ion Collider at Brookhaven National parts, the predominant prototyping method
Laboratory in Upton, NY, the Tevatron was machining glass fiber I epoxy, though
Collider at Fermi National Accelerator options considered included highly
Laboratory in Batavia, IL, the Tokomak crosslinked BMI doughs and other polyimide
Physics Experiment in Princeton, NJ, and the resins with chopped glass, fabricated by low-
European International Thermonuclear pressure techniques such as structural RTM.
Experimental Reactor). 'Magnets were built for SSC that were never
For SSC dipole magnets (each nearly 18 288 seen before in the world', recounts John
mm (60 ft) long and weighing 11 340 kg (12.5 Morena, who served as the primary compos-
tons)) and smaller quadrupole magnets, COill- ites materials and processes advisor on the

Fig. 45.9 Glass fiber/epoxy end saddles were fabricated and tested for Superconducting SuperCollider
magnets. (Source: Brookhaven National Laboratory.)
 Applications 979

project. 'There were many "lessons learned" as Under the National Maglev Initiative in
a result of the years of composite materials 1992, four maglev teams developed concept
development and applications,' he adds. 'This definitions with funding from the Intermodal
initial work has provided the composites com- Surface Transportation Efficiency Act (ISTEA).
munity with the ability to predict the All concepts were geared toward transport
performance of advanced composites and vehicles capable of 482 km/h (300 mph)
other polymer materials in cryogenic and high speeds and incorporated some utilization of
energy physics applications. These original composites. Among team members in the
materials are now used in MRI, superconduct- early '90s were aerospace companies Beech
ing magnet energy storage, magnetic Aircraft and Grumman Corp. Drawing on its
levitation systems, fusion, power transmission extensive aerospace design/fabrication exper-
and space applications.' Others involved in tise, Grumman Corp. - Aerostructures Div.
magnet prototyping were General Dynamics, (Bethpage, NY), began development of a 2/3-
Westinghouse Electric, Babcock & Wilcox, as scale electromagnetic system, based on
well as the Fermi and Brookhaven laboratories superconducting iron core magnets spaced
and the SSC on-site laboratory. Morena notes a 50 mm (2 in) from guideway rails. According
particular materials characterization study to Richard Gran, Director of Advanced
conducted as part of SuperCollider R&D: sub- Concepts at Grumman, 'the advantage of this
jecting thermoset and thermoplastic resin design as compared to a repulsive magnet sys-
systems to high proton irradiation at cryogenic tem is that no secondary suspension system is
temperatures (10 9 rads at 4.2K). One intriguing required; passenger compartment acceleration
result: undercuring some thermoset compos- is monitored (Shaw, 1993). Though Grumman
ites can increase their longevity in the cold intended to finish construction and testing of
irradiation environment without sacrificing its first magnet by April, 1994, funding issues
performance (Morena, 1994). precluded completion: The company merged
with Northrop Corp. (Los Angeles) and then
with Lockhead Martin by 1997, no maglev
45.2.5 MAGLEV: TRANSPORT SYSTEMS OF THE
development existed within the newly framed
FUTURE
business entities.
One potential application of SSC research on Also intent upon prototyping a maglev sys-
superconducting magnets could be magnetic tem is Maglev 2000 of Florida Corp. (Stuart,
levitation (or maglev) of high-speed trains EL.) formed by SSC materials/processes con-
over elevated guideways. Maglev is consid- sultant and co-inventor John Morena with
ered by many to be the transportation/ freight physicist Dr. John Danby and nuclear engi-
solution of the future. Composites offer a neer / rocket scientist Dr. James Powell. Danby
whole list of attributes in vehicle/guideway and Powell are considered the pioneers of USA
structure and in superconducting coils and maglev technology. Prior to publishing their
electromagnetic shielding: high strength and first papers on magnetic levitation (Powell,
stiffness with low weight, nonconductivity, 1996), Danby reports that concepts of electro-
nonpermeability, corrosion resistance, low dynamic levitation and propulsion had
thermal conductivity and high heat capability surfaced periodically but were not considered
at low temperature. Overall weight savings of practical. Early in his career, he worked on
carbon fiber/epoxy or phenolic vehicle floor building particle accelerators while Powell was
beams and shell structure (sandwich panels of involved in advanced reactors. 'Fixed field
honeycomb core with carbon fiber/epoxy magnets on a moving vehicle can induce cur-
skins) could better aluminum by 15% of body rents underneath or around the vehicle
weight (Cope, 1993). sufficient to levitate it,' explains Danby. 'This
980 Scientific applications of composites

requires a strong force to keep the vehicle sep- REFERENCES

arated from the guideway in a stable way. We Blair, Christopher and Petrie, Brian C. 1992. Low
formulated the geometries to accomplish that earth orbit environmental effects on composite
strong stability with efficient inductive, materials: results from LDEF. Intern. SAMPE
shorted metal-loop elements on the guideway Tech. Con! Proc., October 20-22, Toronto, (24):
and superconducting magnets on the vehicles.' T186-T200.
The highly successful Japanese maglev devel- Cope, David B. 1993. Why does maglev need com-
posite materials? Intern. SAMPE Tech. Con! Proc.,
opment is based on the ideas of Powell and October 26-28, Philadelphia, (25): 141-151.
Danby (McConnell, 1993a). Friebele, Elaine. 1992. The ghost of LDEF-future:
In 1990, they came together with Morena to EOIM-3 continues the quest. LDEF Newsletter,
build a proposed 32-80 km (20-50 mile) route PO Box 10518, Silver Spring, MD 20914. May 15:
in Florida with inital design featuring a unique 19-22.
guideway with 8 inch spacing that dampens George, Peter and Dursch, Harry W. 1992. Low
noise and vibration, and completely enclosing earth orbit effects on organic composite materi-
als flown on LDEF. Paper read at
electromagnetic and dynamic fields. It will NASA-sponsored 2nd LDEF Materials work-
demonstrate full speed (300 mph) nonmechan- shop in October, Huntsville, AL.
ical switching capability. Unique to the Maglev Hiel, Clement, Dittman, Dan and Ishai, Ori. 1993.
2000 design is the ability to carry and individ- Designer's corner: composite sandwich con-
ually deliver heavy freight (such as tractor struction with syntactic foam core. Composites
trailors) to stations on major highways. 24(5): 447-450.
Specific utilization of composites will likely Hiller, M.W. et al. 1991. SSC quadrupole magnet
cryostat design alternatives. Adv. Cryogenic
draw upon Morena's SSC experience with low- Engineering 37A, New York: Plenum Press, pp.
cost fabrication processes such as RTM and 535-542.
resin film infiltration. Guideway and vehicle Hiller, M.W. and Waynert, J.A. 1992. A cryogenic
test articles are being built in 1997. The cost of support post for SSC quadrupole magnets. Proc.
Phase 1 is estimated at $4 million with comple- IISSC 4, New York: Plenum Press, pp. 297-300.
tion slated for June 1998; in Phase 11, a full Hiller, M.W. et al. 1993. A low-cost support post for
scale 120 ft long advanced transport freight sse quadrupole magnets and other cryogenic
applications. Paper read at International
and passenger vehicle would be constructed Materials Conference July 12-16, in
with composite outer secondary shell. Albuquerque, NM.
Kilpatrick, Mark c., Girard, Joseph D. and Dodson,
Kelly J. 1992. Design of a precise and stable
45.3 CONCLUSIONS
composite telescope structure for the UVCS,
The scientific journey continues, survlvmg April SPIE conference proceedings, Design of
budget cuts, continual goal assessments, and a Optical Instruments, (1690): 196-215.
Lubin, George and Dastin, Sam 1982. Aerospace
changing global culture. Composites are Applications of Composites. Chap. 28 of
intrinsic to reaching any progressive destina- Handbook of Composites, (ed. George Lubin),
tion, whether that be in space or moving New York: Van Nostrand Reinhold, pp.
people and goods over this planet. Science- 722-743.
based programs have proven invaluable in Lynch, Ted. 1993. The Technology Reinvestment
generating technology transfer for new prod- Project - the government and industry as part-
ucts that enhance essential quality of life, such ners. Advanced Composites, July/August: 36-42.
McConnell, Vicki P. 1989. Advanced composites
as medical care. Advanced materials are key make a case in space. Adv. Comp., Sept/Oct: 34-
enablers, then, for both public and private 45. 1990. National Aerospace Plane: a new
enterprise, and for the ultimate science of regime in flight. Adv. Comp., Nov/Dee: 37-45.
shaping the future. 1991a. Superconducting supercollider: design
changes. Composites Industry Monthly 19, item 3.
 References 981

1991b. Progress report: composites in space- trains. Paper 66-WA/RR-5 read at ASME
craft. Adv. Comp., July / Aug: 26-34. 1992a. Flight Winter Annual Meeting Nov. 27-Dec. 1, in New
vehicles of the future. Adv. Comp., Jan/Feb: York City.
28-34. 1992b. Tough promises from cyanate Shaw, P. 1993. Overview of maglev vehicle struc-
esters. Adv. Comp., May/June: 28-37. 1993a. tural design philosophy, material selection, and
Composites ride the rails. Adv. Comp., manufacturing approach. Paper read at Maglev
March/ April: 28-37. 1993b. Surviving space: a 93 conference May 19-21, at Argonne National
program and materials quandary. High- Laboratory.
Performance Composites, Nov/Dec: 22-28. 1993c: Sondericker, J.H. 1991. Alternate concepts for struc-
Focus on design: telescope structure. High- turally supporting the cold mass of a
Performance Composites, Nov/Dec: 54-56. superconducting accelerator magnet. Proc.
Morena, John. 1994. Invaluable materials science IISSC 3, New York, Plenum Press, pp. 175-189.
from Super Collider. High-Performance Stein, Bland A. 1993. LDEF materials overview.
Composites, Jan/Feb: 12-13. Second Post-Retrieval Symposium, LDEF-69
Nicol, Thomas H. 1992. SSC 50mm collider dipole Months in Space, NASA Conference Publication
cryostat single tube support post conceptual 3194, (ed. Arlene S. Levine), Vol. 3, pp. 741-789.
design and analysis. Proc. IISSC 4, New York: Tennyson, R.c. and Manuelpillai, G. 1993. Analysis
Plenum Press, pp. 747-755 .. of LDEF micrometeroid/ debris data and dam-
Nicol, Thomas H. 1993. Single tube support post age to composite materials. Second
thermal analysis and test results. Paper read at Post-Retrieval Symposium, LDEF-69 Months in
IISSC 5, May 6-8, in San Francisco, CA. Space, NASA Conference Publication 3194, (ed.
Peterson, Ivars. 1992. Flash tracks-building an eye- Arlene S. Levine), Vol 3, pp. 493-512.
piece for a particle accelerator. Science News 142,
July 4: 8-9.
Powell, James and Danby, Gordon. 1966. High
speed transport by magnetically suspended
CONSTRUCTION 46
Ever J. Barbero

46.1 INTRODUCTION and pultruded products listing concentration,

operating temperature, and frequency of
Composite materials are used by the construc-
exposure for a variety of chemicals. Most fiber
tion industry to replace or complement
reinforcements are usually corrosion resistant.
conventional materials such as steel and con-
crete. The main reasons for the use of . Unlike metals, composites do not produce
mterference with electromagnetic radiation.
composite materials are corrosion resistance,
The resin system can be selected to obtain very
~lectromagnetic transparency and weight sav-
low loss factors, but standard resin systems
mgs. Frequently, structural engineers take
are adequate for most structural applications.
advantage of more than one salient feature of
Buildings for electromagnetic interference
composites to formulate a design that is com-
(EMI) testing must be non-magnetic to avoid
petitive with an alternate design based on
attenuation and interference with the phenom-
conventional materials.
eno~ that is being measured. All computer
Corrosion resistance is the most important
eqUIpment, for example, must be tested in an
advantage of composites with respect to steel
EMI facility. Imaging equipment such as
for construction applications. The selection of
nuclear magnetic resonance (NMR) in hospi-
a composite material usually begins with the
selection of a resin that is capable of resisting tals .must be mounted in a magnetically free
the attack of a corrosive substance. The corro- enVIronment. An electromagnetically trans-
parent cover for communications equipment
sive agent can be anything from spring water
allows for the use of less-expensive, non-envi-
to sulfuric acid. Most composite manufactur-
ronmentally protected electronic hardware
ers provide corrosion resistance guides for
and reduced maintenance costs. Antennae
their products. For example, a table listing the
s~ctures that do not interfere with the signals
ma~imum operating temperature of isoph-
bemg relayed or received by the antennae
thahc polyester and vinylester resins as a
increase the efficiency of the system.
function of the chemical type and concentra-
While .weight saving is the main driving
tion is given by TUFSPAN Technical Data and
force behind the application of composites in
Design Guide (1991). Chemical resistance of
aerospace, it is not so critical in construction
common resins used in pultrusion to various
projects. However, reduction of structural
chemical and concentrations as function of
weight can be exploited as a secondary advan-
operating temperature is given by Pletcher
tage to help offset the higher cost of
(1991). Fibergrate (1992) supplies a Chemical
composites as compared to conventional
Resistance Guide for their molded fiberglass
materials. Lightweight structures require less
foundation and supporting structure. In the
case of bridges, a noncorrosive bridge deck
Handbook of Composites. Edited by S.T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 0 412 540207 can be built to replace existing steel reinforced
 Current applications 983

concrete decks that corrode rapidly under the vice for more than twenty years atest the excel-
attack from de-icing chemicals. For example, a lent corrosion resistance of fiberglass
pultruded deck was used to construct the reinforced composites. The following exam-
Wick Wire Run vehicular bridge on public ples of fiberglass reinforced isopolyester resin
road 26, in Taylor County, West Virginia (com- applications, described extensively in the
pleted August 1996). An added advantage of a excellent review by Adams and Bogner (1993),
composite deck would be the weight reduc- illustrate the feasibility of constructing com-
tion that supposedly would allow the user posite structures and using them for many
(highway department) to re-rate some bridges years. A three mile pipe of diameter 254 mm
for a higher live load without major modifica- (10 in) is reported in service since 1971. Fifteen
tions to the existing superstructure. The live miles of piping, carrying saline water with
load could be increased by approximately the temperature up to 50°C (112°F), pressure up to
same amount of dead load saved with the use 1()6 N 1m2 (10 bar), and exposed to sunlight has
of the composite deck minus adjustments for been in operation in Saudi Arabia since 1983
dynamic effects. Other applications where without problems. An old sewage duct was
weight savings are important are cladding of lined in 1971, then inspected in 1991 showing
buildings, rehabilitation of chimneys, etc. no sign of deterioration. More than 300 000
underground fuel tanks are in use in North
America alone. Some of these have been in
46.2 CURRENT APPLICATIONS
operation for more than 26 years without
Current applications of composites in con- problems. Internal or external lining of steel
struction can be classified by the major tanks has been common practice for more than
advantage of the composite material that is twenty years in the oil industry to protect and
exploited. The main ones are corrosion resis- reinforce the bottom of steel tanks that are cor-
tance and magnetic transparency. An roded internally because of corrosive
alternative classification may be based on the substances in the oil or externally because of
type of construction. Composites are used in contact with soil. Wine tanks have been in
the form of structural shapes (similar to steel operation for more than 20 years without
construction), as reinforcement for concrete, problems. Ducts carrying chlorine gases and
cables, and for rehabilitation of existing struc- sulphur dioxide, in use since 1962, and tanks
tures. Reinforcement of concrete may be in the holding hydrochloric acid, in use since 1964,
form of conventional reinforcement, pre- remain in perfect condition. A chimney
stressed concrete, or post-tensioned structures. exposed to organic chemicals, water vapor,
Rehabilitation applications include repairing and temperature up to 60°C, in operation since
deteriorated structures as well as increasing 1968 is reported. Chimneys are important con-
the load carrying capacity of sound structures struction applications because they are load
to re-rate them for higher load capacity. carrying structures designed for large wind
Some of the recent applications of compos- loads. A detailed account of recent applica-
ites in construction of civil infrastructure will tions of composites in construction, classified
be described in the next section. Further exam- by the type of construction, follows.
ples of applications can be found in previous
reviews (Barbero and GangaRao, 1991,
46.2.1 REINFORCEMENT OF CONCRETE
GangaRao and Barbero, 1991), professional
journals (e.g. SAMPE Journal, ASCE Journal) Concrete can be reinforced with fiber rein-
and edited books (Mufti, Erki and Jaeger, forced composites, with fibers mixed in the
1991a, 1991bi Iyer, 1991, Neale and Labossiere, concrete, or by polymers added to the concrete
1992). Applications to pipes and tanks in ser- mix. This article will concentrate on the use of
984 Construction

fiber reinforced composites, i.e. a combination Rebars

of fiber and polymer matrix, to substitute or
Reinforcing bars (rebars) have been used for
complement the traditional use of steel rein-
reinforc~g concrete structures that require
forcing bars (rebars) in concrete. The addition
ma~nehc transparency, such as imaging
of polymer or fibers to cementitious materials
eqUlpment at hospitals (Fig. 46.2). The major
sometimes better classified as a ceramics, bear~
expected application is to replace steel rebars
ing limited resemblance to regular, low cost
in co~crete bri~ge decks exposed to de-icing
concrete, falls beyond the scope of this article.
chemIcals. The first use of composite rebars for
The interested reader may wish to consult
a ~ehicular, public ~ridge is the McKinleyville
Mu~ty, Erki, and Jaeger (1991b, Chap. 2). Use
Bndge, located m West Virginia, USA.
?f fIberglass and carbon fibers mixed directly
Construction was completed August 1, 1996,
mto concrete has been reported (Mufty, Erki
by Orders Construction Co., of St. Albans, WY.
and Jaeger, 1991a). Seibu Construction Co. Ltd.
The bridge is a 54 m (177 ft) long concrete deck
used carbon fibers, produced by Mitsubishi
over steel stringer accommodating two lanes
Kasei Co., for the exterior walls of the
of vehicular traffic. The reinforcement of the
Kitakyushu Prince Hotel in Japan. Kajirna Co.
concrete is exclusively made of composite
also used carbon fiber reinforced concrete pan-
rebars with E-glass fibers. It was designed by
els for the exterior of its head office building in
the West Virginia Department of
Japan. Polymer concrete is used in a variety of
Transportation - Division of Highways (pro-
applications, such as in highway parapet walls
Ject S305-27/4-0.03) and the Constructed
developed by Morrison Molded Fiber Glass
Facilities Center at West Virginia University.
(Fig. 46.1). The installation of the lightweight
The project was supported by the Federal
polymer concrete panels reduces to anchoring
Highway Administration, composite rebar
the panels to an existing parapet, then pouring
concrete into the panel that act as a stay-in-
place concrete form. Composite reinforcement
of regular, low cost concrete can be done using
rebars, grids, pre-stressing tendons and post-
tensioning cables.

Fig.. 46.2 Fiberglass reinforced composite rebars

Fig. 46.1 Installation of polymer concrete panels as durmg construction of reinforced concrete struc-
highway parapets. (Courtesy of the Quazite ture. (Courtesy of Reynolds, Schlattner, Chetter, and
Division of MMFG.) Roll Inc.)
 Current applications 985

manufacturers, and the Corps of Engineers. sion strength and, for most practical pur-
E-glass polyester rods are cost competitiveposes, negligible tensile strength it is
with steel rebars at the present time. These advantageous to pre-stress concrete so that a
rebars have evolved from smooth pultruded state of compressive stress is created before
rods to engineered rebars with improved bond the actual load is applied. (The American
strength to concrete. Further improvements in Concrete Institute recommendation ACI 363 R
suggests a value for the modulus of rupture of
bond strength, tensile strength, and durability
are certain to occur in the near future. concrete Ir = 11.7 (fy/2 which is a very low
E-glass composite tanks have been in ser- value of tensile strength in relation to the
vice for over 20 years, in permanent contact compression strength of concrete, !c.) Then,
with highly acid environments and under con- applied loads only reduce the amount of com-
stant stress. The use of the SPI liner (ASTM pressive stress without producing tensile
D-3299 and D-4097) proved very successful to stresses. In this way, concrete cracks are pre-
vented, which in turn reduces moisture intake
protect glass fibers from attack by chemicals in
storage tanks. Also, composites have been in and degradation. High strength steel tendons
contact with concrete for many years without are currently used for pre-stressing concrete.
deterioration. Investigation of the possible Even though cracks are arrested by the pre-
stress, concrete is porous and water and
degradation of fiberglass rebars in concrete is
under way. The possibility of moisture intake chemicals may reach the prestressing tendons.
of Aramid fibers may be a problem for rebars Composite tendons may replace the steel ten-
dons for added durability. Pre-stressed
expected to be in service for at least 50 years.
Carbon fibers have superior properties but concrete is usually pre-cast at a factory, then
transported to the site. The tendons are pre-
their cost limits their potential as a replace-
ment for construction steel. stressed by a hydraulic jack, the concrete is
Rebars are placed in the form-work to pro- poured and left to cure. After the concrete is
vide for reinforcement of concrete in the sameset, the tendons are cut to remove the chucks
way as fibers are used to reinforce polymers. used to apply the pre-stressing load. Some of
Usually, two perpendicular orientations are the pre-stress is lost because of the compres-
used, with more rebar· area in one direction, sion, creep, and shrinkage of concrete. The
higher the elastic modulus of the tendon, the
according to the requirements of the structure.
Placement of rebars requires relatively inex- higher the pre-stress loss. For this reason, low
pensive labor. The grid-like structure necessary
modulus glass fiber tendons experience less
to reinforce concrete can be obtained pre- pre-stress loss than steel or carbon fiber ten-
assembled in the form of grids. Grids are dons. Pre-stress loads are high and induce
rectangular networks of rebars produced with large strains in the tendons, which may accel-
continuous fibers. Although more expensive erate degradation if the tendons are exposed
than rebars, they may be convenient if the to highly alkaline environment (Sen, 1992).
labor cost is high or the installation difficult, as
Two important requirements for pre-stress-
in the case of tunnels. Grids can also be pro-ing applications are a good bond strength
between the tendon and the concrete and
duced at or near the construction site by tying
rebars into a grid. The light weight of the availability of an effective temporary anchor-
age system. The anchor is not very critical
resulting grid makes transportation to the site
and installation simple. because it is temporary. Therefore, it does not
have requirements of non-corrosiveness, cost,
size, etc., but it must be able to transfer the
Prestressing
load and sustain it until the concrete sets.
Since concrete has a relatively high compres- Anchorage systems are reported by Noritake
986 Construction

et al. (1992), Mochida, Tanaka and Yagi (1992),

Kakihara et al. (1991), etc. for various types of
tendons based on glass, aramid, and carbon
reinforcement. Sen (1992) used the anchorage
developed by Iyer at South Dakota School of
Mines and Technology. Efficient utilization of
composite tendons calls for large pre-stressing
forces to be applied. Permanent levels of stress
in the tendons should be under the stress rup-
ture limit of the composite (see Section 46.3.4).
The pre-stress forces applied to composite ten-
dons may induce cracking of the polymer
matrix. Matrix cracking may be detrimental to
the fatigue life of the composite. Also, cracking
Fig. 46.3 Post-tensioned concrete-deck bridge over
facilitates the ingress of moisture that may
steel stringers. (Courtesy of Owens Coming.)
cause degradation of organic fibers or precipi-
tate alkaline reaction of glass (see Section
46.3.4).
Pre-stressing has been successfully used
by Iyer (1993) in a bridge deck (Fig. 46.3) pre- stirrups. The Sumimoto Construction Co. built
cast and pre-stressed, then transported to the a demonstration bridge in 1990 in Oyama
bridge site. The bridge construction was using three concrete box girders pre-tensioned
sponsored by Owens-Corning, AMOCO, with aramid composite rods, produced by
South Dakota Office of Economic Teijin Co. Grouted anchorages were used to
Development and South Dakota Cement pre-stress the rods at a permanent stress of
Plant; with the participation of Clark 70% of their static tensile strength. While the
Engineering, Glenn C. Barber and Associates, stirrups were also aramid composite rods,
FMG Engineering, Polygon, Shell, Central epoxy coated steel rebars were used as shear
Mix, Carloni Ace Hardware and South connectors between the girders and the rein-
Dakota Concrete Products. A 180 mm (7 in) forced concrete deck, which was reinforced
concrete deck was post-tensioned with glass with epoxy coated steel rebars. The deck was
and carbon fiber tendons which were subse- post-tensioned with aramid composite rods
quently grouted with epoxy based mortar. which were then permanently grouted in the
The deck is supported by steel girders 2.6 m deck. The Birdie bridge was built by Kajima
(8.5 ft) apart and the bridge spans 9 m (30 ft). Co. in 1990 at the Ibaragi Prefecture using a
Aramid bars were used in a standard pre- variety of composite materials. Carbon fiber
stress pre-cast concrete factory to build composite rods, produced by Mitsubishi
concrete barges (Noritake et ai., 1992). At least Kasei, were used to anchor the abutments. Pre-
10 concrete bridges have been constructed in cast concrete panels reinforced with vinylon
Japan with some kind of composite reinforce- short fibers are connected with a grid of car-
ment (Mufti, Erki and Jaeger, 1991a). The bon fiber composite cables produced by Tokyo
Shinmiya bridge is a pre-tensioned concrete Rope Manufacturing Co. The bridge is pre-
bridge built in 1988 in a coastal area. The pre- stressed with aramid composite flat bars
cast concrete girders were pre-stressed with commercialized by Nippon Aramid Limited.
seven strand, carbon composite tendons pro- Half of the aramid bars are used as pre-ten-
duced by Tokyo Rope Manufacturing Co. sioning tendons and the remaining act as
Epoxy coated steel rebars were used for the post-tensioning tendons,with a permanent
 Current applications 987

stress of one third the static tensile strength. on the transfer of the axial load in the cable,
by shear in the grouting material, to the
anchor. Attaching the anchor to the tendons is
Post-tensioning
a labor intensive process. The resin used to
Post-tensioning of concrete with steel or com- bond the tendons to the anchor may creep,
posite tendons is performed to induce a state leading to loss of pre-stress. This anchor was
of compressive stress in the concrete similarly used to partially post-tension several bridges.
to pre-stressing, but post-tensioning is per- Cases have been reported where individual
formed at the construction site. A hole or tendons slipped and broke at the anchor dur-
some kind of access is left in the concrete to ing post-tensioning at the bridge site. Porter
thread the post-tensioning cables through. and Barnes (1991) report several anchorage
Tensile force is applied to the tendons against systems. To avoid the problems that all pot-
the concrete structure before the structure is ted type anchors have, Ahmed and Plecnik
loaded by the service loads. The anchorage (1989) developed a filament wound cable
system is subject to severe requirements. First, where continuous fibers are wound around
it must sustain the tension load for the whole the end eyes of the cable. Each cable must be
life of the structure, which requires very care- custom made for the required length, but the
ful design against creep in the anchor. Second, problems of potted type anchors are elimi-
the anchor should be resistant to corrosion as nated.
the tendon itself if the non-corrosive proper- The Bachigawa-Minami-Bashi bridge (Koga
ties of the system are the objective of replacing et ai., 1992) uses both pre-stressing and post-
steel tendons. The reduced weight of the ten- tensioning carbon fiber tendons. The
dons is not very important because the heavy Schiessbergstrasse bridge in Germany and the
weight of the concrete structure. While the Notch bridge in Austria (Wolf and Miesseler,
stress losses of fiberglass tendons are smaller 1992), use partial post-tensioning with fiber-
than those of steel cables, it is difficult to jus- glass cables. Some of the tendons that form a
tify the use of more expensive and novel cable contain sensors (copper wire or fiber
composite tendons for this reason alone. optical gauges) to monitor the strain level,
Therefore, corrosion resistance is the main integrity of the tendon, and location of even-
objective of using composite tendons. In this tual damage. A demonstration bridge was
case the anchorages should be resistant to cor- built in 1990 by the Sumimoto Construction
rosion. Also, the structure to be post- Co. in Oyama, Japan, using a single concrete
tensioned is usually made of reinforced con- box girder, post-tensioned with aramid com-
crete (although the reinforcement is not posite rods produced by Teijin Co. The
sufficient to carryall the load). If corrosion is internal tendons are placed in a parabolic
a problem, the conventional reinforcement housing in each web and post-tensioned with
(not pre-stressed or post-tensioned) may also steel grouted anchorages to a permanent stress
have to be made of composite material rebars. of 25% of the static tensile strength of the ten-
Anchors for post-tension applications are dons. The external tendons are placed at the
still being developed and evaluated. bottom of the box girder and post-tensioned,
Meisseler and Preis (1989) report on an with grouted anchorages built with composite
anchor developed to hold glass reinforced casings for added corrosion protection, to a
tendons. The anchor is of the potted type, in permanent stress of 10% of the static strength.
which the tendons comprising a cable are Besides the composite post-tension cables, the
spread at each end and potted in a steel regular reinforcement of the girder uses epoxy
anchor with some grouting material, usually coated rebars. Further details are given by
a polymer. The potted type anchors are based Mufti, Erki and Jaeger (1991a).
988 Construction

Rehabilitation stressed and glued laminated deck (Davalos

and Salim, 1992). Some of the timber beams
There is ~ignificant interest in using compos-
were reinforced with carbon reinforced epoxy
Ite m~t~nals for rehabilitation and upgrading
plates. Rostasy, Hankers and Ranish (1992)
of eXIstIng structures. These structures may
report on the reinforcement of the
have been damaged as a result of corrosion of
Kattenbusch bridge (Germany) using steel
the ste~l reinforcement or they may need
and glass reinforced composite plates. The
upg~admg to new seismic standards, larger
main reason cited for the selection of com-
traffIc loads and volume, etc. Interest in com-
posite plates is the documented corrosion of
posites is motivated by the ease of
the steel plates at the bonded interface.
installation of the reinforcement, its corrosion
Co~posite jackets have been proposed to
resistance (mainly at the bond surface), and
retrofit columns for seismic solicitations. Scale
the ~ossibility of selecting from a variety of
samples (40%) were tested (Priestley, Seible
e~a~tIc modulii which improves the compati-
and Fyfe, 1992) and promising results were
bIlIty between the reinforcement and the
found. A fiberglass epoxy jacket is built
existing structure Meier et al. (1992),
around the hinge region of the column. On an
Saadatmanesh and Ehsani (1991), and others
~ctive confinement jacket, hoop stress is
have demonstrated the feasibility of rehabili-
mduced by pressure grouting the space
ta~ing concrete, steel, and wood structures by
~e~een the ~oncrete column and the compos-
remforcement with composite plates. Meier et
Ite Jacket WIth either epoxy or concrete. A
al. report a variety of failure modes that may
passive confinement is obtained when no
be encountered in concrete beams reinforced
pressure grouting is used. Passive confine-
with composite plates. The reinforcement
ment is used in regions of high compression
may be a cured composite plate bonded to
stress, over the region of lap-spliced longitudi-
the beam (Munipalle, 1992), or a room-tem-
nal bars, or for regions of high shear stress.
perature cure prepreg directly applied to the
Significant concrete dilation and consequent
structure. The reinforced plates can be
lo~gitudinal mic~ocracking is necessary to
applied with or without pre-stressing. While
activate the confinement effect in passively
pre-stressing the composite plate increases
encased concrete. Active confinement has the
the efficiency of the reinforcement, it also
advantage that the confinement effect is
complicates the rehabilitation process. Since
always available.
del~mination of the reinforcing element is of
C~eys ~ave been rehabilitated in Japan
major concern, special reinforcement details
by MltsubIShI Kasei using carbon composite
are used at the ends of the reinforcement.
tape and strands (Mufti, Erki and Jaeger,
Composite materials can also be bonded to
1991a). The composite tape is applied along
the sides of beams to improve the shear
the length of the chimney to provide addi-
strength. While using prepreg materials, both
tional bending strength. The strand is wound
shear strengthening and resistance to delami-
around the chimney to provide hoop rein-
nations can be obtained by partially or
forc~ment. A fire protection mortar is finally
completely wrapping the beam with the rein-
applied to protect the polymer composite from
forcement. Meier et al. (1992) reports on the
fire~ to limit the heat gain, and to avoid degra-
rehabilitation of two bridges in Switzerland.
dation of the polymer matrix that otherwise
The Ibach bridge was repaired using carbon
would occur if the carbon composite were
fiber reinforced plates bonded to the concrete
exposed to the environment.
bridge. The rehabilitation of the timber
bridge in Sins involved replacement of the
wood pavement with transversely pre-
 Current applications 989

46.2.2 STRUCTURAL SHAPES Platforms

Composite structural shapes resemble steel Industrial platforms built entirely of pul-
hot- and cold-rolled structural sections. truded fiberglass reinforced plastics are
Composite shapes are produced by pultrusion widely used because of their corrosion resis-
in a variety of sizes and shapes (Creative tance. An area of application of special interest
Pultrusions, 1989; MMFG, 1992). The rein- is offshore platforms for oil production. Two
forcement of choice is E-glass mainly because views, from above and below, of a fiberglass
of low cost. Most shapes are produced with well bay platform are shown in Figs 46.4(a)
isophthalic polyester and vinylester, but some and (b). Steel well bay platforms corrode
pultrusions using epoxy, phenolic, and even quickly in the marine environment. Painting
thermoplastics are available at higher cost. jobs are difficult because sand blasting may
Even for the most common polyester and release paint into the ocean unless costly pre-
vinylester, there is a wide variety of resins sys- cautions, such as scaffolding, are taken. Old
tems, differentiated by mechanical properties, paint may be lead-based and cannot be
thermal behavior, and cost. While vinylesters stripped unless it is completely captured and
are considered to have superior corrosion dumped at a hazardous waste site. Well bay
resistance and mechanical properties than platforms are installed after the production
polyesters, high grade polyesters may match wells are in operation. The area is usually clut-
the mechanical properties of vinylesters. tered with equipment, which makes difficult
Structural shapes are used primarily the access of heavy lifting equipment needed
because of their excellent resistance to chemical for installing heavy steel platforms, and weld-
degradation. In the case of electromagnetically ing cannot be done without shutting down oil
transparent structures, structural shapes are production. Fiberglass platforms weigh typi-
routinely used for building construction in cally 30% of their steel counterpart, allowing
very much the same way as steel shapes. for installation by a smaller crew, in less time,
However, E-glass reinforced shapes have a with less demand for space for lifting equip-
much lower modulus of elasticity than steel, ment, and reduced transportation costs to the
which causes design problems when direct site. The composite platforms are assembled
replacement of steel shapes by composite with mechanical connections eliminating the
shapes is attempted. Specialized sections, with need for welding and they are virtually main-
cross sections different from steel shapes, are tenance free in the corrosive marine
available for use in building systems with con- environment, without need of painting. A 6 m
struction characteristics (e.g. joist spacing) x 12 m (20 ft x 40 ft) well bay platform was
very similar to steel frame buildings installed in 1986 on Shell's Southpass 62 pro-
(Composite Technology, 1992). Unlike steel duction platform in the Gulf of Mexico by a
shapes, the mechanical properties of compos- crew of four in two days and it is estimated
ite structural shapes largely depend on the that a similar steel platform would have
internal reinforcement, and the thermal and required five days by a crew of eight because
corrosion response varies drastically with the of the lack of room for heavy lift equipment.
resin system used. Since the reinforcement
and the resin systems used are not standard-
Building systems
ized, very different set of properties are
possible for identical cross sections, from dif- Complete building systems, including the
ferent manufacturers. structural frame, cladding and roofing are
available in fiberglass reinforced composites
(Composite Technology, 1992). Electromagnetic
990 Construction

(a)

(b)

Fig. 46.4 Well bay platform in an off-shore oil production facility (a) seen from above, (b) seen from below.
Composite materials manufactured by MMFG. (Courtesy of MMFG.)
 Current applications 991

Interference (EM!) testing buildings must be

built of materials free of magnetic interference,
with all metals ruled out and the use of timber
limited by the need to use steel connectors.
Pultruded structural shapes, fiberglass
cladding, and foam core panels offer an attrac-
tive alternative since all components can be
connected with fiberglass bolts and or glued
together to form an electromagnetically trans-
parent structure. Innovative combinations of
glued laminated timber (GLULAM) with com-
posites have also been exploited. A computer
testing facility (Fig. 46.5) was built for IBM in
Poughkeepsie, NY, by Corflex International Inc.
of Warren, OH, and Haines Lundberg Waehler
of New York, NY, using pultruded structural
shapes from Creative Pultrusions Inc. (1989).

Fig. 46.6 Fiberglass spire installed atop the 55-story

Nations Bank in Atlanta, GA. Composite materials
manufactured by MMFG. (Courtesy of MMFG.)

Fig. 46.5 Computer testing facility using pultruded Pedestrian bridges

structural shapes and fiberglass reinforced panels.
(Courtesy of Creative Pultrusions, Inc.) A number of pedestrian bridges have been
built with composite materials. Early bridges
are reported by Meier (1991), Barbero and
The Nations Bank building in Atlanta, GA, GangaRao (1991) and others. While a few
shown in Fig. 46.6 features a 11 m (36 ft) tall, bridges have been built by hand lay-up, the
all fiberglass spire at its top. Electromagnetic majority are built with pultruded sections.
transparency of the composite material used Concrete bridges partially reinforced with
allows the spire to house valuable communi- composite tendons are discussed in previous
cation antennae. Light weight, molded-in sections on concrete reinforcement. In this sec-
color, and timely delivery were cited as advan- tion, we concentrate on bridges where most of
tages of composite materials for this highly the materials used in the construction of the
visible application. The spire is mounted bridges are composites. Because of cost, E-glass
312 m (1023 ft) above ground, and it is reinforcement has been used most.
designed to sustain a wind pressure of 550 kPa Johansen et al., from E. T. Techtonics of
(80 psi). Philadelphia, PA, report on the construction of
992 Construction

Fig. 46.7 Devil's pool bridge is a 15 m (50 ft) bow-string truss built with standard pultruded structural
shapes. (Courtesy of E.T. Techtonics.)

three bridges of 6.1 m (20 ft), 9.75 m (32 ft), of the pultruded glass-polyester deck (Fig.
and 15.24 m (50 ft) (Fig. 46.7) using their 46.8). The bridge is 113 m (370 ft) long and
design and construction method, called 2.23 m (7.3 ft) wide, with a main span of 63 m
PRESTEK. Given the low stiffness of E-glass (206 ft). The deck and towers are constructed
reinforced pultruded composites when com- with Maunsell's interlocking panels (ACCS
pared to steel, conventional steel or concrete System by Designer Composites Technology
designs are not efficient when implemented Ltd), pultruded with 70% volume of fiberglass
with composites. The PRESTEK system uses a reinforcement and isophthalic polyester resin,
beam-truss geometry of pultruded tubes pre- then bonded together with an epoxy adhe-
stressed with aramid or steel cables. The three sive. The cable stays are Parafil ropes, a
bridges are a king-post truss, a queen-post Kevlar fiber core in a polyethylene sheath.
truss and a bow-string truss respectively. The bridge deck was designed to be modular
Maunsell Structural Plastics in Beckenham, so that every component could be handled by
Kent, England, with the participation of GEC hand. The heaviest module, a 6 m (19.7 ft) by
Reinforced Plastics (pultrusion manufactur- 0.6 m (2 ft) plank weighed only 66 kg (145Ib).
ing), Scott Bader Co. Ltd (resin supplier), The weight of the deck is only 150 kg/m
Vetrotex UK (glass reinforcement), Ciba Geigy (100Ib/ft) and each tower weighs only
Plastics (adhesives), Linear Composites 2500 kg (5511lb), facilitating erection and
(Parafil cable stays), R. O'Rourke and Sons reducing the cost of the foundation. The com-
(construction management), and University posite system has class 2 fire resistance rating.
of Dundee, constructed a pedestrian bridge
over the river Tay in Aberfeldy, Scotland. The
Bridge enclosures
pultruded deck is cable stayed from two A-
frame towers, 17.5 m (57 ft) high, to provide The A19 Tees Viaduct in Middlesbrough, UK, is
for adequate stiffness despite the low stiffness a steel-concrete bridge with 117 m (383 ft) span
 Current applications 993

Fig. 46.8 Pedestrian bridge across the river Tay in Aberfeldy, Scotland, features an all-composite deck cable
stayed with Parafil cables from two composite A-frame towers. (Courtesy of Maunsell Structural Plastics
Ltd.)

(Head, 1988) where rapid deterioration of the Fig. 46.9, over the railway station. Pultruded
steel plate girders was taking place. plank and connector sections are joined to form
Maintenance (e.g. painting) and rehabilitation a floor system suspended from the steel girders
work are difficult since the viaduct spans over of the bridge. Pleasant appearance was
railroad tracks. A composite material enclosure required since the bridge is located in a resi-
was built in 1989 to prevent further deteriora- dential area and over the railway station. Light
tion by isolating the bridge from the weight and low maintenance costs were cited
environment and to facilitate maintenance and as additional advantages of composites for this
rehabilitation. A floor area of 16 000 m 2 application.
(172 200 ft2) using 250 metric tons (275 US tons)
of composite material was created under the
Cooling towers
bridge by interlocking pultruded panels
wrapped around the steel girders of the bridge The resistance of composite materials to
to create a box. The enclosure system was humidity and creep under sustained loads has
designed by Maunsell Structural Plastics Ltd, of been demonstrated by their successful appli-
Beckenham, Kent, UK. The system uses inter- cation to cooling towers, in operation for more
locking panels designed by Maunsell and than twelve years (Fig. 46.10). Cooling towers
fabricated by GEC Reinforced Plastics Ltd, are permanently loaded with the heavy
while the main contractor was Fairclough weight of the ceramic filling used for cooling.
Construction Ltd. The enclosure system has These towers are built by Ceramic Cooling
been also used in the construction in 1992 of the Towers of Forth Worth, TX, entirely of com-
enclosure to the Bromley South Bridge, seen in posites except for the ceramic filling. The
994 Construction

':. ----- .:-" ~ - .

~.(' :. .p.7'i "'stL. '-: .~ - .. A -
~ .. ~T " •••

... --
~
-
___ - '"_ • - _ . . . . "0 .... ---

. .

Fig. 46.9 Enclosure of the Bromley South bridge is accomplished with interlocking pultruded panels
bonded together into a floor system and suspended from the steel girders of the bridge. (Courtesy of
Maunsell Structural Plastics Ltd.)

composite material is in permanent contact tributed by Designer Composites Technology

with hot water and humid hot air circulating Ltd, UK, is shown in Fig. 46.11. Finger piers of
through the tower. Reduced maintenance of up to 15 m (49 it) are built with cellular inter-
the tower is cited as an important advantage of locking pultruded panels, joined to foam filled
using composites. Light weight of the struc- flotation units. The surface of the pontoon is
ture of the tower is an additional advantage covered with a non-slip polyurethane coating
when the towers must be located on top of which is acid, solvent, and heat resistant.
buildings. Construction of large industrial
cooling towers motivated the development of
46.3 DESIGN CONSIDERATIONS
the UNILITE Modular System which takes
advantage of modular construction to reduce Current composite design practice emphasizes
cost. Composites facilitate modular construc- simultaneous design of the structure (beams,
tion because of the variety of shapes that can plates, frames, etc.) and the material (compos-
be easily produced. ite) for optimum performance. Design practice
for composites used in construction differs
from aerospace applications in the sense that
Marine construction
standardization of components is required
Potential applications for composites in the because of cost and safety. Therefore, design is
construction and rehabilitation of marine and divided into member design, usually per-
waterfront constructed facilities include formed by the manufacturer, and system
docks, piers, harbors, etc. An all-composite design, carried out by the structural engineer.
pontoon designed by Maunsell Structural Codes of practice do not yet exist for structural
Plastics Ltd, of Beckenham, Kent, UK and dis- design with composites, but they are being
 Design considerations 995

deflections are controlled by the bending stiff-

ness D, equivalent to EI for steel beams. Since
composite beams have different values for
modulus of elasticity at various points in the
cross section, it is not possible to completely
define the stiffness with respect to both axis of
bending by the product of the modulus and
the corresponding moment of inertia (EI and
E1xx ) Instead, two bending stiffnesses V; and
Dx are defined, with respect to the strong axis
and weak axis respectively. Shear deforma-
tions are neglected for steel beams because the
shear stiffness of steel is high (G / E = 0.4) while
for composites is low (G/E < 0.1). The shear
stiffnesses of a composite beam with respect to
the two axis of bending are denoted Fy and Fx'
The values of the bending, shear, and axial
stiffness (Dy, /CFy, A z ) can be obtained from
tables supplied by the beam manufacturer,
obtained experimentally (Bank, 1989), or com-
puted if detailed information about the
constituent materials is known (Barbero,
1998). Maximum deflections can be computed
using the formulae in Fig. 46.12. Experimental
values of shear stiffness usually refer to the
Fig. 46.10 All-composite cooling towers like this, product of /CFy where /C is the shear correction
developed and manufactured by Ceramic Cooling factor.
Tower, have been in operation for more than twelve
years.

developed. Some methods of analysis and

design recently developed are described in
this section.

46.3.1 BEAMS
Beams are the most common structural com-
ponent in civil engineering applications. Both
deflection and strength are equally important
in the design of composite beams. Composite
beams are thin-walled and composed of an
assembly of flat panels. Most beams are pris-
matic but they can have taper. They are
Fig. 46.11 The pontoon pier shown is built of inter-
produced by pultrusion, filament winding, locking pultruded panels and foam filled flotation
hand lay-up, automated lay-up, etc. devices. A polyurethane coating protects the surface
Deflection of composite beams has two from acid, solvents, and heat. (Courtesy of
components, bending and shear. Bending Maunsell Structural Plastics Ltd.)
996 Construction

1* it *it iti
w

~ L ~
IttttttttL w

~ L ~

3
.I\~ = ~ PI: + i PL
VIf. 48 Dy 4 KFy

_ 19 P~ + 1 PL
cS~ - 1536 Dy "8 KFy
IE-~--L ---~~

CK~~--L---~~!~
M ~
~ - SDy
_ MI.?

i P

11E--~ _~~
-L

4 2
Oe = wL + wL
SDy 2KFy
IE-~ --L --~~
Fig. 46.12 Center deflection of composite beams including bending and shear effects.
 Design considerations 997

The significance of the shear deflection with the length of the member. Lateral forces (e.g.
respect to the bending deflection varies with wind forces) and bending moments (e.g.
the span, the larger the span the lesser the influ- eccentric loading) are considered secondary
ence of shear. Sometimes, properties of beams forces, which are dealt with separately.
are reported without distinction between the Column performance is limited by one of two
bending and shear components, using an failure mechanisms, crushing and buckling.
ap~arent value. of m~dulus EaI;p = D/ Iyy ,where Crushing is the failure of the material because
Dy IS the bending stiffness WIth respect to the of excessive compressive stress, similar to
strong axis. The bending stiffness with respect yield of steel. Buckling is more frequent in
to the weak axis Dx cannot be accurately composite columns because thin-walled sec-
obtained as Dx = Eap Ixx. The apparent stiffness tions are preferred. A thin-walled section may
Eaw is then used in the classical deflection for- experience at least three different types of
mulas (for steel) that do not account for shear buckling, which are described next.
deformations. The reported values are usually Long and slender columns fail in a global
based on three-point bending tests, performed sense when the axial load reaches a critical
at the factory with a specific span, which is sel- value PEU For load values lower that the criti-
dom reported. The results of using the classical cal load the column remains straight. When
(steel) deflection formulas for spans or loading the load reaches critical value, the column
other than that of the test are only approximate. experiences sudden lateral deflection. The
The main modes of failure of beams in axial stiffness after buckling is much lower
bending are: (a) compression crushing of the than the stiffness before buckling. Therefore,
compression flange; (b) local buckling of the the lateral deflections are quite large and they
compression flange; (c) tensile rupture of the usually precipitate another mode of failure
tension flange; (d) shear failure of the web; like crushing of parts of the cross section,lead-
and (e) web buckling. Since each part (panel) ing to collapse.
of the cross section can be built with different The concept of slenderness allows us to
materials, the failure mode can be controlled compare members of different cross sections
by design. Local buckling modes can be elimi- and column lengths for their tendency to
nated by increasing the thickness and buckle. For a composite column, the slender-
choosing the fiber orientations properly ness is defined as

Le~( ;~2 )
(Barbero and Raftoyiannis, 1993). The com-
pression strength of composites is lower than A= (46.1)
the tensile strength. Therefore, a symmetric
section is not the most efficient cross section. where PL is the local buckling load, D is the
Symmetric sections fail in the compression bending stiffness, and Le is the effective
flange first. Tensile failure may occur in length of the column, which is used to
unsymmetric sections, when the compression account for different end conditions (Gere
flange works with composite action with a and Timoshenko, 1990, p. 589).
deck, etc. Shear failure is less likely to occur in Short and stubby columns, which have a
sections with multiple webs. Incorporation of low slenderness value, are less likely to buckle
off-axis fibers (cloth, mats, etc.) increases the in a global mode as described previously.
shear strength. However, individual parts (flange or web) of
the cross section may buckle locally. Local
buckling is very likely to occur in composite
46.3.2 COLUMNS
columns because they are commonly thin
Columns are structural members subjected walled. The compressive stress required to trig-
primarily to compression forces acting along ger local buckling increases with the thickness
998 Construction

of the cross section and the local stiffness of the 0.5 and 1.5 show some type of interactive phe-
material. The narrower the flange or web nomenon. The interaction occurs between the
under consideration, the higher the local buck- local mode, the global mode, and crushing.
ling load PL' The wavelength of local buckling Interaction results in lower buckling loads
is independent of the length of the column for than those predicted by any of the modes act-
columns of practical lengths. ing alone.
Flange buckling of open section members is There are many situations of practical inter-
triggered by pure compression. It occurs during est for which the buckling loads required to
compression of columns and on the compres- produce two or more failure modes (Euler,
sion flange of beams in bending. Web buckling local, crushing) may be very close. In this case
is initiated by shear and it occurs during bend- the failure modes interact. That is, the proxim-
ing of beams. Web buckling of open section ity of the stresses to more than one mode of
columns is not common because most open sec- failure causes the structure to fail at a lower
tions have wide flanges that buckle first. There stress value that predicted by either of the
is of course no distinction between flanges and modes involved should they be acting sepa-
webs in closed section columns. Unlike steel rately. Euler and local modes interact to give
structural shapes, composite closed sections are an overall strength deterioration. Interaction
easier to produce and structurally more effi- must be taken into account because the
cient than open section members. strength values predicted by either isolated
Local buckling can be prevented by choos- mode fit are not conservative. The failure load
ing the section geometry and material of a column (PCR) taking into account local,
properties. Global buckling can be prevented global, and interaction phenomena simultane-
by a combination of section geometry, material ously can be obtained from the following
properties and bracing. If all buckling modes design equation (Barbero and Tomblin, 1993)
are prevented, the strength of the member is
limited by the crushing strength of the mater- = 1+1j').} _.1[(1+1/A.2)2_~1
ial itself, which plays a role similar to the yield
PCR
PL 2c ~ 2c cl2 (46.2)
strength of metals. The crushing strength is a
material property which is independent of the The column properties needed to use this
thickness of the flange or web and the geome- design equation are: the local buckling load PL;
try of the section. As for any composite the interaction constant c; and the bending
property, it depends of the constituents (fiber stiffness D, which along with the length of the
and resin) and the arrangements of the fibers column enter in the computation of the slen-
inside the material (orientation, fiber volume derness A. (46.1). All these properties can be
fraction, stacking sequence). The crushing determined experimentally or predicted ana-
strength is usually determined experimentally lytically. The bending stiffness D and the local
but the main factors that influence its value buckling load PL' can be predicted analytically,
can be highlighted by predictive equations while an analytical study of the interaction
(Tomblin, 1994; Barbero, 1998). phenomenon is presented by Raftoyiannis
A slender column buckles in a global (1993). The design equation does not have any
(Euler) mode. A not-so-slender column may safety factor included. Typical properties of
fail in a local buckling mode. A thick-walled wide-flange pultruded structural shapes are
stubby column may fail due to crushing. given in Table 46.1 where the interaction con-
Columns with a slenderness ratio less than 0.5 stant is c = 0.84 for all sections reported.The
fail in a local buckling mode. For slenderness length at which maximum interaction occurs
larger than 1.5 the mode is purely global is denoted by C.
(Euler). Columns with slenderness between
 Design considerations 999

Table 46.1 Column properties obtain the Euler buckling load PEU as the
inverse of the slope in the til P vs. ti plot,
Section PL (kN) D (kN em2) L" (em) where ti is the lateral deflection and P is the
102 x 102 x 6.4 223.25 6094.67 105.9 load.
152 x 152 x 6.4 175.12 20954.42 221.5 Interaction testing can be performed with
the same setup described for global (Euler)
testing. The objective is to determine the inter-
action constant c in (46.2). Once two or more
modes of failure interact, the Southwell
method cannot be used. Therefore, only the
Testing of short columns is performed to iden- maximum collapse load is reported for tests
tify the local buckling load PL in the column performed on columns having slenderness
design (46.2). Buckling of the flanges is seldom values between 0.5 and 1.5. The collapse load
a sudden phenomenon as described by the is lower than the critical load that would occur
theory because of the imperfections of the should any of the modes involved act isolated
material. Therefore, flange lateral deflections from the others, as the experimental data
are observed from the onset of the test and clearly indicates. The test is conducted with a
they grow as the load is applied. The buckling column length that exhibits maximum interac-
load Pu to be used in the design corresponds tion (L'), which occurs for a column
to the asymptote of the hyperbolic curve of slenderness A. = 1. The interaction constant is
load vs. deflection. Since it is not practical to computed as c = (q + s -l)lqs, where q = P IPu
conduct the test up to the large deflections s = PIPEU' P is the collapse load, PL is the local
needed to realize the asymptotic value of load, buckling load (predicted or previously mea-
a data reduction technique (Tomblin, 1991) is sured on short columns) and PEU is the Euler
used. buckling load (predicted or previously mea-
Columns having slenderness larger than 1.5 sured.)
will buckle in a global mode (Euler). The Euler
buckling load is controlled by the bending
46.3.3 REINFORCEMENT OF CONCRETE
stiffness D. A column with pinned-pinned end
conditions is subjected to an axial load (prefer- Reinforcement of Portland cement concrete
ably under axial displacement control). Pinned can be accomplished with composite reinforc-
end conditions are the only conditions that can ing bars (rebars) instead of steel bars to
be achieved with any degree of certainty. A minimize the corrosion of steel and cracking of
100% degree of fixity required by a clamped concrete caused by the expansion of the cor-
end condition is not achievable on composite roding steel rebars. Composite rebars should
columns because of the difficulties associated have good bond with concrete and adequate
with connecting composites. Data from corrosion resistance. Pultruded rebars are the
pinned-pinned columns can be used for other most common and least expensive alternative
end conditions by using the effective length for the reinforcement (not pre-stressed) of con-
concept (46.1). Weak axis tests are simpler to crete. Pultruded rebars have an angle
perform, since strong axis tests require the use overwrap and I or a sand coating to improve
of lateral support. Because of the imperfec- the bond with concrete. Fiberglass rebars,
tions in the material and loading fixture, the aramid rebars, seven-wire carbon cables,
load deflection plot has an hyperbolic shape, grids, and even gratings have been used in
the buckling load being given by the asymp- research studies as reinforcements of concrete.
tote of the plot. A data reduction technique Fiberglass rebars have lower modulus of
known as the Southwell method is used to elasticity than steel rebars, causing larger
1000 Construction

deflections than in steel reinforced concrete Creep is observed as increasing elongation

beams. Typically, the deflection of reinforced under constant load. Relaxation is the reduction
concrete (RC) beams is controlled by two fac- of stress over time for a constant elongation.
tors: the amount of concrete that remains Permanent levels of stress in the composite
un-cracked (essentially in compression); and should be under the stress rupture limit (also
the modulus of elasticity of the reinforcement. called static fatigue) of the composite. Martine
The extension of cracking in the tension side of (1993) reported stress rupture of E-glass rein-
RC beams depends largely on the strength and forced composite at 10 000 h at 58% of the initial
uniformity of the bond between the rebar and strength. Glaser, Moore and Chiao (1983) report
concrete. Good bond translates into uniformly survival of S-glass reinforced epoxy specimen
spaced cracks in the tension side of the beams. to a 10 year sustained load test: 90% of the spec-
Therefore, a larger area of concrete remains imens survived at 50% of static strength; 98%
un-cracked, thus contributing to the bending survived a 40% loading; and 100% survived a
stiffness and limiting deflections. Fiber rein- 35% loading. To account for stress rupture and
forced composite rebars have a linear other factors in the design of fully overwrapped
stress-strain behavior up to failure, thus they pressure vessels for compressed natural gas,
experience a brittle behavior when compared stress ratios (burst over service fiber stress) of
to steel rebars that have a yield plateau (i.e. 3.5 for fiberglass, 3 for aramid, and 2.25 for car-
plasticity) before they rupture. Brittle behavior bon reinforced composite were proposed by the
is not convenient from a safety point of view American National Standards Institute (AGA
because of possible catastrophic RC member NGV2 1992). These are indications that the sta-
failure. However, concrete reinforced with tic strength of fiber reinforcements cannot be
fiberglass rebars experiences very large deflec- utilized for long duration loads like those
tions up to failure, larger than comparable encountered in construction applications.
steel reinforced beams in their plastic regime, Glass fibers, if unprotected, deteriorate
because of a combination of low modulus of when exposed to an alkaline environment.
elasticity and high tensile strength of the fiber- Therefore, glass reinforced composites rely on
glass rebars. the protection provided by the resin to resist
No code or standard regulating the design the attack of alkaline environments. Alkaline
of RC beams, reinforced with fiberglass com- degradation is of particular concern in con-
posite reinforcing bars (rebars) is in place at crete reinforced with fiberglass composites
the time of this publication. Faza and since the concrete mix is alkaline. While con-
GangaRao (1993) suggest that it is possible to crete alkalinity is well documented at the time
proceed along the lines of the American concrete is poured, only now are studies being
Concrete Institute guidelines ACI 318-89 com- conducted to investigate the level of alkalinity
plemented by the recommendations ACI 363 of concrete with time. Water is needed to
R, properly modified to account for the prop- establish a link between the glass fibers and
erties of the fiberglass rebars. the alkaline concrete environment. If water is
present, the link can be established through
cracks in the resin. Permeability of the poly-
46.3.4 ENVIRONMENTAL EFFECTS
mer matrix resulting from voids or moisture
Dead loads (e.g. the weight of the structure) diffusivity of the polymer are being consid-
will be applied to the composite material for the ered as possible additional mechanisms that
life of the structure. Sustained loads induce two may place the alkaline concrete environment
major effects in composites, creep and stress in contact with the glass fibers.
rupture. Creep and relaxation are two alterna- S2-glass reinforced Shell Epon 9310 com-
tive descriptions of the same phenomenon. posite rods produced by pultrusion were used
 References 1001

by Sen, Issa, and Mariscal (1992) to pre-stress many years of service, mostly in adverse con-
concrete piles. Seven 3 nun (0.125 in) diameter ditions. Composites are being used
rods were twisted with one turn every 30 cm extensively for applications where the advan-
(12 in) to create a seven wire strand. The tages significantly justify their higher initial
strands were used to pre-stress concrete piles cost. Composites have been successful when
to be used in marine environment. To simulate the design and manufacturing of a product
marine environment, the piles were subjected was performed by a single company or group
to wet and dry seven-day cycles in a 15% of companies, using the integrated design
sodium chloride solution, then tested to failure approach typical of the aerospace industry.
in bending. To facilitate the moisture ingress, The use of composites in the traditional civil
four out of eight glass composite reinforced engineering environment, where individuals
piles were pre-cracked at the mid-span prior create unique structures from standard com-
to testing. The bending strength of specimens ponents, has been slow, with the exception of
subjected to the wetting cycles, specially those markets where corrosion resistance offsets
pre-cracked, reduced significantly over time. higher costs and less than optimum perfor-
From the pre-stressing force applied, it is pre- mance. Large markets have not developed
dicted that 1.6% strain was induced in the rod, partly because of lack of design codes and
while larger values of strain may have been specifications and lack of awareness of com-
induced during the precracking of the beams. posites advantages by structural engineers. A
Additionally, radial stresses produced by the number of factors, including the decline of
twist of the rods in the strand and by the mois- military industry and the need for rehabilita-
ture ingress in the resin may have caused tion of USA infrastructure, have produced a
cracking. Although the neat resin failure strain flurry of activity in this area, which undoubt-
is reported at 4%, the strain applied to the edly will produce significant progress in the
composite rods may have produced cracking application of composites in construction.
of the matrix, thus facilitating the ingress of
moisture carrying the alkaline solution to the
glass fibers. Scanning electron microscopy REFERENCES
micro-photographs clearly show degradation Adams, R.c. and Bogner, B.R. 1993. Long-term use
of the glass fibers along with some cracking of of isopolyesters in corrosion resistance. Proc.
the matrix. An additional pile was fabricated 48th Annual ConI, Section I-C. New York:
with pultruded fiberglass rods for which an Composites Institute, The Society of the Plastics
additional coating resin was added to provide Industry Inc., pp. 1-5.
a smooth finish. This pile performed signifi- AGA NGV2. 1992. Proposed American National
cantly better than the other four after a nine Standard, basic requirements for compressed
natural gas vehicle (NGV) fuel containers. Draft
month exposure, perhaps because of the 8. American National Standards Institute.
added protection of the finishing resin. It is Ahmed, S.H. and Plecnik, J.M. 1989. Transfer of
concluded that the alkalinity of the environ- composite technology to design and construc-
ment (e.g. concrete), the availability of water tion of bridges. Washington, DC: US
or other solvent, and the protection provided Department of Transportation, Federal
by the resin must be evaluated for each spe- Highway Administration, Office of University
Research, Contract DTRS 5683-C00043.
cific application.
Bank, L.c. 1989. Flexural and shear modulii of full-
section fiber reinforced plastic (FRP) pultruded
46.4 CONCLUSIONS beams. J. Testing Evaluation, 17(1),40-45.
Barbero, E.J. and GangaRao, H.V.S. 1991. SAMPE J.
Composite materials have been used for many Part I (12) 1991. Part 11(1) 1992
structures that have performed well over Barbero, E.J. 1998. Introduction to Composite Materials
1002 Construction

Design. Washlngton, DC: Taylor and Francis. CFRP rods to a post-tensioned pre-stressed con-
Barbero E.J. and Raftoyiannis, I. 1993. Local buck- crete bridge, in Advanced Composite Materials in
ling of FRP beams and columns. ASCE J. Mater. Bridges and Structures, (eds. K.w. Neale and P.
Civil Engng, 5(3), 339-355. Labossiere). ~ontreal: The Canadian Society for
Barbero, E.]. and Tomblin, J. 1992. A phenomeno- Civil Engineering, pp. 405-414.
logical design equation for FRP columns with ~artine, E.A. 1993. Long-term tensile creep and
interaction between local and global buckling. stress rupture evaluation of unidirectional fiber-
Thln-Walled Structures. Composite Technol., 18, glass-reinforced composites. Proc. Composites
117-131. Institute 48th Ann. ConJ., pp. 9-A/l-4.
Composite Technology Inc. 1992. ~eisseler, H-J and Preis, L. 1989. High performance
Creative Pultrusions. 1989. Creative Pultrusions glass fiber composite bars as reinforcements, in
Design Guide. Alumn Bank, PA: Creative Concrete and Foundation Structures. Strabag Bau-
Pultrusions Inc. AG information brochure.
Davalos, J.E and Salim, H A. 1992. Design of stress- ~G. 1992. EXTREN Fiberglass Structural Shapes
laminated T-system timber bridges. Timber Design Manual. Bristol, VA: ~orrison ~olded
Bridge Information Resource Center, USDA Fiberglass Co.
Forest Service, N ortheastem Area. ~ochida, 5., Tanaka, T. and Yagi, K. 1992. The
~organtown, ~. Development And Application Of A Ground
Faza, S and GangaRao, HV.S. 1993. Pultruded fiber Anchor Using New ~aterials. Advanced
reinforced plastic bars, an alternative to steel Composite Materials in Bridges and Structures,
reinforcement of concrete, Section 13-D, Proc. (eds. K.W. Neale and P. Labossiere), ~ontreal:
48th Ann. Conf, pp. 1C-5C. New York: The Canadian Society For Civil Engineering,
Composites Institute, The Society of the Plastics pp. 393-402.
Industry, Inc. ~ufti, A.A., Erki, ~-A. and Jaeger, L.G. (eds). 1991.
Fibergrate. 1992. Structural Fiberglass Products and Advanced Composite Materials With Application To
Systems. Dallas, TX: Fibergrate. Bridges. ~ontreal: The Canadian Society for
GangaRao, H.V.s. and Barbero, E.J. 1991. Civil Engineering.
Construction: structural applications, in ~ufti, A.A., Erki, ~-A. and Jaeger, L.G. 1992.
Encyclopedia of Composites (6), (ed. S. Lee). New Advanced Composite Materials In Bridges And
York: VCH Publishers. Structures In Japan ~ontreal: The Canadian
Glaser, R.E., Moore, R. and Chiao, T.T. 1983. Life Society for Civil Engineering.
estimation of an E-glass epoxy composite under Munipalle, U.~. 1992. Analysis And Testing Of
sustained tensile loading. Composites Technology Wood-Glass Fiber Reinforced Plastic Adhesive
Review, 5(1),21-26. Interface. ~. S. Thesis. West Vrrginia University,
Head, P. 1988. Use of fiber reinforced plastics in ~organtown, ~.
bridge structures. Helsinki: :xm International Neale, K.W. and Labossiere, P. (eds). 1992. Advanced
Association for Bridge Structures Engineering Composite Materials In Bridges And Structures.
(IABSE), pp. 123-128. ~ontreal: The Canadian Society For Civil
·Iyer, S.L. 1993. First composite cable pre-stressed Engineering.
bridge in the USA. Proc. 38th Int. SAMPE Symp. Noritake, K., Kumagai,S., Wzutani, J. and ~ukae,
~y 10-13, pp. 1766-1771. I K. 1992. Construction of a pre-stressed barge

Iyer, S.L. (ed). 1991. Advanced Composite Materials in using aramid FRP rods, in Advanced Composite
Civil Engineering Structures, (ed. S.L. Iyer). New Materials in Bridges and Structures, (eds. K.W.
York: American Society of Civil Engineers. Neale and P. Labossiere), ~ontreal: The
Kakihara, R., Kamiyoshi, ~., Kumagai, S. and Canadian Society For Civil Engineering, pp.
Noritake, K. 1991. A new aramid rod for the 533-541.
reinforcement of pre-stressedconcrete struc- Pletcher, D. 1991. Consider structural composites,
tures, in Advanced Composite Materials in Civil Chem. Eng. Progress, November, 44-49.
Engineering Structures, (ed. S.L. Iyer). New York: Porter,~.L. and Bames, B.A. 1991. Tensile testing of
American Society of Civil Engineers, pp. glass fiber composite rod, in Advanced Composite
132-142. Materials In Civil Engineering Structures, (ed. S.L.
Koga, ~., Okano, ~., Kawamoto, Y., Sakai, H. and Iyer). New York: American Society Of Civil
Yagi, K. 1992. Application of a tendon made of Engineers, pp. 123-131.
 References 1003

Priestley, M.J.N., Seible, F. and Fyfe, E. 1992. Gere, J.M. and Timoshenko, S.P. 1990. Mechi1nics of
Column seismic retrofit using fiberglass-epoxy Materials, 3rd Edn. Boston: PWS-KENT
jackets, in Advanced Composite Materials in Publishing Co.
Bridges and Structures, (eds. K.w. Neale And P. Tomblin, J. 1991. A Universal Design Equation For
Labossiere). Montreal: The Canadian Society Pultruded Composite Columns. M.S. Thesis. West
For Civil Engineering, pp. 287-298. Virginia University, Morgantown, WV'
Raftoyiannis, L. 1993. Buckling Mode Interaction in Tomblin, J. 1994. Compressive Strength Models For
Fiber Reinforced Composite Structures. Ph.D. Pultruded Fiber Reinforced Composites. Ph.D.
Thesis. West Virginia UniverSity, Morgantown, Dissertation, West Virginia University,
WV. Morgantown, WV.
Rostasy, F.S., Hankers, C. and Ranisch, E-H. 1992. TUFSPAN. 1991. Technical Data and Design Guide,
Strengthening of R/C- and P /C-structures with Forth Worth: Tufspan, p. 8.
bonded FRP plates, in Advanced Composite Wolf, R. and Miesseler, H-J. 1992. Experience with
Materials in Bridges and Structures, (eds. K.W. glass fiber pre-stressing elements for concrete
Neale and P. Labossiere). Montreal: The bridges, in Advanced Composite Materials in
Canadian Society For Civil Engineering, pp. Bridges and Structures, (eds. K.w. Neale And P.
253-263. Labossiere). Montreal: The Canadian Society
Saadatmanesh and Ehsani. 1990. Fiber composite For Civil Engineering, pp. 425-433.
plates can strengthen beams. Concrete
International. Farmington Hills, Ml: American ACKNOWLEDGEMENTS
Concrete Institute, pp. 65-71.
Sen, R, Issa, M. and Mariscal, D. 1992. Feasibility of My sincere gratitude to Prof. G. Turvey for his help
Fiberglass Pretensioned Piles in a Marine researching applications in Europe, to all the con-
Environment. Report No. CEM/ST /92/1, tributors of information for this article, and to West
University of South Florida, Tampa, FL Virginia University for the support of this project.
AEROSPACE EQUIPMENT AND 47
INSTRUMENT STRUCTURE
Gary C. Krumweide and Eddy A. Derby

47.1 INTRODUCTION aluminum, or very expensive beryllium fur-

It is necessary to discuss GFRP (graphite fiber ther justified the material selection.
reinforced plastics) thoroughly in order to Consequently, numerous mirror bezels,
understand the importance of composites in telescopes, optical benches and reflectors were
the manufacture of aerospace equipment and designed and built from GFRP in the early
instrument structures. Before composite mate- 1970s. As a result of these efforts, today more
rials such as GFRP became viable candidate and more structures are being fabricated from
materials for aerospace primary structures, or GFRP materials, principally because specific
even for sporting equipment, they were first materials matching specific property require-
used for aerospace equipment and instrument ments are now available.
structures. Because composites exhibited both Although GFRP has dominated composite
unique and superior properties, designers materials applications, DuPont's Kevlar-49
were willing to pay the prevailing high prices has been found to be ideal for antenna reflec-
to achieve their design goals. For space hard- tors because of its extremely light weight and
ware, where a pound of weight saved was RF transparency. Many communication satel-
worth thousands of dollars, designers were lites utilize this type of Kevlar reflector, such
motivated to characterize composite materials as the SatCom-F, Telstar, ANIK-E, SpaceNET,
suitable to their applications. Understandably, and Superbird SCS (Fig. 47.1).
composites for primary and secondary struc-
ture (e.g. launch vehicles, aircraft frames, wing
spars and skins, etc.) were a 'hard sell', and
temperature extremes were too severe for
'thermosetting' GFRP to be used on missiles.
The quantity of GFRP required for a particular
piece of aerospace equipment or instrument
structure was usually minimal, so relative
costs were low, and composite materials' supe-
rior properties compared to heavy Invar, or
high coefficient of thermal expansion (CTE)

Handbook of Composites. Edited by S.T. Peters. Published

in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 Fig. 47.1 Superbird SCS Kevlar dual-shell reflector.
 Historical perspective and progress 1005

For aerospace equipment and instrument and/or are going to be used for aerospace
structures, generally, thermoplastic versions of equipment and instrument structures.
GFRP have not seen as much application as Table 47.4 compares the mechanical and
the thermosets due to high investment costs in thermal properties of candidate materials for
tooling and facilities. Considering the rela- aerospace equipment and instrument struc-
tively small quantities that are usually bought, tures. One could ask, from the obvious
thermoplastic applications are not often cost property advantage of beryllium, why all such
effective. Also, the required high temperature structures are not made from beryllium? If not
cures subject the laminate to microcracking beryllium, why not metal matrix composites
instabilities. or carbon-carbon (C/C) or silicon carbide
Metal matrix composite applications have (SiC)? The answer is that raw material cost,
been limited, and knowledge of material prop- fabrication cost, practical size, and other criti-
erties and processes has been restricted due to cal properties all come into play for any
their use on classified programs. Materials like specific application. Table 47.5 shows typical
silicon carbides and carbon-carbon find lim- design requirements for various applications
ited, but important, use in aerospace and indicates which materials typically satisfy
structures, particularly mirrors. Design selec- the critical requirement.
tion criteria plays an extremely important role The primary reason that GFRP is a material
in determining what type of material is used candidate for most applications is the wide
on any particular component of aerospace range of material systems that are currently
equipment and instrument structures. The pri- available that can compare with Kevlar, alu-
mary reasons for choosing a composite minum, Invar, beryllium, metal matrix
material, rather than a metal, are weight, composites, silicon carbide, and carbon-car-
dynamic stability and thermal stability. bon.
Table 47.1 categorizes some of the desirable
and undesirable characteristics that con-
47.2 HISTORICAL PERSPECTIVE AND
fronted the early users of GFRP. Table 47.2
PROGRESS
addresses the undesirable properties listed in
Table 47.1 and indicates how the aerospace From a historical perspective, and to illustrate
equipment and instrument designers have rec- where significant progress has been made in
ognized the barriers or problems with GFRP the use of composites for aerospace structure,
and found work-around techniques to allow the following areas of interest will be
their usage. Table 47.3 illustrates several struc- addressed:
tural applications where GFRP have been used

Table 47.1 GFRP properties

Desirable Undesirable
Low density Low short transverse properties
Coefficient of thermal expansion (CTE near Hygroscopic .
zero) High material cost
High specific strength High fabrication cost
High specific stiffness Low impact strength
Readily formable Subject to microcracks (translaminar stress
Crack growth resistant relief)
Adaptable laminate properties Low peel strength
Easily repairable
1006 Aerospace equipment and instrument structure

Table 47.2 GFRP design barriers and work-around techniques

Undesirable properties Barrier to design Work-around techniques Reference

Anisotropic Low strength in 1. Radius blocks Stumm, 1981

behavior angles 2. Substitute material Campbell, 1981
(metals) locally Krumweide, 1988
3. Use mortise and
tenon joints
Thermal instability 1. Use mortise and Stumm,1981
of angles tenon joints Campbell, 1981
2. Butt edges/insert Krumweide, 1988
3. Back-to-back splice
High through- 1. Egg crate joint Stumm,1981
thickness expansion 2. Mortise and tenon Krumweide, 1988
of laminate 3. Add local Krumweide, 1991
compensation
4. Maximize in-plane
material orientation
5. Fitting Boss through
skin
Hygroscopic nature Expansion/ distortion 1. Moisture barrier Hertz, 1977; Stumm, 1979
2. Define exposure/ Walrath, 1979; Levy, 1984
drying scenario Krumweide, 1989
3. Cyanate ester Telkamp, 1990
Krumweide, 1991; Brand, 1992
Material cost/ High basic part cost 1. Proper material Krumweide, 1977
fabrication cost selection Krumweide, 1988
2. Tooling/fabrication
techniques
3. Assess cost/weight
ratio
4. Oven cure
5. Minimize pieces
6. Eliminate molded
parts
7. Tailor joints/load
Susceptibility to Micro properties 1. Use low Stumm, 1981
microcracks affected temperature Krumweide, 1991
curing resin systems Brand, 1992
2. Low stacking angle
3. Small ply thickness
4. Reduce fiber modulus
5. Thermal cycle parts
6. Cyanate ester
Susceptibility to Low impact strength 1. Protect surface with Dunbar, 1978
impact damage Kevlar /honeycomb Herrick, 1984
2. Material design
(thickness, orientation)
Low peel strength Low joint allowables 1. Use mortise and Dunbar, 1978
tenon joints Krumweide, 1979
2. Fasteners and angle Stumm, 1981
clips
 Historical perspective and progress 1007

• problems encountered/work-around ites, principally GFRP, and indicates a variety

techniques; of work-around techniques which have been
• construction methods; implemented successfully to address these
• tooling development; design concerns. To help explain and clarify
• major milestones. these techniques, some examples of specific
design solutions are discussed below.
47.2.1 PROBLEMS ENCOUNTERED /WORK-
AROUND TECHNIQUES Moisture effects
Aerospace designers have constantly been For GFRP, it has been determined that the pre-
seeking the ideal material - the one material dominant mechanism of water penetration
which meets design requirements with the into the laminate is through the resin by a dif-
fewest compromises. Of course, few materials fusional process. The moisture diffusion is
are completely ideal for a given set of require- governed by Fick's second law which is simi-
ments and design trade-offs are almost always lar to Fourier's equation for thermal
necessary. conductivity:
Table 47.2, above, summarizes typical prob-
lems, presents some design concerns, and/or (47.1)
'barriers to design' associated with compos-

Table 47.3 GFRP Applications for aerospace structures

Equipment structures Instrument structures

RFSystems: Metering structures (LANDSAT, Thematic

-Reflectors (Viking, Nimbus-G, ACTS) mapper; SOHO, UVCS; HEAO-B, COSTAR and
-Feed horns and waveguides (INTELSAT) Hubble Telescopes)
-MUX cavities
Camera housings (Mars Observer Camera (MOC),
-Diplexers
-Phased arrays Hubble WFOV Camera)
Optical benches (MAGSAT, UARS HRDI, Hubble
Large segmented reflectors
FGS Optical Benches)
Solar panel (substrates) (BS-3, RADARSAT,
Mars Observer) FPA or relay optics (Hubble Telescope;
THEMATIC MAPPER)
Booms, Stardust (Shuttle RMAB, M-SAT)
Mirror bezel
Bus (Mighty Sat I&II, SMEX/WIRE, Jndostar Forte)
Mirrors (Microwave Limb Sounder, NGST, INM
Electronic chassis (E01, SMT5-0BC, MARS 98, Scan Mirror)
Stardust)
Lens holders
CCA cardguides (NAWC Coldplate)
Support benches (Hubble FGS)
Solar Concentrators
Startrackers (Hubble Equipment Shelf)
Submillimeter reflector
Helicopter mast mounts
Laser comm gimbals
1008 Aerospace equipment and instrument structure

where c = moisture concentration; t = time; Dz maximum moisture pick-up for the composite
= moisture diffusivity; and z = thickness coor- (MCm)may be calculated by the following equa-
dinate. tion:
Diffusivity is dependent on resin type and
MCm = (Mm)r x Wr (47.2)
temperature but is independent of moisture
concentration or laminate orientation. where Wr = resin weight % and (Mm)r is the
Figure 47.2 illustrates the importance of maximum moisture pick-up for the neat resin
temperature in determining the diffusivity of a for a given relative humidity condition.
laminate. Figure 47.3 illustrates the moisture The moisture content as a function, RH, can
absorption behavior as a function of tempera- be represented by:
ture for this same laminate. Note that this is
MCm = A (RH)B (47.3)
for a 100% relative humidity (RH) exposure.
Typical moisture absorption values (primarily where A and B are coefficients obtained from
by the resin) are 3.H% by weight for epoxy empirical data. Figure 47.4 depicts the per-
and 1-1.3% by weight for cyanate ester. The cent of moisture pick-up for various levels of

Table 47.4 Typical mechanical and thermal properties of isotropic/ quasi-isotropic materials

Material Modulus E Density CTE Cp K

(CPa) (Mg/m3) (XID-6/K) (kJ/kgIO (W/mK)
Copper 117 8.86 16.6 0.38 398
Beryllium 289 1.83 11.5 1.88 179-207
Aluminum 70.3 2.68 23.8 0.96 138-237
Titanium 113.7 4.43 9 0.54 16.9-20.8
Stainless steel 193 8.03 16.5 0.50 16.26
Super Invar 144.7 8.03 0.18 0.50 13.8
Fused silica 73 2.19 0.5 0.71 1.33
VIe fused silica 67.5 2.19 0.029 0.75 0.86
Zerodur 90.3 2.54 0.108 0.92 1.7
P75S/EP 103.3 1.72 --0.14 0.88 55.3
PlOOS/EP 143.3 1.80 --0.72 0.88 116
PI2OS/EP 170.9 1.83 --0.85 0.88 190
PI3OS/EP 186 1.83 -1.08 0.88 334
T300/EP 55.1 1.58 2.7 0.88 2.4
T50/EP 82.7 1.58 0.72 0.88 20.8
AS-4/EP 51.7 1.58 2.52 0.88 5.2
XN50A/EP 103.4 1.80 --0.072 0.88 48.4
XN70A/EP 144.7 1.83 --0.54 0.88 98.6
M55J/EP 103.4 1.66 --0.27 0.88 20.8
M60J/EP 120.6 1.69 --0.45 0.88 24.2
FT500/EP 82.7 1.77 0.36 0.88 41.5
FT/700/EP 137.8 1.80 --0.72 0.88 81.3
GY70/EP 110.2 1.66 --0.18 0.88 NA
Boron/EP 89.6 2.05 5.4 0.88 NA
S-GL/EP 26.9 2.02 9.7 0.71 0.34
KV-49/EP 29.6 1.38 6.3 NA 1.04
KV-149 40.6 1.38 1.8 NA 1.04
NA = Not Available
 Historical perspective and progress 1009

Table 47.5 Material selection criteria

Aerospace application
Critical Equipment structure Instrument structure
requirements
RF Solar panel Bus Booms Stable Mirrors
reflectors substrates structures structures
Mass GFRP (1) GFRP (1) GFRP GFRP GFRP(5) GFRP (6)
Kevlar Kevlar Aluminum Aluminum Beryllium Beryllium
M/M Beryllium Si/C,C/C
M/M
Dynamic GFRP (1) GFRP (1) GFRP GFRP GFRP(5) GFRP(6)
stability Kevlar Kevlar Aluminum Aluminum Beryllium Beryllium
Aluminum M/M Beryllium Invar Si/C,C/C
M/M Aluminum
Thermal GFRP GFRP (1) GFRP(5) GFRP GFRP (5) GFRP (6)
stability Kevlar Invar Beryllium Invar Invar
M/M Beryllium Beryllium
Si/C, C/C
Dynamic GFRP(l) GFRP (1) GFRP GFRP GFRP(5) GFRP(6)
loads (Gs) Kevlar Kevlar Aluminum Beryllium Beryllium Invar
Aluminum M/M M/M Invar Beryllium
Aluminum Aluminum
Temperature GFRP(2) GFRP (2) GFRP (2) GFRP(2) GFRP(2) GFRP (2)(6)
extremes Kevlar Kevlar Aluminum Beryllium Invar Si/C, C/C
Aluminum M/M M/M Beryllium Beryllium
Aluminum Aluminum
Hygrostability GFRP (2) GFRP(2) GFRP(2) GFRP (2) GFRP (2) GFRP (2)(6)
Aluminum Aluminum Beryllium Beryllium Invar
M/M M/M Invar Beryllium
Aluminum Aluminum
Si/C,C/C
High thermal GFRP(3) GFRP (3) GFRP (3) GFRP(3) GFRP(3) GFRP (3)(6)
conductivity Aluminum Aluminum Beryllium Beryllium Invar
M/M M/M Invar Beryllium
Aluminum Aluminum
Si/C, C/C)
Low thermal GFRP (4) GFRP(4) GFRP(4) GFRP (4) GFRP (4) GFRP (4)(6)
conductivity Kevlar
RF transmissibility Kevlar Kevlar
Cost GFRP GFRP (1) GFRP GFRP GFRP Aluminum
Kevlar Kevlar Aluminum Aluminum Invar Invar
Aluminum
(1) Combinations of Kevlar and GFRP is used when mass and dynamic stability is important.
(2) Cyanate resins have been shown to handle temperature extremes (T high, no microcracking).
(3) Pitch fibers (especially ultra, ultra high modulus have high thermal ~onductivity).
(4) PAN fibers have low thermal conductivity.
(5) Metals shown may be applicable if a small size structure.
(6) GFRP has been used successfully as mirror substrates and for some submillimeter reflectors core and skin.
1010 Aerospace equipment and instrument structure

1.DOOE,""
or
I:!.L/L == PI:!.M.
t._E~ --" Once P is known, the expected strain can be
calculated for a given percent moisture con-
tent in the laminate.
'" The significance of Fig. 47.4 cannot be over-
emphasized. Up until 1989 designers of aero-
" ..........
space structures had to utilize the
work-around techniques 1 and 2 mentioned in
1.OOIE-8 "-... Table 47.2 in order to handle the excessive
expansion and the corresponding distortion
associated with the hygroscopic nature of
D." 0.0D28 O.DOJ2 o.ooa 0._ epoxy based GFRP material systems. Table
TEMPERATURE C11 K •
47.7 indicates work-around techniques uti-
lized on various programs before cyanates
Fig. 47.2 Diffusivity against temperature for were introduced. Cyanate ester resin systems
P75S/cyanate; P75S/epoxy system. now being used (Brand et al., 1992) in GFRP
materials offer increased stability at reduced
cost for aerospace structures, primarily due to
relative humidity and shows the significance the possible elimination of expensive mois-
between an epoxy and cyanate ester resin. ture barriers (10% of manufacturing costs)
Table 47.6 provides values for the swelling and the even greater cost of facilities for dry-
coefficient (fJ) for typical GFRP materials out processing.
used to fabricate aerospace structures, where
Pis calculated from:
P == (~) (47.4)

....
0.40 J, -..... ,......,.,
.,. '.00 1!IIL.... AElIIN

iL
.... /; '/
tV
~
~
a.30 (......,.,
P7.'ERLt8G

II I--- ~

I!
/ V"- II! 0..
D.2D

L! r- ~
/ ~
1\ 0".
iIi !lAO
.#
~
l/
EM' ERA URE
~
I 17 r- r--... r--- ~.II I .... / ~V
vv
0.10

II 1/ V
-
............
t- 21 C
y I---
~

0." kV D.OO
D.oo 20.00 4G.ao 10.00
RElATIVE HUMIDITY (% •
8II.DD 100.00
0.00 2.00 4.00 e.oo 8.DO 10.00
TIME (HOURS'1I2)

Fig. 47.3 Percentage moisture change against time Fig. 47.4 Percentage moisture pick-up against rela-
(hI/2) for 0.5 mm thick P75S/cyanate at 100% RH tive humidity (%). (a) ERL1999 epoxy resin; (b)
exposure. P75S/ERL1962 epoxy; (c) P75/ERL1999 cyanate.
 Historical perspective and progress 1011

Table 47.6 Typical material CMEs for pseudoisotropic laminates

Material f3 (PPM/%M) MM@100%RH

T300/EP 400 1.13
T50/EP 215 1.13
P75S/EP 162 1.03
P100S/EP 103 1.03
P120S/EP 86 1.03
P75S/CY 105 0.32

Microcracking Figure 47.5 illustrates the effect of composite

microcracking. As the temperature is lowered
Microcracking of a composite which causes
and reaches the threshold of microcracking, as
dimensional instability should no longer be an
evidenced by an abrupt strain change, the
issue if the proper material selections are
curve changes to a new slope. Dimensional
made. A need to thermal cycle to gain dimen-
instability manifests itself in the following
sional stability (causing translaminar cracking
ways:
or microcracking to progress to completion)
was a requirement in the past, e.g. in Hubble • A hysteresis effect is produced in the struc-
Telescope tube components. This was due, pri- ture.
marily, to of th~ relative brittleness • Changes in CTE occur (CTE becomes more
(non-toughened) of the resins and the unavail- negative with increased thermal cycling).
ability of thin cure ply thickness prepregs. • Moisture response rates and moisture levels
Also, at the time, the phenomenon was not increase.
fully understood. • Irreversible expansion of the GFRP material.

Table 47.7 Prior work-around technique for addressing the hygroscopic nature of graphite/epoxy

Instrument Work-around technique Annotations (Ref)

(from Table 47.2 -
Hygroscopic nature)
Teal ruby 1 Stumm, 1979. Cryogenic application,
indium/bismuth deposited metal-type moisture
barrier - space application
Thematic Mapper 2 Walrath, 1979. Two flight units operational,
bake-out and environmental controls during
assembly and test very well controlled and
monitored - space application
Mars Observer camera 2 Telkamp, 1990. Pre-flight unit test provides
dimensional change data, compensated for by
preset of optics - space application
HRMA Cylinders HEAO-B 1 Hertz, 1977. Aluminum foil-type moisture
barrier; space application
1012 Aerospace equipment and instrument structure

.. Anisotropic behavior
'00 The anisotropic behavior of GFRP offers a
. challenge in the design of aerospace struc-
. " .. tures. Some typical areas of concern are joints,
. 17,1' 1.
springback, cutouts in cylinders and bowing
of panels and assemblies .
.u .10-11
L ..

.. /'- -:--7~---­
" ' ' -......... 1
Joints
Early in the 1970s, designers discovered that
FORIITACTIlATFII
~~-.---.---.----r---.---.---
... ...
OK
... ... the through-the-thickness (translaminar)
properties exhibited by GFRP reduced
strength and increased CTE values for joints,
Fig. 47.5 Interpretation of common events occur- i.e. the CTE in-plane maybe 0.lOx10-6rC and
ring during measurements illustrating 34x10-6rC through the thickness. In the fol-
microcracking of pseudoisotropic GY-70/930 lowing years, however, improved
(0° /45° /90° /135°) •. 12mm/ply. mechanical! thermal joint designs have
evolved which optimize joint characteristics
relative to thermal distortion/ cost/weight
requirements. Figure 47.6 illustrates this evo-
lution for both fixed and removable joints with
the development path moving from A to D.
Today the following procedures can be used
by designers to select appropriate fiber and
resin systems to eliminate or minimize micro-
cracking.
• Establish the thermal range of the exposure
environment.
• Determine the stiffness, CTE, and weight
requirements of the application.
• Use a thinner cure ply thickness (CPT)
prepreg, if high modulus materials are
required (e.g. 0.0063 mm or less CPT) or a
woven fabric. © 004 00lIIlE CLIP ® HllATISE , TEIIM

• Select a resin system compatible with the

thermal and mechanical requirement (e.g.
cyanate esters such as Fiberite 954-3 or YLA
RS-3 do not microcrack when used with
UHM pitch fibers (Amoco P75S) even at 0 IETII.f1.lII6E ® IWlIUS BlOCK
cryogenic temperatures).
• Establish the optimal fiber orientation (e.g.
(0° /45° /90° /135°).).
Of course, adequate testing of the material
laminates before they are employed in the © RECESSED INSERT - TYPE I ® RECESSED INSERT - TYPE 2
structure will ensure that correct material
selections have been made. Fig. 47.6 GFRP joint design evolution.
 Historical perspective and progress 1013

Springback Flat panel bowing/warping

Springback occurs when composite laminates When laying up a flat panel, the typical fiber
are formed into angles during processing. angle tolerance from the desired angle is 0±2°.
Springback, or angle closure, occurs to the This manufacturing tolerance is a major con-
extent of about 1S upon removal of the lami- tributor to the bowing or warping of a panel.
nate from the tool (General Dynamics The thicker the laminate, the more a problem
Corporation, 1985). Exposure to moisture con- these variations are, in that substantial forces
ditions will open this angle (via moisture are required to make the panel flat. The lami-
absorption) to return to, or even exceed, the nate can straighten upon moisture absorption,
original angle. Through-the-thickness expan- and drying the laminate to its original cured
sion of the laminate is the source of this state will bring back the prior laminate geom-
dimensional instability. Using flat laminates etry.
and butt bonding at right angles is the work- The best solution for eliminating bowed or
around technique. Very thin clips can be used warped panels is to use a 'rotate and fold'lam-
at the joint. inate technique, specifically developed, to
Thermally, structures can be very sensitive produce essentially balanced laminates. This
to this problem as was discovered in the late technique is depicted in Fig. 47.7. In essence,
1970s in the testing of the GEMS Forward ply lay-up tolerances are canceled by becom-
Mirror Support (Campbell et al., 1981). This ing symmetrical, with the resulting laminate
problem has often been overlooked by design- properties truly isotropic throughout the lami-
ers over the years and needs to be considered nate. Note that this process is only applicable
whenever composites are utilized on dimen- to pseudoisotropic laminate configurations.
sionally stable structures. Back-ta-back angles
or more simply mortise/tenon joints have Ed,. A

been the work-around technique.

.. Half A D .. Half 6 D

Cut-outs in curved surfaces (cylinders) ! Top

! !' Top

Some design approaches require holes in Ed,. C

cylinders. The degree of distortion when a
large hole is cut in a cylinder is surprisingly
extensive and the stresses generated can force
the cylinder to take a different shape (oval or
hour-glass). Even the ends of a cylinder that
has a large hole in it will not be flat or parallel
to each other. The thicker the cylinder wall, the
more difficult it is to correct this condition.
Stiffening rings usually correct the problem on
thin walled cylinders. Springback and cutouts,
in formed parts, may reduce the benefit gained
by a reduced number of piece parts if tight
dimensional control is necessary. Parts may
not fit together. Again, through-the-thickness
expansion of the laminate is the source of this
dimensional instability.
Fig. 47.7 The rotate and fold laminate construction
technique.
1014 Aerospace equipment and instrument structure

Assembly bowing
By maximizing the use of flat stock and mak-
ing dual skin, eggcrate core structures, very
stable assemblies are possible. This brings
about an additional concern with through-the-
thickness expansion in that when an egg-crate
core is used in a structure, severe bending is to
be expected unless a specific design technique
is employed. Since through-the-thickness
expansion is WOO-times the basic in-plane
expansion of the laminate, bending of the
structural assembly occurs. Figure 47.8 illus-
trates the correct and incorrect slotting
method. If done correctly the structure (Fig.
47.9) stays relatively straight through thermal
Fig. 47.9 FGS keel. The structure remains stable
cycling and moisture absorption. because the correct slotting method was employed.
Mortise and tenon joints, used instead of clips, also
promote stability.

INCORRECT SLOTTING METHOD

[ ]
47.2.2 CONSTRUCTION METHODS

~ ~ ~ ~ Some important design options for aerospace

WET OR T ) R.T. equipment and instrument structure are:
• truss (tubes) or cylinders (for telescopes);
~ ~ ~ ~ • monocoque skins or honeycomb sandwich
DRY DR T = R.T. (for reflectors);
• bonded, or bonded and bolted (for typical
joints);
CORRECT SLOTTING METHOD • molded unibody or flat laminate/bond
~ ~ assemblies (for typical structures).
~ ~
WET OR T ) R.T.
Truss compared with cylinder (telescopes)
~ ~ Material availability, size, weight, interface,
~ ~ loads and stiffness requirements or considera-
ORY OR T = R.T.
tions may drive the designer to a particular
configuration for a particular application.
Designers who favored tubular truss struc-
tures in the 1970s might not choose such an
approach today. Design techniques are differ-
ent, material choices are many, and
manufacturing methods have changed. Today,
J I-- CROSSING RIB THICKNESS for example, the Hubble Telescope could be
designed as a faceted, dual-shell structure
AOHESIVE+
with a ribbed core. The faceted dual-shell
Fig. 47.S Correct and incorrect slotting methods. (skins) could have water-jet cutouts with a
 Historical perspective and progress 1015

latticework appearance. Then, again, the tele- a molded part is excessive on a particular
scope might look the same but be fabricated design, a molded part may not be better. Then
from a non-microcracking, low moisture again, flat laminate construction may mean
absorbing, thin-prepreg, cyanate ester prepreg that more parts have to be made, inspected
system material. and handled, and if this cannot be done effi-
ciently by the manufacturer, then it may not
be the best approach. Some manufacturers
Monocoque skins compared with
can handle flat laminate construction very
honeycomb sandwich (reflectors)
efficiently, if they have experience of assess-
Large GFRP honeycomb reflectors were suc- ment of cost, weight, risk. Many flat laminate
cessfully fabricated in the mid-1980s for the 100% bonded assemblies for aerospace equip-
Advanced Communication Technology ment and instrument structure have been
Satellite (ACTS) spacecraft reflector. At the fabricated and successfully flown, both on
same time that large membrane reflectors aircraft and spacecraft. Apparently the previ-
were made for a Direct Broadcast Satellite ous high risk, thought to exist with this
(DBS) spacecraft reflector. They both met the approach, has not been proven true in prac-
desired design requirements. The tie-down tice. Here again, designer experience and
configurations and weight constraints, in part, preference will determine the course to fol-
dictated which approach was utilized. Here low.
again the requirements of size, weight, inter-
face, loads, stiffness and material availability
47.2.3 TOOLING DEVELOPMENT
will influence the design. Designer experience
and preference will also determine what con- The number of cures, the size of the part and
cept is used in the future. facilities available influence the choice of tool-
ing to such a degree that each product
fabricated must be thought-out thoroughly to
Bonded joints compared with bonded and
select the proper tooling.
bolted joints
An advantage of flat laminate construction
Experience with all bonded structures, cost is that it eliminates a need for fabricating pro-
considerations, and successful application (i.e. duction molds. A designer may want to look at
hardware on-orbit) are driving designers of this approach first in order to cut costs, if expe-
aerospace structures away from fasteners alto- rience and fabrication techniques support this
gether. The short cyclic load duration for method of fabrication.
spacecraft today are such that little, if any, ben- The high cost of production molds can be
efit is gained from bolts or fasteners and their significantly reduced through the use of alter-
use should be minimized. Aircraft applica- nate techniques. For example, for cylinders,
tions for equipment and instruments may be thin-walled rolled and welded aluminum
another story and each application needs to be molds create dramatic savings as their light
reviewed thoroughly, assessing cost, weight, weight permits envelope bagging which elim-
risk and so on. inates the need to withstand autoclave
pressures.
Monolithic (bulk graphite) molds are
Molded (unibody) compared with flat
expensive and heavy (storage and heat-up
laminate bonded/assemblies
rate concerns), but their CTE match to GFRP
The use of fewer parts may not be beneficial or Kevlar/epoxy is a great advantage.
if they do not match up well at assembly. Greater accuracy and better replication of
Also, if the cost of tooling and touch labor for parts is possible.
1016 Aerospace equipment and instrument structure

The tooling concept and materials used to ment and their application to aerospace equip-
fabricate the tooling can be as important to the ment and instrument structures (Tables 47.8
success of the hardware as the hardware and 47.9). The photograph of the Hubble Space
design itself. The aerospace structural designer Telescope (Fig. 47.10) only shows the metering
must stay in touch with technological develop- truss structure for the telescope; substantially
ments in tooling and tooling materials. more GRFP was used in the fabrication of the
Ford Plane Assembly (FPA) and the many
other components. The milestones cited repre-
47.2.4 MAJOR MILESTONES
sent major advances and point the way for
Milestones in the use of composite materials continued use of composites for a wide variety
have been in the areas of materials develop- of aerospace structures.

Table 47.8 Milestones for GFRP materials

Material Achievement
Thin prepregs Reduced weight of hardware components
Minimized microcracking
Supported mirror technology
Increased thermal stability of laminates
Toughened epoxies Increase impact resistance of laminate
Increase bond strength (interlaminear) of joints
Increase compressions strength of laminate
High modulus PAN fibers Reduce hardware weight
Allow high stiffness, thermal stable, cyanate laminate
Allow high strength application
Ultra-ultra high modulus pitch fibers Increase thermal conductivity
Minimum weight, stiff structures
Increased EMI capability
Cyanate ester resins Reduced moisture levels
Reduced strain response to moisture level
Reduced microcracking
Increased use temperature (high Tg)

Table 47.9 Milestones for GFRP applications

Application Date Achievement

Hubble telescope 1970-1980 Design and fabrication of a very large
(2.4 m diameter X 5.2 m length) and
thermally stable structure
Nimbus-G 1970-1980 No fasteners - membrane shell 100%
bonded microwave reflector; Qual
unit qual-tested twice. Flown in 1990
as TOPEX.
Shuttle RMAB 1970-1980 Reusable structure. Multiple shuttle
missions.
 Current applications 1017

Fig. 47.11 ACTS reflector.

Fig. 47.10 Hubble space telescope metering truss

assembly.
47.3.2 HRDI OPTICAL BENCH (UARS)
This 1.24 X 1.8 X 0.11 m thick, 49.9 kg optical
bench for NASA's Upper Atmosphere
Research Satellite (UARS) mission was nearly
47.3 CURRENT APPLICATIONS
100% bonded (Fig. 47.12). The only fasteners
were anti-peel fasteners at the four comers
47.3.1 ACTS REFLECTORS
where graphite/epoxy (T50/ERL1962) panel
The Advanced Communications Technology sides terminate at titanium fittings. The bench
Satellite (ACTS) reflectors were manufactured carried 3.2 times its own mass and had a CTE
by Composite Optics, Inc. (COl), for General of 0.36 X lO-6;oC (Dodson, 1989).
Electric/ Astro-Space Division under contract
with NASA Lewis Research Center (Fig. 47.11).
47.3.3 UVCS (SOHO)
These large reflectors (2.2 and 3.3 m aper-
ture) had measured surface RMS accuracies of The Ultraviolet eoronagraph Spectrometer
approximately 0.071 mm over their entire sur- (UVeS) for the European Space Agency's Solar
face after thermal cycling. These reflectors and Helospheric Observation (SOHO) space-
were launched into orbit in September 1993 craft (1995 launch) is 100% bonded
and are functioning well. Their GFRP skins graphite/epoxy (P75S/ERL 1962) (Fig. 47.13).
(P75S/ERL1962) and Kevlar honeycomb core This truss structure is the first of its kind in that
(Kevlar 49/934) resulted in a 3.44 kg/m2 the basic truss panels are bonded assemblies
weight for the completed reflector (Rule, 1989). that have an I beam cross section. Eighty
1018 Aerospace equipment and instrument structure

optical stability requirements (CTE -{l.09 X 1~ jOq

had to be satisfied along with strength and fre-
quency (70 Hz) requirements. The structure
will hold 90.7 kg of equipment and only weighs
21.8 kg itself (Kilpatrick et al., 1990). Most of the
surfaces on the structure were clearable and
inspectable (minimal closed areas).

47.3.4 COSTAR (HUBBLE TELESCOPE

CORRECTIVE OPTICS)
The Corrective Optics Space Telescope Axial
Replacement (COSTAR) was developed by
Ball Corporation by NASA/GSFC. This struc-
ture (Fig. 47.14) supports a Deployable Optical
Bench (DOB) that serves as the corrective optic
for the Hubble Space Telescope. Built by
Hercules Aerospace Company, this composite
Fig. 47.12 HRDI bench (UARS). structure has to have a low CTE «0.15) and

Fig. 47.13 UV Coronagraph Spectrometer (UVCS) telescope structure assembly (SOHO).

 Predictions 1019

Fig. 47.14 Corrective Optics Space Telescope Axial Replacement (COSTAR).

high resonant frequency (100 Hz for the DOB). • new materials;

The long-term stability effects due to moisture • new manufacturing processes;
absorption (ground) and desorption (space) • new applications;
were addressed by optimizing the composite • economics.
laminate orientation. The long-term CTE con-
This, of course, does not mean other composite
cerns were more important in meeting
materials will not be used, but their use will be
line-of-sight requirements (>15 arc-sec) than
in similar proportion as is the current practice.
was a need to optimize CTE. A removable
cover and mounting of aluminum electronic
box required very unique design solutions 47.4.1 NEW MATERIALS
(Neam and Gerber, 1992).
GFRP is a constant evolution of material varia-
tions. Fiber makers continue to change their
47.4 PREDICTIONS fiber manufacturing processes to create a new
fiber to satisfy the needs of the user. Properties
The dominance of GFRP as the preferred
like fiber strength, modulus, and thermal! elec-
choice composite material in the fabrication of
trical conductivity are continuing to be
aerospace equipment and instrument struc-
adjusted for various application requirements.
tures will continue into the next century
The demand for properties changes also are
because of a number of reasons:
requested of the resin manufacturers and! or
1020 Aerospace equipment and instrument structure

prepreg suppliers. Cyanate ester resin formula- 47.4.2 ECONOMICS

tions are in their infancy, current GFRP
A 1993 conference held in Logan, UT, spon-
cyanates have levels of moisture absorption
sored by the AIAA and Utah State University,
less than one-third the GFRP / epoxies. Soon
had a basic theme for small satellite structures
moisture absorption for GFRP / cyanate may be
of 'cheaper, faster, better'. The 'better' was in
one-tenth that of GFRP epoxies. In the future,
some cases interpreted as lighter for better
we may have thermoset resins that will have
performance. .
no moisture absorption and their operating
One of the primary messages to the mdus-
temperatures may be over 240°C (464°F).
try from that conference was that econoInics
dictates that satellites become smaller and
New manufacturing processes lighter so that the 'on-orbit' cost for a particu-
lar satellite is much less, can be launched by a
It is hard to imagine new manufacturing meth-
small, low-cost launch vehicle (e.g. asc
ods for processing GFRP that are not in use
Pegasus or Lockheed LLV1). GFRP is currently
today, but the demand for less expensive
the most logical choice to satisfy the require-
materials, and hardware may bring develop-
ment of cheaper, faster, and better.
ments. One method currently being developed
Manufacturing processes are developed that
by Composite Optics, Inc. in San Diego, CA,
will allow lightweight GFRP to be utilized for
called SNAPSAJ'I'M is predicted to reduce the
bus structures, solar panel substrates, reflectors
manufacturing cost to less than half current
and instruments. Weight reductions of 50-60%
levels.
are possible with GFRP. Package size may be
the only liInitation to allowing three or four
New applications satellites to be launched rather than one or two.
Creative packaging and miniaturization will
New applications stimulate the need for new
solve this problem and launch booster cost will
materials, and new materials, in turn, create
then be minimal (25-35% of the current cost
new applications. Future applications for level of a single satellite).
GFRP are mirrors of all sizes, surface figure
This same idea (cheaper, faster, better) can
and accuracies in the IR and visible range,
also apply to aircraft equipment and instru-
large antenna reflectors (9.1-27.4 m diameter),
ments, namely, airborne avionics, tracking
electronic chassis and cardguides, plated RF
and targeting instruments, and phased
components (MUX Cavities, Dirle~ers), arrays. Today, a military or commercial air-
phased arrays, to name a few. The list IS on-
craft maker may not be able to sell a new
going but it is hard to tell whether new GFRP
aircraft but he could redesign the avionics
materials bring about these applications or the
(lightweight) to improve the performance of
applications bring about the material. For
that aircraft (including helicopters).
instance, Amoco's KllOOX fiber is being exper-
imented with for electronic chassis, but the
original demand was for high thermal conduc- 47.5 CONCLUSIONS
tive material for heat sinks (thermal straps). In
The technology of composite materials, espe-
another example, thin prepregs were devel-
cially GFRP, is constantly evolving to allow
oped to make lighter skins for honeycomb
increased aerospace application. This
panel structures, but in the future Inirrors for
increased use of GFRP is for both current type
visible range optics will also mostly use these
applications as well as genuinely new applica-
thinner prepregs.
tions where GFRP was not previously used.
The future for composites, in general, looks
 References 1021

promising because of the need for aerospace Krumweide, G.C, Brand, RA. 1991. Attacking
equipment and instrument structures to be Dimensional Instability Problems in
lighter which increases performance and saves Graphite/Epoxy Structures. Composite Design,
Manufacture and Application. ICCM/B.
costs. Manufacturing methods are being
Honolulu. July 15-19.
developed that allow GFRP hardware, in par- Krumweide, G.C 1977. Development of a graphite/
ticular, to be processed less expensively with Epoxy Reflector: A design-to-cost project.
reduced cycle times. It is not hard to imagine SAMPE Quarterly. 8(3).
GFRP material replacing most metallic mater- Krumweide, G.C, Chamberlin, D.N., Rule, J.E.
ial applications for aerospace equipment and 1988. Adaptation and innovation in high-modu-
instrument structure once' old paradigms' are lus graphite/epoxy composite design: Notes on
Recent Developments. SPIE O-E LASIE 1988.
given up and awareness of new GFRP mater-
Los Angeles.
ial technology is the norm. Krumweide, Derby, E.A., Chamberlin, D.N. 1989.
The performance of effective moisture barriers
REFERENCES
for graphite/epoxy instrument structures.
SAMPE. Atlantic City, NJ.
Brand, RA., Derby, E.A., Chamberlin, D.N. 1992. Krumweide, G.C, Hoste, J.H. and Staats, J.R 1979.
Evaluation of high-modulus pitch/cyanate Structural Development of the Thematic
material systems for dimensionally stable struc- Mapper Optical Metering Structure. The
tures. Design of Optical Instruments, Orlando, FL: Enigma of the Eighties: Environment,
SPIE. Economics, Energy. SAMPE J. 24(2), 1343--1355.
Campbell, M.P., Hoste, J.H., Kedward, K.T., Levy, D.J. and Arnold, CR 1984. Metal Moisture
Krumweide, G.C 1981. Designing Composite Barriers for Composites. 29th Nat. SAMPE
Structures for Thermally Stable Applications. Symp. April 3-5.
Fifth DOD/NASA. Con! Fibrous Composites in Rule, J.E. 1989. Thermal Stability and Surface
Structural Design. New Orleans, LA. January Accuracy Considerations for Space-based
1981. Single-and-Dual Shell Antenna Reflectors. ESA
Dodson, D.J., Rule, J.E. 1989. Thermal Stability ESTEC. Noordwijk, Netherlands.
Considerations for Space Flight Optic Benches. Stumm, I.E., Pynchon, G.E. Krumweide, G.C 1981.
Tomorrow's Materials Today. Vol. 34. Graphite/Epoxy Materials Characteristics and
Dunbar, D.R, Robertson, A.R, Kenison, R. 1978. Design Techniques for Airborne Instrument
Graphite/Epoxy Booms for the Space Shuttle Applications. 309. Airborne Reconnaissance. V,
Remote Manipulation. ICCM II, Toronto, Canada. SPIE.
General Dynamics Corporation (GDC Manual). Stumm, J.E., Pynchon, G.E., Pepi, J.W. and Bovenzi,
1985. Design for Cost and Quality Manual. EG. 1979. Low Temperature/High Stability
Herrick, J.W. Multi-Directional Advance Applications of Composites. The Teal Ruby
Composites for Improved Damage Tolerance. Experiment. Conf. Advanced Composites. EI
Composites in Manufacturing 3. Anaheim. Segundo, CA.
January 10--12, 1984. Telkamp, A.R., and Derby, E.A. 1990. Design
Hertz, J. Moisture Effects on Spacecraft Structures. Considerations for Composite Materials used in
1977. The Enigma of the Eighties Environment: the Mars Observer Camera. Advances in Optical
Environment, Economics, Energy. SAMPE, 24(2). Structure Systems. 1303. Orlando, FL: SPIE.
Kilpatrick, M.C, Girard, J.D., Dodson, K.J. 1990. Walrath, D.E. and Adams, D.E 1979. Moisture
Design of a Precise and Stable Composite Absorption Analysis of the Thematic Mapper
Telescope Structure for the Ultraviolet Graphite/Epoxy Composite Structure. Modem
Coronagraph Spectrometer (UVCS). Advances in Developments in Composite Materials and
Optical Structure Systems. Vol. 1303. Orlando, FL: Structure. ASME Winter Meeting.
SPIE.
AIRCRAFT APPLICATIONS 48
Richard N. Hadcock

48.1 INTRODUCTION metal aircraft structures. However, there are

some basic differences between the generic
Advanced composites, composed of high-
civil and military requirements and further
strength, high modulus, low density
variations associated with the intrinsic differ-
continuous fibers embedded in polymer
ences between the structural behavior of
matrices, first became available some 30 years
metals and composites.
ago. Since then, composite aircraft structures
The US Federal Aviation Regulations (FAR)
have transitioned from laboratory curiosities
for civil aircraft and the US Air Force (USAF)
into low-risk, light-weight alternatives to
and US Navy (USN) military aircraft require-
metal structures. Thousands of safety-of-flight
ments have differences. The European Joint
composite components are flying in regular
Airworthiness Requirements GAR) are similar
service on military and civil aircraft.
to the FAR and have a similar numbering sys-
Major advantages of high-performance
tem, but, again, there are some differences.
composite structures include weight savings,
FAR sections related to design and certifica-
material tailorability, improved fatigue and
tion of aircraft, aero engines and propellers
corrosion resistance. Disadvantages are pri-
contain more than 600 pages1. The equivalent
marily cost related. Almost all the composite
USAF documents contain almost 400 pages
structures currently in production and service
and distribution is restricted.
have thermoset matrices. A few aircraft parts
Although the requirements are applicable
are currently being made from thermoplastic
to both metal and composite structures, certifi-
matrix composites. Metal matrix composites
cation of composite structures is very much
are still in the development stage.
more extensive and requires many more tests.
'Conventional' aircraft structural materials
The contractor must fabricate and test thou-
now include polymer matrix composites in
sands of specimens and hundreds of
addition to aluminum, titanium and steel. The
subcomponents to qualify a single new car-
chronology of utilization of different rein-
bon/ epoxy system and associated structural
forced plastics is shown in Fig. 48.1.
details.
As an example, the material qualification
48.2 DESIGN AND CERTIFICATION f1rogram for the all-composite Beech Starship
REQUIREMENTS mcluded collection and analysis of thousands
of data points from element, panel and sub-
The general structural design and certification
component tests. Finally, fatigue and residual
requirements apply to both composite and
strength tests were made of a complete air-
frame 2,3. A similar test program was used by
CASA to certify the Airbus A320
Handbook of Composites. Edited by S.T. Peters. Published
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 carboni epoxy horizontal stabilizerA.
 Advanced composite materials 1023

Cottonl x· - - -x
BakolKo
Propellers r Co.posite Material x· -R&D--x

Flax/Phenollc x- - -x Wing Spar, Fuselage Production

P.per/Phenollc x - x Drop Tanke - -

Glu./Poly"t.r x- - -x Fuselage-
Radom•• , Drop Tanka

GI..a/Phonolic x·· x Flrowalls - - -

Asbestos/Phenolic x-·· -x Wings

GI ..../Epoxy x···x Radom". fin •• Ruddo .... Rotor Blado. - - - - - - - - - - - -

fairings. Fu•• lag ... Wing •• Etc.

GI ..s/Polylmldo x···x Radomo. ------------------1

CArbon/Epoxy x- - - - - - - - -x Stabilizers, Wingl, Fu ••lag •• , Control Surface. - - -
Fairing., Floor., Propell.rs, Etc.
Boron/Epoxy x- - - -x Rudde,.., Stabilize,.. -

Kevlar®/Epoxy x- - -x FaIring., Radom •• ,

Loading & Trailing Edge.

Carbon/Polylmlde x- - - - -x Hot Structur•• , Englnelnlet.,-------

Englno Flap •• Wing Skin.

Carbon/Thormopl.sllc x·· ··Spoilers·········x Wing Skins, - - - - I

Stablllz..... 000 ...

Carbon/BI.malolmldo x···x Flaps, Strako., Englno C.. lngs •. - - - - - !

Wing & Fus.lago Skins. Thru.t Rovo ... o",

t910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Fig. 48.1 Aircraft structural reinforced plastics. (© 1994 RNH Associates.)

Certification of military aircraft composite 48.3 ADVANCED COMPOSITE MATERIALS

structures also requires a major test program.
Boron fiber reinforced plastics (BFRP) and car-
The 'Building Block' approach, which starts bon fiber reinforced plastics (CFRP) advanced
with static, fatigue and damage tolerance composite materials were first used for aircraft
tests of elements, followed by tests of sub-
structures in 1966. Epoxy matrices were used
components and full-scale components, has
for these materials, which were produced as
been very successful in identifying and solv-
prepregs in the form of unidirectional prepreg
ing design and manufacturing proble~s tape.
before the design is finalized and frozen. LlVe Use of BFRP has been limited by the very
fire tests are also required for US military air-
high material price (more than $900/kg in
craft certification. The 'Building Block'
1993).
approach was for certification of the
Since the early 1970s, CFRP prepregs have
Grumman/US Air Force B-1A composite hor-
become available in the form of unidirectional
izontal stabilizer and for the McDonnell
tape, woven broadgoods and other forms with
Douglas/US Navy F-18 and AV-8B composite
many different types of carbon fibers and with
wings S,6.
many different epoxy, polyimide and bis-
maleimide (BM!) thermoset matrices. CFRPs
have also been available with PEEK and other
thermoplastic (TP) matrices.
1024 Aircraft applications

Almost all CFRP aircraft structural compo- Because the price of aviation fuel has
nents have been made with thermoset dropped dramatically since 1978 relative to
matrices (1993 prices ranged from $60/kg for material and labor costs, aircraft prices have
epoxy matrix unidirectional tape to $250/kg now become far more important a design
for BMI matrix materials). Thermoplastic requirement than weight savings. As such, it
matrix applications have been limited by the should now be assumed that the customer will
high price (about $250/kg) and low compres- pay very little premium for weight savings
sion strength of these materials. and the prices of composite components
Glass fiber reinforced plastics (GFRP) con- should be comparable to their metal counter-
sisting of E-glass, and the higher strength parts. Early in the design phase, weight
S-glass fibers in epoxy, polyimide and pheno- savings can be transformed into cost savings
lic matrices are being used for helicopter rotor by reconfiguring the airplane, but this is not
blades, for many secondary structures such as possible once the overall design is frozen.
radomes and fairings. Glass/phenolic ther- Weight savings are most valuable early in the
moset and glass/polyethersulfone and other design phase, but the value of weight savings
thermoplastic matrix composites are used for diminishes as the design becomes finalized.
fire-resistant cabin and freight hold panels and The design of the structure is primarily
liners because of their low heat and smoke dictated by production costs and by main-
release properties. Prices of glass/epoxy tainability and repair considerations.
prepreg range from $5/kg to $lO/kg. Composite material prices are very high and
DuPont introduced Kevla~ 49 aramid fiber costs of design, certification testing, tooling,
in 1971. Aramid fiber reinforced plastics inspection, material storage, waste material
(AFRP) have low density, high tension disposal and repair are all higher than for alu-
strength and excellent impact damage resis- minum structures. However, these higher
tance, but low compression strength. AFRP costs can be offset by designing large integral
materials are being used for radomes, leading composite components which reduce part
edges, fairings, floors and other secondary count, the number of joints and assembly
structure applications. AFRP has also been costs.
used with some local CFRP reinforcement for Most aerospace companies are using the
the complete airframe of the Avtek 400A busi- 'Concurrent Engineering' approach, where
ness aircraft. engineering, manufacturing, quality control,
Most of the composite primary and safety- logistics support and cost estimating person-
of-flight structure components (wings, nel are formed into collocated teams. The more
fuselages, empennage and control surfaces) important interfaces are listed in Fig. 48.2.
currently in service are made from The costs and prices of aircraft components
carboni epoxy prepreg unidirectional tape or vary considerably with the type of aircraft and
woven broadgoods. Some light aircraft are the type, construction, size and materials used
made almost entirely from GFRP woven for the component.
broadgoods. Current (1992-4) civil aircraft prices vary
from $70/kg ($30/lb) for a small private air-
craft to between $800/kg and $1000/kg for
48.4 THE DESIGN PROCESS
airliners and business aircraft. Military aircraft
Composites structure design involves many costs are very much higher and are very
different disciplines and is far more complex dependent on the total numbers of aircraft pro-
than metal structures design because of the duced. These range from a cumulative average
anisotropic strength and behavior of the com- cost of $2100/kg for 3000 General Dynamic
posite materiaF. F-16 fighters to more than $12 ODD/kg for 20
 The design process 1025

Requirements
External geometry
Interface geometry
Extemalloads
Structural requirements
Interface requirements
Operational environment
Weight & cost targets

I
Design engineering
Internal loads (from stress)
Materials selection
(with M&P and stress)
Cost/weight tradeoffs
Detail design
Joints and attachments
Detail drawings
(Paper / computer)

I
Engineering Production Other
Loads & dynamics Materials & processes Quality control
Loads Material specifications Material control
Stiffness requirements Process specifications NDI equipment
Flutter analysis, etc. M&P standards NDI/DI requirements
Accept/reject criteria
Stress Manufacturing (with stress and M&P)
Finite element models Manufacturing engineering Dimensional inspection
Design allowables Manufacturing methods MRR action/ damage repair
(withM&P) Shop instructions (with stress, design and M&P)
Structuraloptimuation Subassembly
(with weights) Assembly Cost estimating/control
Stress analysis Manufacturing
Facilities/equipment Materials (including costs of
Test requirements
Stress reports Autoclaves/presses/ ovens waste and waste disposal)
Refrigerators Engineering
Weights ATMs/TPMs Quality control
Weight targets NC ply cutters MRRactions
Weight control RTM/RIM equipment IE, facilities and equipment
Machining/ drilling/ trimming Rates and overheads
Structural test
Element tests Tooling Product support
Subcomponent tests Tool design Maintenance & repair manuals
Static and fatigue tests Moldform tooling Inspection requirements
Environmental tests Bonding tooling Repair materials and processes
Flight test support Assembly tooling Special repairs

Fig. 48.2 Design/manufacturing interfaces. (© 1994 RNH Associates.)

1026 Aircraft applications

..
MIUTARY. F/NB{T

100 ~==
•
MIUTARY TRANSPORTS
en
z
T-45 A320 0.1012 $l00/kg •
AIRUNERS
o
::J
.J 10
//
~Lll.228
,t,I1 ~/~ /'"
•
BUSINESSNC

•
ATR72
~
o
I

PRIVATENC
uj
()
a:a.. ~BARON
0.1 ;ONANZA

0.01 +---+--+-H-++H+-+-+-+-+++I-H---+--+-++I+++I
1000 10000 100000 1000000
OPERATING EMPTY WEIGHT, kg
a

100
AH-64
ICff.531
$3OOOIkgl
$2OOOIKgl
$1500/I<g 1 ..
CIVIL
RAH-66
•
I./IIIVI/
1//171== $1 ooo/l<gl
en II MIUTARY
z
o 10
TIGER
~ [L$75O/I<g 1

3 8-76
UH-60
~I BK 117 r.i V' Bel1214
o
"
8911212
uj
() 1 Bell 222
a:a.. .......
Dauphin
80105
Bell 206
LMcDD500
0.1 I11I I
100 1000 10000 100000
b OPERATING EMPTY WEIGHT, kg

Fig. 48.3 Aircraft costs and prices. (a) Fixed wing aircraft; (b) helicopters. (©1994 RNH Associates.)

Northrop B-2A bombers. Some prices of civil range from about $70/kg to $130/kg, prohibit-
aircraft and costs of military aircraft are shown ing much use of CFRP materials.
in Fig. 48.3. Airframe prices are 50-60% of the Prices are based on information from
aircraft prices, or from $400/kg to $600/kg for References 8-13 and have been adjusted to
airliners and business aircraft, depending on constant 1992 US dollars utilizing Aircraft
the type and size of the airplane. Private and Price Deflators8.
competition aircraft airframe fly-away prices
 Structural applications 1027

48.5 STRUCTURAL APPLICATIONS grown during the past 30 years. Published

weight fractions are based on either airframe
During the past 30 years, structural applica- weight, structure weight (airframe plus landing
tions of composite materials have grown from gear), or operating empty weight; and in some
a few radomes and fairings to the entire air- cases the base is not defined at all. Because of
frame. Many hundreds of CFRP wings, these differences, the structural composite
stabilizers and flight control surfaces are cur- material weight fractions shown in Fig. 48.4
rently flying in regular service. have been normalized in terms of per cent
The composites weight fraction has also structure weight (airframe plus landing gear).

*'W
II:

G
::J

~
~
rJJ
o
a...
oo~ 1

1980 1 1 2000
YEAR (FIRST FUGHT)

1975 1 1 1
YEAR (FIRST FUGHT)

Fig. 48.4 Composite material weights. (a) Tactical aircraft; (b) civil transports. (© 1994 RNH Associates.)
1028 Aircraft applications

The composites weight fraction for tactical upper branch includes most of the Western
aircraft, Fig. 48.4(a), appears to have leveled European airliners and the Boeing 777; the
between 22% and 26%. The Lockheed F-22A lower branch includes the McDonnell MD-ll,
airframe is composed of 26% composites, 30% Ilushin Il-96 and the Boeing 737X. The latter
titanium and 14% aluminum; the McDonnell aircraft is reported to have less composites
Douglas F / A-18E structural weight is 22% than the Boeing 737-300 because many of the
composites, 15% titanium, 29% aluminum and operators will be the smaller airlines which do
14% steeP2. not have composite maintenance and repair
Composite weight fractions for other pro- facilities 12.
duction and technology demonstrator military Weight savings provided by composites
aircraft, such as the Grumman A-6E (which vary considerably with the type of aircraft and
has a CFRP wing designed and manufactured component. Weight savings, in terms of com-
by Boeing), Northrop B-2A bomber and YF- posite weight fraction, are shown in Fig. 48.5.
23A fighter, General Dynamics F-16XL, These tend to decrease as the overall compos-
Grumman X-29A, Rockwell/MBB X-31A, IAI ite weight fraction increases. These data were
Lavi, the British Aerospace EAP, the obtained from different sources, primarily
Eurofighter 2000 and the Mitsubishi FS-X, are References 7, 9 and 12-18. It should be noted
not included because of space limitations or that the weight savings for the AV-8B are
availability of weight data. based on a weight estimate for comparable
The composites weight fraction for civil aluminum wing, and not the AV-8A aluminum
transports, Fig. 48.4(b), has two branches: the wing, which is smaller15.

t-
f·14AI
J:
(!}
[j
~
45
...
TACTICAL AIRCRAFT
en 40
~ •
en 35 TRANSPORT AIRCRAFT
0
a. 30
:!:
0 25
()
#. 20
~ 15+-r:::-::::1---11----lA
~ 10+-------+-------+-------+-----~
~
en 5+--------r------~--------+-------~
t-
:c
(!} 00 10 20 30 40
~ COMPOSITES, % STRUCTURE WEIGHT

Fig. 48.5 Composite weight savings. (© 1994 RNH Associates.)

 Structural components 1029

48.6 STRUCTURAL COMPONENTS combat and technology demonstrator aircraft

wings have composite skins which are bolted
Many composite components are currently
being produced in the US and overseas for to metal substructure (e.g. the McDonnell
military and civil aircraft. Most of the major Douglas F/ A-18, Grumman X-29, Rockwell
X-31, General Dynamics F-16XL, lAl Lavi),
commercial aircraft primes subcontract out the
other wings have composite skins which are
production of more composite components
bolted to composite substructure (e.g.
than they build in-house.
McDonnell Douglas AV-8B, Bell/Boeing V-22,
Composite structural and semi-structural
SAAB Gripen, Grumman/Boeing A-6E,
components which are currently being pro-
Dassault Rafale, British Aerospace EAP,
duced are listed in Table 48.1. Some specific Eurofighter 2000)6,9,12,14,15,17.
aircraft components are described later in this
At least three different forms of construc-
section.
tion have been used for stabilizers: composite
Different forms of construction have been
skins bonded to metal/honeycomb substruc-
used for many of these components. Some
ture (Grumman F-14, McDonnell Douglas F-15
Table 48.1 Aircraft composite structural components

Wings: Fuselages: Propulsion systems:

Box beam skins Radomes Engine fan blades
Box beam substructure Forward fuselage Engine casings
Wmglets Canopy frames (helicopters) Nozzle flaps
Leading edge flaps/slats Mid fuselage Thrust reversers
Ailerons /Flaperons Rear fuselage Engine nacelles and cowlings
Flaps Speedbrakes Fan cowls
Flap vanes Tailcone Engine doors
Spoilers Floor beams Fences
Fixed leading edges Floors Transcowls
Fixed trailing edge panels Rotordomes Firewalls
Flap track fairings Crew compartment armor Turbine blade containment
Actuator fairings Rotary launchers rings
Cabin doors (helicopters) Drive shafts
Empennage: Cabin and freight hold Pylon fairings
Horizontal stabilizers/ canards: linings and partitions Fuel tanks
Skins Overhead baggage Drop tanks
Substructure compartments Propeller blades
Elevators Air ducts
Flight refuelling probes Miscellaneous:
Leading edges
Fixed trailing edge panels Crew seats
HelicopterlRotorcraft:
Tips Toilet waste tanks
Main rotor blades Filament-wound air bottles
Vertical stabilizers: Tail rotor blades Cargo containers
Skins Rotor drive shafts Cabin door balance springs
Substructure Tooling
Doors and fairings: -moldforms
Rudders
Leading edges Landing gear doors - assembly/bonding fixtures
Fixed trailing edge panels Landing gear fairings - checking fixtures
Ventral fins Landing gear pods Honeycomb core (GFRP and
Tips Wing-fuselage fairings CFRP)
Antenna housings Stabilizer fairings Damage repair (metal and
Equipment access doors composites)
1030 Aircraft applications

and F-1S, Mikoyan Mig-29), composite skins stabilizer of the Grumman F-l4A, Fig. 4S.6.
bolted to metal substructure (General The Navy requirement that the stabilizer
Dynamics F-16), and composite skins bolted to should be fully qualified by full-scale static
composite substructure (McDonnell Douglas and fatigue tests one year before the first flight
AV-SB, Bell/Boeing V_22)9,lS,17,20,21. was satisfied by successful tests in 1969. More
Elevators, rudders, ailerons and other con- than 1500 stabilizers had been built when F-14
trol surfaces have generally been made by production ended in 199117.
bonding composite skins to substructure com-
posed of metal spars and full-depth
honeycomb core. Later designs are replacing
the honeycomb core with composite ribs and
spars.
All composite sheet-stiffened construction
is used for the forward fuselage of the
McDonnell Douglas AV-SB and the complete
fuselage of the Bell/Boeing V_229,14,16.
There are also many variations in the con-
struction of composite components used for
civil aircraft and helicopters.

48.6.1 MILITARY AIRCRAFf COMPONENTS

Development of advanced composites tech-

nology was initiated in 1964 by General
Bernard Schriever of the US Air Force who
tasked the Air Force Materials Laboratory
(AFML) at Wright Patterson Air Force Base,
Ohio, to contract some US aircraft companies
to design, fabricate and test various structural
components made from boron/ epoxy.
McDonnell made 45 rudders which were
installed on F-4 fighters for service evaluation; Fig. 48.6 (a) Grumman F-14; (b) F-14 boron/epoxy
General Dynamics was awarded a major con- horizontal stabiliser. The stripes on the lower cover
tract to develop a boron/epoxy horizontal are aluminium foil for lightning protection.
stabilizer for the F-ll1 in 1966; and Grumman (Courtesy of Grumman Corporation.)
won a contract to design and fabricate a wing
box beam for a conceptual advanced fighter in
1965. The US Navy awarded a contract to Boron/ epoxy covers were utilized for the
Douglas to develop a carbon/epoxy horizon- McDonnell F-15 horizontal and vertical stabi-
tal stabilizer for the A-4 in 1971. Douglas had lizers and rudders. The large speed brake,
previously made and flown a carbon/epoxy which is located on the fuselage behind the
wing flap on an A-4. By May 1971, more than canopy, is carbon/epoxy-aluminum honey-
SO different advanced composite aircraft com- comb sandwich. The F-15 was still being
ponent programs had been initiated22,23. produced in 1994, 20 years after the first flight.
The first production application of The principal production use of
advanced composites for a safety-of-flight boron/epoxy was for the F-14 and F-15 stabi-
component was the boron/epoxy horizontal lizer covers. Boron/epoxy was selected
 Structural components 1031

because a comprehensive design database was and access panels. The aircraft is currently
available for the material from Air Force being extensively redesigned and enlarged.
development programs and was less expen- Designated the FI A-1SE/F, carboni epoxy
sive than carboni epoxy. During the accounts for lS% of the structure weight com-
intervening years, this situation has changed. pared to 10% for the F I A-I SA. Because of its
Extensive design data have been generated for higher strength and stiffness, the new
many different carboni epoxy systems and IM7/977-3 carbon/epoxy system is being
prices are now significantly lower than for used for the wing and empennage skins
boronl epoxy. instead of AS4/3501-69,12. FI A-1SE/F material
Carboni epoxy was selected for the hori- usage is shown in Fig. 4S.721 •
zontal and vertical stabilizers covers of the In 1976-77, McDonnell Douglas extensively
General Dynamics F-16. These covers are redesigned the British Aerospace AV-SA
bolted to aluminum substructure. More than Harrier vertical take off and landing (VTOL)
3000 F-16s have been produced between since fighter I attack aircraft for the US Marine Corps
1976. to improve range/payload capability.
McDonnell Douglas, with Northrop as the Carboni epoxy was utilized for the complete
major subcontractor, was awarded the Navy wing, horizontal stabilizer and forward fuse-
Air Combat Fighter contract in 1976 with the lage. Designated the AV-SB Harrier II, the
FI A-ISA an enlarged and extensively modi- composite wing weight is about the same as the
fied version of the Northrop YF-17. To save aluminum wing of the AV-SA but has 19% more
weight and offset the increased weight of the area, a 50% increase in internal fuel capacity
landing gear and other components required and a supercritical airfoil. Range-payload capa-
for carrier operation, carboni epoxy materials bility was increased by 100%. Because of high
were used for the wing, tait control surfaces temperatures from the exhaust in the VTOL

Percent of Structural Weight

F/A-18C/O F/A-18E1F
DAlumlnum ______ 49 ________ 31
~ Steel ___________ 15 ________ 14
Titanium ._______ 13 ________ 22
Carbon Epoxy .__ 10 ________ 18
G ather ___________ ~ ________ ~
100 100

Fig. 48.7 McDonnell Douglas F / A-18E/F material usage. (Courtesy of McDonnell Douglas.)
1032 Aircraft applications

mode, carbon/bismaleimide is used for the account for 44% and 7% respectively of the
skins of the inboard flaps and the strakes, structure and rotor weight of the aircraft, are
which are mounted under the fuselage 9,12. estimated to have provided a 25% weight sav-
Usage of composites, which account for 26% of ing relative to a comparable metal airframe12,14.
the airframe weight, is shown in Fig. 48.8. 21 Because weight is so critical to performance of
Composites are also used on other US tacti- this aircraft, much of the structure is currently
cal aircraft. The Navy funded Boeing to design being redesigned to further reduce weight and
a CFRP wing for the Grumman A-6E in 1987 costs 12 •
because fatigue of the aluminum wings lim- The outer skin and much of the substruc-
ited the life of the aircraft. Boeing built 179 ture of the large Northrop B-2A bomber is
composite wings in Seattle, which were made from carbon/epoxy. The B-2A was
installed on the last batches of 21 A-6 aircraft. designed and produced by a team composed
An additional 158 aircraft were retrofitted of Northrop, Vought and Boeing. The B-2A,
with CFRP wings. A follow-on program for which has a wing span of 52 m (172 ft)' a wing
120 wings was canceled by the Navy in area of 464 m 2 (5000 ff) and a gross weight of
September 1993. 180 metric tonnes (400000 lb), first flew on
The first production application of car- July 17, 1989 and is by far the largest 'compos-
bon/thermoplastic composites was the ite' aircraft to fly to date 9,12.
stabilizers of the Lockheed F-117A interdictor. McDonnell Douglas is using 6800 kg
The complete fleet of F-117A aircraft are being (15 000 lb) of composite materials for control
filled with these stabilizers to extend the flight surfaces, stabilizer leading edges, the tail cone,
envelope9,12. landing gear doors and pod, engine nacelles,
The wing, fuselage, tail and the rotors of the and the wing-fuselage fairing of the US Air
Bell/Boeing V-22 Osprey multi-mission VTOL Force C-17A transport.
tilt-rotor aircraft are almost entirely made These components, shown shaded in Fig.
from CFRP and GFRP. Composites, which 48.9, account for 8% of the structure weight

Fig. 48.8 McDonnell Douglas/BAe AV-8B material usage. (Courtesy of McDonnell Douglas.)
 Structural components 1033

and are made by McDonnell Douglas and 26% weight saving26 • The Dassault Mirage
many different American and European sub- 2000, which first flew in 1978, has a CFRP fin
contractors including Grumman (ailerons, designed and built by Aerospatiale, a
elevators, rudders), Heath Techna (fairings), CFRP /BFRP hybrid rudder, and CFRP
Beechcraft (landing gear doors), elevons, nose landing gear doors and some
Aerostructures Hamble (flap hinge fairings equipment doors. Composites account for
and trailing edge panels) and Northwest about 12% of the structure weight and provide
Composites (main landing gear pods). a 25% weight savings of that total. About 350
Since 1970, more than 7500 fixed wing US Mirage 2000 had been delivered by the end of
military aircraft have been produced which 199313,26.
utilized composites for safety-of-flight compo- Dassault Rafale C and M tactical combat air-
nents, such as horizontal stabilizers, wings craft are now in production with CFRP wings,
and control surfaces8• forward fuselage, vertical stabilizer, fin, and
Composite materials have also been used control surfaces with an AFRP radome and jet
for many structural components of European pipe fairings, accounting for about 25% of the
and other military aircraft. In France, Avion structure. Other advanced materials include
Marcel Dassault-Breguet Aviation and superplastic formed-diffusion bonded tita-
Aerospatiale have been involved in composite nium leading edge flaps and canards and
structures development programs with the aluminum-lithium fuselage panels9,13.
French government support since 1972. A British Aerospace (BAe) designed and built
boron/ epoxy rudder was flown on a Dassault a CFRP wing for the SEPECAT Jaguar and
Mirage ill in 1975; horizontal stabilizers were teamed with MBB, in Germany, to design,
flown on the Mirage F I in 1976 and composites build and fly a CFRP taileron for the Panavia
first entered production in 1978 with CFRP Tornado in the early 1980s. The British gov-
ailerons for the Mirage F I, which provided ernment authorized BAe to go-ahead by with

_ CarbonlEpoxy
~ CarbonIKevlarlEpoxy
_ GFRPlNomex Cor.
KlIvlar/Foam Cor.
~ KevlarlNornex
m Ce.bonINomex

-
Fig.48.9 McDonnell Douglas C-17A composite applications. (Unpublished: all rights reserved under copy-
right law by McDonnell Douglas. Reprinted with permission.)
1034 Aircraft applications

development of the EAP (Experimental

Aircraft Programme) in 1982. The EAP had a
carboni epoxy wing with carboni epoxy sub-
structure co-bonded to the lower cover. BAe
made the right wing and Aeritalia, who con-
tributed about 15% of the costs of the program,
made the left wing9,13. The EAP flew in 1986
and the CFRP wings provided much of the
technology and background of experience
used currently by BAe Aeritalia (now Alenia),
CASA and MBB-Dornier (now DASA) for
Eurofighter 2000 (EFA) carboni epoxy wings,
fin, rudder and control surfaces. The first EFA
flew in 19939.
The EAP wing also provided BAe with the
experience to assist SAAB in the design of the
CFRP wing for the Swedish JAS 39 Gripen
multi-mission fighter. The first three and a half
wing sets were manufactured by BAe, subse-
quent wings and all other CFRP components
are manufactured by SAAB in Sweden. CFRP
components include the wings, canards, fin,
outboard elevons, landing gear doors, fuse-
lage and fin fairings and some fuselage access
doors, shown in Fig. 48.10. The Gripen first
flew in 1988 and entered service with the
Swedish Air Force in 19939,13,17.
Other military aircraft which have The .... _ '*"'*'
""""',,. lor hogII ~ one! IO..-m
advanced composite components and are cur- "_10"-"""' _ _ _
.,.". t_ _ To '-!I ",.~ ...., 0Il0Ut 2S "' .. 1110

rently being produced include the Russian

Mikoyan MiG-29 fighter and the Ukrainian
Antonov An-124 heavy transport. The rear
portions of the horizontal stabilizers, fins, rud-
ders, ailerons and flaps of the MiG-29 have
carboni epoxy covers, which account for about
7% of the airframe weight. Mixed composites
are used for many secondary structure appli-
cations on the An-124 stabilizer including
trailing edge panels, landing gear pods, wing-
fuselage and flap track fairings, floor beam (c)
caps, and nacelles. About 5500 kg (12 125 lb) of
CFRP, AFRP and GFRP are used on the An-
124, saving 1800 kg (3968Ib)9,13.
Many technology development and flight
demonstration programs which have incorpo- Fig. 48.10 SAAB JAS39 Gripen (a) Composite wing
rated CFRP wings and other components. assembly; (b), (c) composite applications. (Courtesy
These include the Rockwell International of SAAB Military Aircraft.)
 Structural components 1035

HiMAT, Grumman X-29A, Israeli Aircraft structures. Lockheed designed and made 18
Industries Lavi (the CFRP wing and vertical AFRP fairing panels and 8 AFRP ailerons for
stabilizer were designed and built by L-I011 Tristars. Lockheed changed many
Grumman), the General Dynamics F-16XL and glass/ epoxy panels and fairings to
the Rockwell InternationaljMBB X-31N. Kevlar / epoxy for the extended-range Tristars31.
Aeroelastic tailoring was used on the The experiences gained from the ACEE pro-
HiMAT and X-29A to improve maneuver capa- grams provided the confidence needed by
bility and, in the case of the X-29A, prevent Boeing to select CFRP for the Boeing 757, 767
divergence of the forward-swept wing28,29. and 737-300 control surfaces in the late 1970s.
Domier designed, built and ground tested a Boeing uses about 1500 kg (3300 lb) of
CFRP wing for the Alpha Jet in 198630 • advanced composites in the 757 and 767 and
680 kg (1500Ib) in the 737-300. CFRP compo-
nents include elevators, rudders and many
48.6.2 CIVIL AIRCRAFf APPLICATIONS
secondary structure components made from
Prior to 1972, the major US aerospace compa- CFRP, GFRP and CFRP /GFRP hybrids. By the
nies advanced composite development end of 1988, composite components on Boeing
programs were primarily directed toward mil- airplanes had accumulated more than 9-mil-
itary applications of these materials. Boeing lion flight hours32 • Most of the 757 and 767
flew a boronlepoxy fore-flap on a 707 in 1970, fairings and fixed panels were originally
but commercial aircraft advanced composites aramid/ epoxy or aramid/ carboni epoxy
development essentially began when the hybrid-honeycomb sandwich construction.
ACEE (Aircraft Energy Efficiency) Program Because of surface and matrix cracking, much
was initiated by NASA in 1972. The program of the aramid/ epoxy was replaced with
determined properties of different carboni glass/epoxy, to improve surface protection33•
epoxy material systems after long-term world- Following definition in 1990 of the 777, the
wide environmental exposure and spectrum world's largest twin-engined jet, Boeing
fatigue loading equivalent to 20 years or 36 000 formed design/build teams to develop the air-
flights of airline service. The program also frame and systems. About 15 000 kg
included flight service of numerous composite (33 000 lb) of composite materials are used on
components to obtain confidence in the long- the 777. The empennage, produced by Boeing
term durability of advanced composite at the new Composites Manufacturing Center
structures and materials31 • at Fredrickson, Washington, is made from a
The Boeing carboni epoxy flight service new toughened-matrix carbon/epoxy com-
program, which began in 1973, included 108 posite prepreg, manufactured by Toray in a
Boeing 737 spoilers, ten 727 elevators and four plant next to the Fredrickson facility. The same
737 horizontal stabilizers. The McDonnell material is used for the floor beams, which are
Douglas carboni epoxy program included 20 produced by Rockwell International. Other
DC-I0 carbon/epoxy rudders and a single composite components are supplied by com-
vertical stabilizer. By 1991, one of the DC-lO panies in the USA, Canada, Europe, Australia,
rudders and the vertical stabilizer, which was Singapore, Brazil and Korea. The first 777 flew
installed on a Finnair DC-lO, had respectively in June 1994 and deliveries began in July 1995.
accumulated 58300 and 17580 hoursl8,31. As of September 1996, Boeing had delivered 35
Lockheed made a carbon/epoxy vertical stabi- aircraft and had 281 orders from fourteen air-
lizer for the L-I011, which was tested but not lines 8,12. Composite structure on the 777 is
flown. shown in Fig. 48.1l.
The Lockheed flight service program was The stabilizers have laminated skins with
primarily directed to Kevlar / epoxy (AFRP) cocured stringers, solid laminate spars, simple
1036 Aircraft applications

TT7com.,.,stt....--..
• T~ mot..... ,"',""'""'*' damago r u J _ and damIge -
.,.d petta ... doIgned '''' 11m".., bolted '" _ _ ~..
...

• CorroaIon and fallguo .....tonl

• WoIghl _nlll

Wing fixed _ng . . . - . .

T.-IUng edge pane.. ---fl----J.....,

NoM
..dome

Fig. 48.11 Boeing 777 composite structure. (Courtesy of the Boeing Company.)

honeycomb sandwich ribs and non-structural ment programs, which included components
forward torque boxes. They are designed for for military aircraft and a complete CFRP wing
simple repair. The horizontal stabilizer has a for the Falcon 10, Airbus Industrie began
span of 21.3 m (70 ft) and the projected tail using carbon/epoxy for the A-300-600 spoilers
area is 101 m 2 (1090 ft2), about twice the span in 1983 and for the complete vertical stabilizer,
and four-times the area of the AV-8B wing9,34. rudder, elevators and spoilers of the A310-300
Following successful airline service of the in 1985. The vertical stabilizer is used as a fuel
DC-10 CFRP rudders and fin, deployed under tank on extended range aircraff6.
the ACEE program, and production experi- The entire tail, control surfaces, and cabin
ence from composite components on the floor of the A320, A321, A330 and A340 are car-
MD-80, McDonnell Douglas extended use of bon/ epoxy. The A300 medium-range narrow-
composites on the MD-11 wide-body airliner18• body airliner entered airline service in 1988 the
Almost 5000 kg (11 000 lb) of composite larger A330 and A340 wide-body airliners
materials are used on the MD-11 of which entered service in 1993 and 1991 respectively.
4400 kg (9700 lb) is structural (primarily car- Composite applications on the A320 are
bon/epoxy). Weight savings of 20-30% have shown in Fig. 48.1326 .
been achieved after trading some weight sav- The Airbus CFRP horizontal stabilizers are
ings for improved reliability, maintainability, designed and manufactured by CASA in Spain
durability and producibility. Most of the com- and the vertical stabilizers are manufactured
posite structure is solid-skin or Rohracell foam by Deutsche Airbus in Germany9,26.
sandwich construction. Honeycomb sandwich Avions de Transportation Regionale (ATR),
construction has been avoided because of sus- the Aerospatiale/ Alenia consortium, uses car-
ceptability to impact damage and moisture bon/ epoxy for all the control surfaces and
penetration into the core lS • MD-11 composite Kevlar / epoxy for many components of the
components are shown in Fig. 48.12. ATR 42 regional airliner which was certifi-
As a result of the very successful French cated in 1985.
Dassault-Aerospatiale composites develop-
 Structural components 1037

Wlnglet Trailing Edge

Extended
Tall Cone
~Ele\lato'rs and
Trailing Edge

Wing/Fuselage Fillets Mal n land Ing Ge ar

Door and Struts

Fig. 48.12 McDonnell Douglas MD-ll composite structure. (©1991 by McDonnell Douglas Corporation.
All rights reserved. Reprinted by permission.)

ATR utilizes carbon/epoxy for the complete Many thousands of structural components
outer wing boxes of the ATR 72, a stretched made from composites are currently in airline
version of the ATR 42, the first airliner with a service. Damage assessment and repair has
CFRP wing box to be fully certificated in been a major problem to the airlines. Too many
Europe and the USA. It first flew in 1988 and different fiber-matrix systems are currently
was certificated and entered airline service in being used, even by the same OEM; prepreg
1989. The wing box, which is a fuel tank, is materials are expensive and have a limited life;
made by Aerospatiale at Nantes. The weight of damage assessment requires special non-
the wing box was reduced by 130 kg (286Ib) destructive test equipment and experienced
using CFRP instead of aluminum 14,26. technicians; minor repairs need special skilled
Composite materials used on the ATR 72 mechanics; major repairs must be made in an
and details of the CFRP wing box are shown in autoclave and take many days to complete;
Fig. 48.14. cost of replacement parts or leasing spares are
Dornier chose CFRP for the complete tail much higher than metal parts35•36 •
and rear fuselage of the Dornier 328 regional lATA sent a questionnaire to the airlines to
airliner, which also has an AFRP pressure bulk- obtain information on composites maintain-
head. The 328 was fully certificated and ability in December 1991. The ATA/IATA/SAE
entered airline service in 1993. Material distrib- Commercial Aircraft Composite Repair
ution and details of the CFRP rear fuselage/fin Committee (CACRC) was formed the follow-
box structure are shown in Fig. 48.15. ing year. Committee members include
The de Havilland Canada Dash 8, Embraer representatives from the FAA, airlines, OEMs,
Brizilia, SAAB 340 and other regional airliners NASA and material suppliers36•
are using AFRP or CFRP for many secondary
structure components.
1038 Aircraft applications

. CFRP
_ AFRP
c:J CFRPIAFRP
GFRP

Floor

Lead ing and

Landing gear doors trailing edge
FaIrIngs acceSS
panels

Fig. 48.13 Airbus A320 composite applications.

III carbon I nomu londwich

fTIII (Qrbon monolithic Itructure
II kt!l"liar I nomex sandwich
II :t:ti';:n';~;~~n~1j~
~ fibregrou I nOme); saI1dwich

(a)

CARBON SPARS
(FRONT AND REAR)

(b)
PANEL/PIANO JUNCTION

CARBON PANELS
(TOP AND BOTT().\!)

Fig. 48.14 ATR72 (a) composite applications; (b) CFRP wing box. (Courtesy of ATR.)
 Structural components 1039

_Al__
-".....
filii ........

_ CFAP
--
(a)

_ ""AP

Domier328 CFC Components of the Rear Fuselage

and the Vertical Stabilizer Box

(b)

Fig. 48.15 Dormer 328 (a) composite applications; (b) rear fuselage and fin. (Courtesy of Dormer.)

48.6.3 GENERAL AVIATION APPLICATIONS Th~ Starship and the one-piece CFRP wing,
which has a span of 16.6 m (54 ft) are shown in
The Windecker Eagle, which flew in 1967, was
Fig. 48.16.
the !irst all-composite (GFRP) airplane to
The complete rear fuselage, tail, and canard
obtam full FAA certification. The Eagle was
of the Italian Rinaldo Piaggio P.180 Avanti is
followed by the Lear Avia Lear Fan 2100,
carboni epoxy; the Avanti was granted full
which had an airframe made from CFRP with
certification in 1990, but only 20 aircraft had
some AFRP secondary StruCture17• The Lear
been sold by the end of 1993. Piaggio was
Fan program was terminated in 1985 because
planning to build 10 aircraft in 1994.
of certification problems and costs.
Most of the airframe of the Avtek 400A is
The Beech Starship 2000, which is almost
Kevlar I epoxy with carboni epoxy reinforce-
entirely made from carboni epoxy-faced hon-
ment. Avtek was initiating the FAA
eycomb sandwich construction, received full
certification program in 199338 •
FAA certification in 198837• By December 1993,
Some ~uropean all-composite (primarily
~eech had produced 50 Starships and produc-
tion was put on hold because of slow sales. GFRP wlth carboni epoxy reinforcement)
1040 Aircraft applications

Messerschmitt-BOlkow-Blom (MBB) in Germany,

Sud-Aviation (later Aerospatiale) in France
and Kaman in the United States. The MBB
BO.105, which has GFRP blades, was first pro-
duced in 1967.
With the exception of Sikorsky, almost all of
the helicopters currently in production world-
wide have composite blades.
Aerospatiale (now part of Eurocopter
International, which was formed in 1991 with
the merger of the helicopter divisions of
Aerospatiale and DASA/MBB) began GFRP
composites in the mid-1950s for cowls and
fairings. About 25% of the structure weight of
the AS 365 is made from composites26•
Sikorsky used GFRP for the canopy frame
of the S-61 helicopter in 1959, and by 1967
composites use had been extended to the rear
fuselage skins, doors, the horizontal stabilizer
and the engine cowlings of the Sikorsky 5-76.
Sikorsky and Bell produced composite compo-
nents for the ACEE program and complete
fuselages for US Army Advanced Composite
Aircraft Program (ACAP).
Composites are used extensively for the
rotor blades and much of the airframes of the
Fig. 48.16 (a) Beech starship 2100; (b) starship wing
new Boeing/Sikorsky RAH-66 Comanche.
upper cover. (Courtesy of Raytheon Aircraft Co.) Compared to the UH-60 Black Hawk, compos-
ite usage in the airframe has grown from 9% to
51 %; has shifted from 50% titanium to 67%
aircraft include the British Slingsby T67 Firefly composites for the rotor system; and has
trainer (USAF T-3A), the German Grob GUS shifted from 22% magnesium to 28% compos-
and FFT Eurotrainer. The Russian Sukhoi Su- ites for the drive system. Carbon/epoxy
26 and Su-31 aerobatic competition aircraft accounts for 44% of the airframe weight; the
have carboni epoxy wings and tails. rotor blades are made from carboni epoxy-
CFRP and GFRP materials have been used glass/epoxy broadgoods; the main gear box
for many experimental and home-built and kit housings are made from carbon/bismaleimide
aircraft. These include the Voyager, designed using resin transfer molding (RTM); and the
by Burt Rutan and built by Scaled Composites, tail drive shaft is filament-wound
which made the first unrefuelled non-stop carboni ep0xy39.
flight around the world in December 1986. Composites are being used extensively for
rotor blades and airframes many other new
American, Western European and Russian heli-
48.6.4 HELICOPTER APPLICATIONS
copters including the McDonnell Douglas
Glass/ epoxy main helicopter rotor blades Explorer, Eurocopter BO 108 and the Kamov
were initially developed in the late 1960s by Ka-629.
 Structural components 1041

48.6.5 PROPULSION SYSTEM APPLICATIONS

I
Composite Thrust Reverser (40 in.)
In addition to engine cowlings, cowl doors
and nacelles, advanced composites are being Jet Pipe
utilized for engine casings, thrust reversers,
translating cowls, fan blades and propeller
blades.
A Grumman-developed carbon/bis-
maleimide aft cowl assembly and thrust
reverser for the Fokker 100 is shown in Fig.
48.17(a). This is 25% lighter and far less
expensive than the current titanium cowl
assembly. A production CFRP transcowl for
the General Electric CF6-80C2 engine is
shown in Fig. 48.17(b). The transcowl assem-
bly weighs 86 kg (180 lb). Grumman has been
contracted to build 600 units for the CF6-
80C2, which are used on the Boeing 747 and
767, the MD-ll and the A320.

48.6.6 CONCLUSIONS

The weights of composite structures pro-

duced annually for airliners and military
aircraft, based on five-year averages, are listed
in Table 48.2.
Twenty years ago, advanced composite
structures were only being produced for com-
ponents of military aircraft. This situation has
changed dramatically during the intervening
years: the total annual production weight has
increased by two orders of magnitude and
most of the composite structure (66%) is being
produced for European airliners (Airbus, ATR, Fig. 48.17 (a) CFRP cowl assembly; (b) transcowl.
Domier). (Courtesy of Grumman Corporation.)

Table 48.2 Composite structure production (five-year averages)

Period 1975-1979 1980-1984 1985-1989 1990-1994

Total fly weight: 27 180 905 2620
metric tonnes/year
Distribution:
US military aircraft 21 (78%) 56 (31%) 158 (18%) 172 (7%)
US airliners 3 (11%) 97 (54%) 334 (37%) 582 (22%)
Other military aircraft 3 (11%) 21 (12%) 93 (10%) 126 (5%)
Other airliners 6 (3%) 320 (35%) 1740 (66%)
1042 Aircraft applications

During the past 20 years, advanced com- 15. Schier, J.P. and Juergens, R.J., They Force a Fresh
posite structures have become light-weight, Look at the Design Process. Aeronautics &
reliable, low-risk alternatives to conventional Astronautics, September 1983.
16. Watson, James c., AV-8B Composite Fuselage
aluminum and titanium structures. Their use Design. J. Aircraft, March 1982.
will continue to grow. 17. Lubin, George and Dastin, Samuel J. Aerospace
Applications of Composites. Handbook of
REFERENCES Composites, (ed. George Lubin), Chap 20, New
York: Van Nostrand Reinhold, 1982.
1. Federal Aviation Regulations (FAR), Subchapter C 18. Ashizawa, Moto, Composite Technology
- Aircraft. 14 CFR Chapter I, US Department of Growth Leading to the MD-11 Application and
Transportation, Federal Aviation to the Civil Transport Aircraft of Tomorrow.
Administration, Washington DC. Proc. SAMPE Meeting, Japan, 1991.
2. Abbott, Ric, Design and Certification of the All- 19. Brenner, Lothar and Johst, Eberhard, The
Composite Airframe. SAE paper 892210, SAB Airframe of the Dornier 328 - Proven Progress.
Tech. Conf., September 1989. DGLR Jahrbuch 1989 I.
3. The All-Composite Airframe Design and 20. Hadcock, Richard N. and Huber, John. Specific
Certification. Aerospace Engng, April 1990. Examples of Aerospace Applications of
4. A. Barrio Cardaba, et al., Design and Fabrication Composites. Lecture Series No. 124, Advisory
of the Carbon Fiber/Epoxy A-320 Horizontal Group for Aerospace Research and Development
Tailplane. SAMPE J., Jan./Feb. 1990. (AGARD), October 1982.
5. Waggoner, Gary and Erbacher, Herman, 21. Information supplied by the author by
Damage Tolerance Program for the B-1 McDonnell Douglas Aerospace, St. Louis,
Composite Stabilizer. AIAA Paper 77-464, Missouri. Reproduced with permission.
AIAA/ ASME 18th Structures, Structural 22. Rosato, Dominic V. and Lubin, George, Plastics
Dynamics and Materials Conf., March 1977. in Aircraft and Aerospace. Handbook of Fiberglass
6. Weinberger, Robert A. et aI., US Navy and Advanced Plastics Composites, (ed. George
Certification of Composite Wmgs for the F-18 Lubin), Chap 29, New York: Van Nostrand
and Advanced Harrier Aircraft. AIAA Paper 77- Reinhold, 1969.
466, AlAA/ ASME 18th Structures, Structural 23. Hadcock, Richard N., Boron/Epoxy Aircraft
Dynamics and Materials Conference, March Structures. Handbook of Fiberglass and Advanced
1977. Plastics Composites, (ed. George Lubin), Chap 24,
7. Hadcock, Richard N. Design and Analysis of New York: Van Nostrand Reinhold, 1969.
Advanced Composite Structures. Handbook of 24. Hadcock, Richard N., Status and Analysis of
Composites, (ed. George Lubin), Chap 20, New Advanced Composite Aerospace Structures
York: Van Nostrand Reinhold, 1982. Programs. Composites Technical Note No. CTN-
8. Aerospace Facts and Figures 1992-1993 (and ~y 466-35. Grumman Aerospace Corp. May 1971.
previous editions). Aerospace IndustrIes 25. Chaumette, D. Flight Qualification of
Association of America, Washington, D.C. Composite Structures at Avions Marcel
9. Jane's All The World's Aircraft, 1993-94 (and pre- Dassault - Breguet Aviation. AIAA Paper 82-
vious editions), (ed. Mark Lambert). Coulsdon, 0755,1982.
England: Jane's Information Group Ltd. 26. Information supplied to the author by Avions
to. Ted G. Nicholas. US Military Aircraft Data Book. de Transport Regional, Blagnac Cedex, France.
Various editions. Data Search Associates. Used with permission.
11. Current Aircraft Prices '88 (and other editions) 27. Information supplied to the author by SAAB
Interavia Publishing. Military Aircraft, Scania AB, Linkoping,
12. Aviation Week and Space Technology: Various Sweden. Used with permission.
issues. 28. DeAngelis, V. Michael, In-Flight Deflection
13. Defense News, Flight International, Aerospace, Measurement of the HiMat Aeroelastically
Aeronautics & Astronautics, Aerospace Tailored Wmg. J. Aircraft, December 1982.
Engineering, Air International: Various Issues. 29. Hadcock, Richard N., X-29 Composite Wmg.
14. Aerospace Composites & Materials, 1988-1991 and Evolution of Aircraft/Aerospace Structures and
Aerospace Materials, 1992-1993, Burnham, Materials Cont Air Force Museum/ AlAA, April
England: The Shephard Press. 1985.
 References 1043

30. Rose D. et al., Design of a Carbon-Fiber- Commercial Aircraft Group, Seattle,

Reinforced Wmg for the Alpha-Jet Major Panel Washington. Used with permission.
Tests. J. Aircraft, June 1986. 35. The Cost of Composites. Air Transport World,
31. ACEE Composite Structures Technology: July 1992.
Review of Selected NASA Research on 36. Harris, Charles, Assessment of Practices in
Composite Materials and Structures. NASA Supporting Composite Structures in the
cont Pubn 2321, 1984. Current Transport Fleet. 4th NASA/DoD
32. Harradine, Peter J. and Quinlivan, John T., Advanced Composites Technology Cont, Salt Lake
Composites and the Commercial Jet - A Boeing City, Utah, June 1993.
Viewpoint. Paper No. AIAA-89-2126. 37. Information and photographs supplied to the
AIAAI AHSI ASEE Aircraft Design, Systems author by Beech Aircraft Corporation, Wichita,
and Operations Conference, Seattle, 1989. Kansas, 1993. Used with permission.
33. Hyatt, M. Caton, R. and Lovell, D., Advanced 38. Information and photographs supplied to the
Materials Development in Commercial Aircraft. author by Avtek Corporation, Camarillo
Paper No. AlAA-89-2127. AIAAI AHSI ASEE California, 1993. Used with permission.
Aircraft DeSign, Systems and Operations 39. Garbo, Samuel P. and Rosen, Kenneth M.,
Conference, Seattle, 1989. Composites Usage on the RAH-66 Comanche.
34. Information supplied to the author by Boeing Vertiflite, Marchi April 1992.
COMPOSITES IN THE SPORTING GOODS 49
INDUSTRY
Brian E. Spencer

49.1 INTRODUCTION went into sport and recreation applications.

In 1990 Frost & Sullivan reported that over $800
Worldwide sport and recreation applications
are the third largest user of advanced compos-
million dollars were spent by domestic
defense/ aerospace contractors for advanced ites behind defense/aerospace and elastomer
reinforcement (tires, hose and belts)l. In the
composites. In the same period, $70 million
USA as well, sport and recreation applications
worth of advanced composites were purchased
~re the third largest users of advanced compos-
from domestic producers for sporting goods.
Ites. One of the major growth markets for
Advanced composites are usually defined as
those composites that use either carbon, advanced composites over the past several
years has been the sporting goods industry.
aramid, S-glass, ceramic, polyethylene, boron,
Although there was a total decline in USA
or other high strength or high stiffness fiber. In
1991, the total worldwide consumption of defense spending of only 14% from 1990 to
1994, the procurement decline was 45%2 (Fig.
advanced fibers was estimated to be 27 200
metric tons 1• Approximately 10% of this fiber 49.1). The defense procurement decline was

350 r==::==:=:============:::::lio
- T o t al decline 1 4 % -

1
300 r---\I .. ".-
',303 ' 304 295 ~
,',' 281
fA 250 , 261 264 255 Manpower, R&D

1
<Ii logistics. other
I"
c:

I o
o 254

I,
~200
CD

I ' I
~150 Procurement
::J
III

I
100

IlIlm,WD,_.
._Imlllimi.
50

o~~~~~~~~~db~~
91 92 93 94 95 96 97
Fiscal year

Fig. 49.1 Defense budget declines. (Courtesy of

Handbook of Composites. Edited by S.T. Peters. Published Westinghouse ESG news and department of
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 defense budget.)
 Introduction 1045

disastrous for the composites industry and 49.2 MANUFACTURING TECHNIQUES

resulted in severe upheavals. Although no seg-
Composite products may be manufactured by
ment of the composites market could offset the
several processes, depending on the shape and
defense-induced decline, the sporting
physical properties required. Common
goods/recreation composite segment along
processes are hand lay-u~, ~achine lay~up,
with most other segments is predicted to show
roll wrapping, filament wmding, pultrus~on,
positive growth in the near future3 (Ta~le 4?1). compression molding, platen moldmg,
This growth trend is expected to continue mto
chopped fiber spray-up, resin transfer mold-
the 21st century. Advances in materials and
ing and braiding. Not all of these processes are
processes have reduced consum~r prices f?r ~e efficient for manufacturing sport and recre-
recreational composite whlle provIdmg ation products.
improved performance for the athlete.
Virtually all fishing poles and most golf
Some of the first applications for compos-
shafts are fabricated by the roll wrap process,
ites in sport and recreation were fiberglass which consists of placing pre-cut sheets of
boat hulls and fishing poles. Now the list of
prepreg unidirectional tape or cloth on a man-
products using composites inc~udes al~~st drel, then rolling the mandrel between platens
every sporting and recreation actiVIty.
to consolidate the composite. Fly fishing rod
Products in golf, tennis, and bicycle racing
lay-ups have in the past had all unidirectional
have brought attention to the superior perfor-
0° plies of one material but golf shafts may
mance of composites in sports and recreation.
have two fibers, such as carbon and boron,
Hand lay-up and roll-wrapping have been
and some angle plies. A release cloth or tape is
the processes generally used for most sporting
usually wrapped around the composite to
goods applications. There has been an
apply pressure during the required heat cure
increased interest in filament winding as a pre-
process and hold the composite in place. After
ferred process for tubular products such as
curing, the composite part is pulled from the
golf shafts, sail masts, ski poles, softb~ bats
mandrel and can be machined as necessary to
and bicycle frame tubing, because filament
the final shape. The part is then painted and
winding can lower labor costs and add a new
graphics are added.
level of design flexibility, product consistency,
Braiding and pultrusion have been used
and quality for these products. successfully for some applications. Tennis

Table 49.1 Annual growth rate of selected materials by region, 1993-2003

Product/end use US average Rest of world World average

annual average annual annual growth
growth % growth % %
Advanced polymer composites
Aircraft 6.5 8.5 7.45
Missiles/ space 1.0 4.0 2.2
Recreation 6.0 3.5 5.0
Armor 6.0 4.5 5.4
Automotive 50.0 2.0 26.7
Industrial/ other 4.0 5.0 4.6
Total 7.1 5.4 6.4
1046 Composites in the sporting goods industry

rackets have been fabricated by braiding. The bath, along the fiber path between the ten-
braiding process for composites is identical to sioning system and the delivery eye, through
braiding textiles, except that the fiber is usu- which the dry fibers are pulled. Both wet
ally braided as a prepreg. Prince introduced winding and prepreg winding are widely
the Vortex tennis racket, braided with co-min- used. The choice of technique depends on the
gled nylon and carbon fibers, in the late 1980s. application, cost and quality requirements.
Ski poles, kite tubes, arrow shafts, and similar Composites that have directionally oriented
hollow or solid shafts are made by the pultru- continuous fibers are analyzed by techniques
sion process. Pultrusion of composites is such as netting analysis, rule of mixtures, clas-
similar to extruding metals. The resin and sical lamination theory, and micromechanics.
fiber material are pulled through a heated die These analysis techniques all account for the
that forms the composite to shape and cures fact that composites have material properties
the resin in one continuous process. Since the that depend on the fiber directions of each ply.
fibers are essentially all aligned down the axis The overall directional properties of the lami-
of the part there is little transverse or hoop nate are calculated by adding the relative
strength unless hoop fibers or other off-axis effects of each ply. The analyses are more com-
reinforcement such as mats are added during plicated than with isotropic materials such as
the process. Arrow shafts have shattered on steel and aluminum. However, the wide-
impact because of this lack of strength in the spread use of computers has made the task of
hoop direction. analyzing composites relatively simple.
Resin transfer molding (RTM) consists of Composites are in many cases much
injecting resin into a mold containing fiber in a stronger and stiffer than metals such as steel,
preform pattern. Resin transfer molding is aluminum or titanium. When one takes into
usually used in conjunction with a fiber place- account the much lower density of composites
ment technique such as hand lay-up, braiding as compared to metals, composites have an
or filament winding. Some tennis rackets are even greater performance advantage over
fabricated using the resin transfer molding metals. To compare material properties on an
process4 • equal weight basis (instead of an equal volume
The filament winding process is the basis), a property such as ultimate strength is
machine controlled technique of applying a divided by the density of the particular mater-
controlled amount of resin and oriented fiber ial. The resultant is termed specific strength.
on a mandrel that provides the finished com- The same can be applied to stiffness (modulus)
ponent shape. The mechanics of filament and is called specific modulus. Comparing
winding are illustrated in Fig. 49.2 The ten- materials this way is very important when
sioned fiber bands are accurately placed on designing weight critical parts.
the mandrel using numerically computer Figures 49.3 and 49.4 show the strength,
controlled or chain and gear mechanical modulus, and specific properties of represen-
winding machines. Fig. 49.2 shows the fiber tative composites and metals. The composite
being pulled from tensioned spools through data are for composites with all fibers oriented
the delivery eye onto the rotating mandrel. in one direction (unidirectional ply data). The
The resin can either be pre-applied on the fiber test loads are applied to the specimens in the
in an off-line operation (pre-impregnation, or same direction as the fibers. Also, all compos-
wet prepreg) or applied during the winding ite data presented here are for 60% fiber and
operation (wet winding). Prepregging usually 40% resin by volume samples. This ratio is
requires storing the material at low tempera- typical for advanced composite laminates. The
ture until required for winding to retard the epoxy resin used to make the samples is not
curing process. Wet winding uses a resin indicated specifically because its effect on the
 Manufacturing techniques 1047

orlzontal and Vertical

Headstock Carrlagea
Tallstoc

000
000

(b)

Fig. 49.2 Layout of a typical filament winding machine (Courtesy of Composites Machines Company.)

unidirectional ply properties is negligible a product. Fatigue describes the phenomenon

compared to the effect of the fibers. of reduced strength over time of load applica-
In Fig. 49.3 it can be seen that composite tion (static fatigue) or the number of times a
plies are up to twice the strength of high load is applied (cyclic fatigue). Aluminum and
strength steel and three to five times stronger fiberglass composites suffer from poor fatigue
than titanium and aluminum. Comparing performance whereas carbon composites excel.
specific strengths, composites are three to Figure 49.5 displays the cyclic fatigue
eight times stronger than the metals. The stiff- strength of various metals and composites.
ness of the materials, shown in Fig. 49.4, range Specimens of each material were loaded
from fiberglass composite at the low end to repeatedly in tension to various predeter-
high modulus carbon composite at the high mined loads and then the loads were removed,
end with metals in between. Composites com- until the specimens failed. A curve of cyclic
pare more favorably to metals in specific stress against number of cycles to failure can
modulus. The specific modulus of the com- be constructed for each material, from which
posites shown are one to six times greater the 10 million cycle failure stress for each
than for metals. material can be obtained. The 10 million cycle
Fatigue resistance is another important life is considered by designers to approximate
material property to consider when designing an infinite life for most applications.
 1048 Composites in the sporting goods industry

Specific Strength, E6 in.

Material 0 2 4 6 8
34 msi carbon

42 msi carbon

50 msi carbon

S-2 glass

E gla s

Kevlar

Spectra 1000

Titanium

4340 Steel

Aluminum
o 100 200 300 400 500
Strength, k i
• •_ Strength ....._ _...1 Specific Strength

Fig. 49.3 Strength of composites; comparison with metals at 60% fiber volume.

::>pecUlc MoaUIUS, co In .
Material 0 200 400 600 800
34 msi carbon

42 msi carbon

50 msi carbon

5-2 glass

E glass

Kevlar

Spectra 1000

Titanium

4340 Steel

Aluminum
0 5 10 15 20 25 30 35
Modulus, Msi
Modulus Specific Modulus

Fig. 49.4 Stiffness of composites; comparison with metals at 60% fiber volume.
 Sporting goods applications 1049

Tension - tension ratio at 10 E+ 7 cycles

Material
34 msi carbon
42 msi carbon
50 msi carbon
S-2 glass
E glass
Kevlar
Titanium
4340 Steel
Aluminum
100 200 300 400 500 600
Alt. stress density, ksi/lb. per cu. in.
Fig. 49.5 Fatigue strength of composites; comparison with metals at 60% fiber volume.

The failure loads shown in Fig. 49.5 are because of the changes in 'feel' of the rod.
given as stress density which is the test stress Many golf shafts incorporate both
divided by the material density. For the mate- carboni graphite and boron reinforcements to
rials tested, composites have two to six times change flex point, reduce shaft weight, increase
the specific fatigue resistance as metals. torque resistance and increase clubhead
It is important to mention that the previous speedS.
comparisons of metals and composites do not
tell the whole story. Often, the benefits of a
49.3 SPORTING GOODS APPLICATIONS
composite part compared to a metal one
would not be as great as indicated by the dif-
49.3.1 RACKETS
ferences in ply strength and stiffness. The
design allowable composite ply strength is There are approximately 22.5 million tennis
10-20% lower than the test specimen players and 7 million racquetball players in
strengths. Other factors including voids, long- the USA. These players purchased 2.45 mil-
term environmental effects, use temperature lion tennis rackets at a cost of $170 million and
and imperfect fiber alignment all reduce the 1 million racquetball rackets at a cost of $53.4
ultimate strength of composites. Finally, when million retail dollars in 1992. Approximately
all the reinforcing fibers are not in the direc- 73% of these rackets were composite4•
tion of the load, the fibers do not carry the load Most all of these rackets are produced in
as efficiently. However, when these factors are the Far East using a compression molding
adequately controlled, there are many applica- process with an internal bladder. This process
tions where composite structures greatly consists of hand lay-up of prepreg around a
outperform their metal counterparts. Sport straight metal mandrel and plastic bladder.
and recreation applications are one market After lay-up the prepreglbladder is removed
area where composites have a decided advan- from the mandrel and placed in a compres-
tage over metals. sion mold. The mold is closed, the bladder is
There are other not easily quantifiable rea- pressurized, and the part is temperature
sons for choosing a particular composite or an cured. A significant portion of the labor in
individual lay-up. Fishing rods generally do manufacturing a racket is in the surface fin-
not have hoop reinforcement or angle plies ishing, painting and labeling.
1050 Composites in the sporting goods industry

Other processes that have been used

include injection molding, braiding, and resin
transfer molding and combinations of these
processes. Filament winding is being consid-
ered because it can allow the use of wet
winding, which reduces the raw materials
costs and labor content in racket manufactur-
ing. This process uses wet resin and dry
fibers and provides product consistency
because the operation is machine controlled.
The preform can be wound on a straight man-
drel to allow standard filament winding
techniques to be used. After winding, the pre-
form can be removed from the mandrel and
placed in the mold. The key technology is to
place the wet fiber preform in the mandrel
without disturbing or kinking the fibers. The
process then continues as the typical
processes now being used. A typical product
of this method is shown in Fig. 49.6. Filament
winding can also be used to dry wind the
fiber to make a cheap oriented fiber preform.
Then using RTM, the resin can be injected
with the prospect of further reducing hand
surface finishing. An additional advantage to
filament winding is the design flexibility of
changing strength and stiffness around the Fig. 49.6 Prince graphite extender racket. (Courtesy
racket head, which is difficult to accomplish of Spyrotech.)
using lay-up techniques.
impact on the acceptance of carbon fiber com-
posites in the bicycle industry. In 1991 more
49.3.2 BICYCLE FRAMES AND COMPONENTS
than 50% of the carbon bicycles produced
Advanced composite bicycle frames first worldwide (> 100 000) were manufactured in
appeared in the early 1970s. These frames Taiwan6.
were not successful due to poor design and Tube and lug designs dominated the mar-
manufacturing techniques and market resis- ket until 1993. The technique consists of
tance to the new materials. In the mid-1980s, bonding premade tubes into forged or cast
composites in bicycles had success in the mar- metal lugs. Trek developed a composite lug
ket place and today are the materials of choice that is offered in some of their frames. Kestrel
for new designs6. For the past few years introduced a molded monocoque frame in
approximately 12 million bicycles were sold 1987. Molded monocoques are manufactured
annually in the USA. Of these, approximately by lay-up of composite material into female
10% were considered high-end (sales price mold halves with the reinforcing metal pieces
over $750) products. as required. The female mold sections are
In 1989 and 1990 Greg LeMond won the brought together and internal bladders are
Tour de France Bicycle Race using a carbon inflated to provide compaction pressure dur-
fiber frame. His success has had a significant ing the curing operation.
 Sporting goods applications 1051

Tube and lug frames are relatively simple to helps to reduce the seam effect on perfor-
manufacture. However, the frames can have mance consistency.
durability problems if the tubes and tube to Grafalloy introduced a filament wound
lug jOint are not properly designed to prevent shaft for the 1993 season. Filament wound
significant stiffness mismatches. Galvanic cor- shafts provide superior performance at low
rosion potential is a concern. Filament weight and cost compared to roll-wrapped
winding offers an inexpensive method for shafts available with comparable performance
manufacturing frame tubes. Wet winding specifications. Those familiar with the indus-
techniques allow using the required materials try believe filament winding will be the
in the lowest cost form. Since the process is process of choice in manufacturing golf shafts
machine controlled, labor requirements are in the next few years.
minimized. Currently the most widespread Player tests and mechanical golfer testing
manufacturing method is roll-wrapping. Roll- have proven the performance improvements
wrapping requires cutting prepreg plies of of a filament wound shaft over roll-wrapped
composite and hand wrapping them around a shafts. Many original equipment manufactur-
metal mandrel prior to curing. ers are evaluating filament wound shafts for
Frame tubes are not the only use of com- their product lines. One factor where roll-
posites in the bicycle industry. Handlebars are wrapped shafts can have an advantage over
filament wound, removed from the mandrel, filament wound shafts is in tip durability.
and then cured in a mold to provide the Filament winding can introduce more voids in
proper shape and curvature. An internal blad- the composite laminate than roll-wrapping.
der provides the compaction pressure. Seat Special attention is needed by filament
posts, swing arms, and handlebar extensions winders to insure that their product has suffi-
can be made using this process. Titanium cient tip strength.
tubes in some front wheel suspension systems ASTM is working to establish test methods
are reinforced with composite tubing. Wheels for golf shafts. Standard tests will allow for a
and spokes are also being made of composite better comparison of products and give the
materials using combinations of roll wrap- customer more confidence in the quality and
ping, hand lay-up and resin transfer molding. performance of the product being purchased.

49.3.3 GOLF SHAFTS 49.3.4 FISHING POLES

In 1992 over 100 million composite golf shafts In 1992, 16 million fishing poles were pur-
were made world wide with a wholesale dol- chased in the USA. VIrtually all were produced
lar value of $900 million7• The majority of the in the Far East using the roll-wrap process. It is
shafts were manufactured in the Far East and estimated that about 20% are high-end prod-
sold for less than $10 each wholesale. At the ucts and may be amenable to the
present time, virtually all composite golf manufacturing costs of automated processes in
shafts are manufactured using the roll-wrap the USA. For filament winding, this market
process. This process relies on manual orien- segment is the least developed. Development
tation of the plies of material on the mandrel work is underway to produce a cost competi-
and ply alignment can be inconsistent from tive high-end fishing pole in the USA.
part to part. The alignment directly affects the
shaft performance; flex, and torque. The cut
49.3.5 OTHER APPLICATIONS
plies also create a seam down the length of
the shaft which causes inconsistency in the Other applications for composites in the sport-
shaft performance. Attention to ply alignment ing goods industry include softball bats, pool
1052 Composites in the sporting goods industry

cues, kite tubing, shoe inserts, ski poles and REFERENCES

model aircraft. The majority of these products 1. Segal, c.L. Worldwide Markets For Advanced
are currently being manufactured by either Fibers, 23rd Intern. SAMPE Tech. Conf.,
roll-wrapping, hand lay-up, or pultrusion. Kiamesha Lake, NY, October 23,1991.
Improvements in material properties, reduc- 2. Anon. Westinghouse ESG News, Oct, 1994.
tion in material prices, and significant 3. Kline & Company The Changing Advanced
improvement in component performance has Polymer Composites Market, 1993.
4. Feeney, B. Composites in Racquetball and Tennis
created a market for composites in the sport-
Rackets, 5MB Conf. Composites Manufacturing
ing goods industryS. Unfortunately for USA and Tooling '94, Anaheim, California, January
manufacturers, the majority of the sporting 17-20.
goods products are fabricated overseas. This 5. Levin, S. The Use of Composite Materials in the
business will only come back to the USA if Bicycle Industry, 5MB-Effective Applications of
companies are cost competitive and/ or offer a Composites in the Sporting Goods and
product of higher value which is superior in Recreational Industry Clinic, August, 1992.
6. Textron Specialty Materials, Boron Backgrounder.
performance. One method to improve cost
7. McConnell, V. Composites in sports: going for the
competitiveness is through the use of automa- gold, Adv. Comp., September/October 1992.
tion. Filament winding, along with innovative 8. McConnell, V. Sports Applications - Composites
engineering, offers the possibility of bringing at Play, High Performance Comp., Jan/Feb 1994.
composites to the forefront in the sporting
goods industry.
TYPICAL PROPERTIES FOR APPENDIX A
ADVANCED COMPOSITES
Kenneth R. Berg

A.I INTRODUCTION material and configuration is selected, a mini-

mum test program would then be initiated.
For a company or institution that is designing
composite material structures, or embarking Having a set of typical composite materials
for the first time into the application of has advantages and disadvantages. For exam-
ple, if one were to design a structure utilizing
advanced composite materials for structural
purposes, it is imperative that material prop- only typical material properties, without the
knowledge of the scatter that may occur in
erties be available. Of course it would be
those properties, structural failure may occur.
desirable to have a complete set of statistical
Design Allowables, such as the statistical '.A: Perhaps not immediately, nor on every struc-
values for properties, or even the 'B' values, ture produced, but on an unknown statistical
basis, at some point in time. However, prior to
(see Chapter 33 for detailed definitions of
a final design for a structure, the normal engi-
these values and Neal and Spiridgliozzi, 1987).
neering procedure is to initiate the test
Since complete statistical Design Allowables
are not available, the next sought after mater- program. The purpose of the test program is
threefold: one, confirmation of the design;
ial properties would be 'typical' properties.
two, determine the scatter that occurs due to
However 'typical' properties are not defined
statistically and may be defined in many dif- variations in materials and the manufacturing
ferent ways. Therefore it is important to process; and three, over a period of time,
discuss typical material properties and also either to confirm the material properties data-
discuss the means to achieve a set of typical base being used, or to accumulate test data for
a material properties database.
properties.
The purpose of having a complete set of
typical properties is to be able to design com- A.2 TYPICAL PROPERTIES - CONSTITUENTS
posite structures with a minimum of testing
confirmation. Having a complete set of typical A.2.l FIBERS
properties will allow design optimization, pre-
liminary design, cost and weight optimization One of the problems of determining typical
and other trade-offs with a number of different properties is the variations that occur in the
materials and candidate laminates with differ- materials making up laminates. In the case of
ent fiber orientations. Once an optimum glass fiber, the types of glass fiber and number
of manufacturers is considerably less than
with carbon fiber.
Handbook of Composites. Edited by S.T. Peters. Published
However, even with this limitation, there are
in 1998 by Chapman & Hall, London. ISBN 0 412 54020 7 at least two major types of glass fiber, E-glass
1054 Typical properties for advanced composites

and 5-2 glass. Within each of these glasses are Table A.l 'TYPical products from carbon fiber man-
variations in chemical composition, fiber diam- ufacturers (available in USA)
eter, fiber finish, fiber sizing, the number of
AMOCO (Thornel)" Toho Rayon (Besfight)
individual fibers in a tow, roving, yarn, etc.
Manufacturers have different names for the P-100 HTAWlOO
similar type of glass, for example the higher P-75 IM600
strength, higher modulus glasses. These fiber T-300 HM35
T650/35
glasses are the older S-glass (no longer avail- T650/42 Toray (Torayea)
able commercially), 5-2 glass and the R-glass T300
by a French manufacturer. Other countries fab- Toho (Celion) T700s
ricate the same type of glass, but with only G30-500 T800H
minor differences in properties. G40-800 M40
For carbon fiber, not only are there the same G40-600 M46J
variations as mentioned above for glass, but in Grafil (Grafil)b AKZO (Fortafil)
addition, there are large variations in strength
34-700 F-5
and modulus and in manufacturers, see Tables 42-650 F-3
Al andA2. 42-750
Based on the large number of variations in
fibers, it would be virtually impossible to Hexeel (Magnamite)
obtain complete statistical material properties AS4
for each variation. Even to obtain typical prop- IM6
erties for each variation would not be IM7
IM8
practical.
To reduce this problem to a practical level, it • Typical product name
is necessary to analyze the usage of glass and b Grafil is a subsidiary of Mitsubishi Rayon Co. Ltd and
carbon fibers (or other fibers). The usage of their fibers are called Grafil or Pyrofil.
advanced composite fibers by 'pounds used
per dollar expended', is estimated to be, in
order of highest usage: E-glass, high strength Table A.2 Torayca fiber types (Toray, 1991)
carbon (modulus of 227 GPa, (33 x 1()6 psi»
Fiber type Number offilaments
and then 5-2 glass. With this list, it is possible
to develop typical properties for composites T300 lK, 3K, 6K, 12K
fabricated from each of these fiber types. An T300J 3K,6K,l2K
T400H 3K,6K
example of the determination of the strength
T700S 12K
and modulus of the typical high strength car- T800H 6K,12K
bon fiber is shown in Fig. AI. The 'typical' TlOOOG 12K
property becomes: Tensile modulus of 227 GPa T1000 12K
(33 x 106 psi), and tensile strength of 4000 MPa M35J 6K,12K
(580 ksi). M40J 6K,12K
M46J 6K,12K
M50J 6K
A.2.2 RESIN SYSTEMS M55J 6K
M60J 3K,6K
The matrix for fiber composites can be classi- M30 lK, 3K, 6K, 12K
fied into two categories, metallic and M30SC 18K
non-metallic. This discussion on typical prop- M40 1K, 3K, 6K, 12K
erties involves only non-metallic resin matrix M46 6K
systems. M50 1K,3K
 Typical properties - constituents 1055

300,---------~----------------_.
Typical PropertY -----'-~
250 H'8treiigii,H4GpaH H'~*~~H

1
Data from vUlaus lber manufacturers
50················ 5 • Epoxy/Amini + Epoxy/Anhydrld. lIE Vlnyl ••t.r
I Po.,..... I!!I 'J'tpIc.1 Pro...rty
OL---~----~----~----~----~ 0
o 2 3 4 5 0 100 2ao 3DD 4DD 5DD 100 7ao 8DD lao I.ODO

Strength OPa Strength MPa

Fig. A.l Typical strength and modulus for high Fig. A.2 Typical strength and modulus for E-glass
strength carbon fiber. (Courtesy of Riggs composite - flexural strength. (Courtesy of Riggs
Corporation.) Corporation.)

As was discussed for fibers, only the high environmental considerations. Key among
usage matrix systems in advanced composites these characteristics are: temperature, frac-
are considered as candidates for typical prop- ture toughness, compression after impact,
erties. In addition, for typical properties of crack propagation, humidity, stress concen-
advanced composites for structural applica- trations, interlaminar shear, mechanical
tions, only structural resin systems are fasteners in laminates, holes in laminates,
candidates. Structural resins are defined as creep, damage tolerance and compatibility
resins that have similar modulus and tensile with fiber finish. In determining typical prop-
strength as standard epoxy systems. For exam- erties, these characteristics are not included
ple, an applicable resin for structural but, as applicable, need to be considered for
composites would have a modulus of approx- the final design.
imately 3.5 GPa (0.5 x 1()6 psi) and a tensile
strength of approximately 100 MPa (15 ksi).
The more popular structural resins are
polyester, epoxy, vinyl ester and phenolic. For 800r----------------------------.
typical composite properties, the use of any of 700
!'rvPIO.' Value 840 MPa ~
." .
these resins will allow a single typical prop- :.:I 800 ... '.~" . .
erty (Fig. A.2) (CertainTeed Corporation,
1989). Isoo
A comparison for composites with different !400 ................. .
epoxy sizing from different manufacturers and j
a typical value is shown in Fig. A3.
Figure A.4 shows a number of different 1 300

~:::Ir-.-C-.I-I~-~-·~-~-~~-·PO~·-·_3----*-~--IIO~n-Q-30-~-_F-----.1
resin systems and the typical values for
strength and modulus (Lubin, 1987). This data I H.rcul•• AS-4/1'ype G X Harcula. AS-4/Treated
is for primarily fiber controlled properties. o~--------------------------~

There are properties in which the resin is

the significant factor. These characteristics are Fig. A.3 Typical compression strength for carbon
associated with stress concentrations and fiber fabric composites. (Courtesy of Riggs
Corporation. )
lOS6 Typical properties for advanced composites

100

\
. Typical. VaLues .
90 Strength 608 MPa

.
.. ·Modulus 7c} GPa . ...... .....
80
ca
a. 70 X"'~ !I
CJ 60
en
::;, 50
'3
'tI 40
0
::IE 30
20 +H.rOUI •• A..1eap *H.rcul•• A47D-8H • Hereul•• A-170-5H
XC.llon W-11U • Flbarlte HMF-1tSlM AFlb.rlt. HMF-.l41/34
10
0
0 100 200 300 400 500 600 700
Tensile Strength MPa

Fig. A.4 Typical strength and modulus for carbon fiber fabric composites. (Courtesy of Riggs Corporation.)

A.3 TYPICAL PROPERTIES - COMPOSITES ence the transverse strength and modulus of
the base unidirectional laminate (100% 0°
A.3.l FIBER CONTROLLED TYPICAL plies). The different strength resins shown on
PROPERTIES Fig. A.S, are as follows:
Fiber-reinforced composite materials are pri-
marily used to take advantage of the high Transverse Transverse
strength and stiffness of the fiber. Therefore in modulus tensile strength
(GPa) (MPa)
most applications, the laminate orientation is
designed so that the strength and modulus are Lower strength resin 8.6 41.4
controlled by the fiber properties. For example, Typical strength resin 10.0 55.2
Higher strength resin 11.4 69.0
for a typical fiber orientation in a laminate of
0° /±fP /90°, the 0° plies control the failure of
the laminate whenever the percentage of 0°
plies is greater than 10%, (with a () greater than 700 r--~La~m~l~n.~I.~1~5"~OOr<.~15:="-=±.::..8°.L'7!.::O-"'-"-"'9O""-r--:-: =..::-;_=....=RMI=.
.•.:7:
tow

±l00). For () less than ±l0°, if the combined per- soo ... ,Fiber f.-ctu,. Cril cal (Ultimate) ....
.. ,:::.:;:::;
~lJpIcaI.lrength ItHln

*" Hlghwlnagth "-til

centage of 0° and ±f)0 plies is greater than 10%,
..
... La• .,8......
~

*'J\'pk:aI .....gtIII"...n
then the laminate is also controlled by fiber
fracture. These limitations are approximate ---------,=
and depend on the fiber strength and modulus,
but are valid for carbon, aramid fibers and
glass. For high () angles and for 90° laminates, 100~~R-
••-m-c~--.I~(R~NP~~-~-IU~-)-------+_------__4

the tensile strength of the laminate is reduced oL-________L -________L -______ ~

o 15 30 45
below the tensile strength of the resin due to
Angle ±8°
stress concentrations between the fibers (Berg,
1967). Figure A.S shows the tensile strength of Fig. A.S Effect of resin properties on tensile
a carbon fiber laminate of 0°/ ±f)0/90° plies, strength of high strength carbon composites.
with the strength of the resin varying to influ- (Courtesy of Riggs Corporation.)
 Selecting the use of typical properties lOS7

As can be seen in Fig. A.S, the typical fiber A.4 SELECTING THE USE OF TYPICAL
fracture composite properties (ultimate fail- PROPERTIES
ure) are not affected by large variations in the
properties of the resin. However, for resin sen- AA.1 WHERE TYPICAL PROPERTIES ARE
sitive properties, for example, the first ply ACCEPTABLE
failure (limit) of the 90° plies, the resin prop-
The acceptability of material properties in the
erty is influential.
design of structures is based on a number of
For the fiber fracture controlled composites,
factors. If a design is being produced for a cus-
which are the main interest in structures, typi-
tomer, the customer is often the final word on
cal composite material properties are valid
the acceptability of the material properties uti-
over a wide variation in resin characteristics.
lized. An alternate possibility is that the
A.3.2 MATRIX CONTROLLED TYPICAL customer is not interested in accepting or
PROPERTIES rejecting the material properties used, but
would rather accept the product against a
The main interest in structural components is specification. Final acceptance is a qualifica-
the high strength and modulus obtainable tion of the product through testing.
from the fiber reinforcing of the matrix. The third case is where a product is pro-
Therefore in the design of the laminate, for duced by the company itself and sold to the
most of the applications, the resulting failure consumer directly. Of course the consumer
modes are fiber fracture critical. Unfortunately, (public) is not interested in accepting or rejec-
there are cases where the matrix is the critical tion the material properties database. An
element in the failure mode for all laminate example of this is the automobile industry.
orientations. This does not mean that the fiber In the case of a company that subcontracts
does not contribute to reinforcing the matrix in the design and fabrication of composite struc-
both strength and modulus, but only that the tures, the company may either want to review
ultimate failure is the result of failure in the the subcontractor's material properties data-
matrix. base, or be able to review the design of the
Among the cases where the matrix is the subcontractor with the use of the company's
critical failure mode are laminates that are typical properties database.
subjected to shear, the first ply failure (limit In each of the above cases, a typical mater-
load) of a laminate (Fig. A.S, first ply failure ial properties database can be used for cost
curves and Berg, 1982) and most cases of the and weight trade-offs, selection of the best
transverse strength property of a laminate materials, optimizations studies and prelimi-
with no 90° plies. Even for these cases, a typ- nary design.
ical property can be determined utilizing the It is important to note that the final design
typical set of unidirectional properties repre- would always go through an extensive test
sented by most of the epoxy systems in use program to verify the material properties
by the prepreg manufacturers as well as most selected, the final design, the manufacturing
of the structural epoxies sold. A typical set of process and to determine the variability of the
unidirectional properties for high strength product.
carbon composites, for example would be: Thus, a typical material properties database
Strength 0° 90° is acceptable and useful to reduce the cost of
Tension 2100 MPa 55 MPa engineering design, reduce the cost of testing
Compression 1380 MPa 190 MPa and allow a more intelligent and less time con-
In-plane shear 55 MPa suming approach to the final design.
Modulus 138 GPa 9.7 GPa
Shear modulus 6 GPa
1058 Typical properties for advanced composites

A.4.2 WHERE TYPICAL PROPERTIES ARE NOT the composite transformations equations (Tsai,
ACCEPTABLE 1988). A typical property database for all ori-
entations and selected materials is available
In the above discussion, the customer require-
(Berg, 1993), but also a typical composite mate-
ments were mentioned as one of the criteria
rial database could be generated by a
for acceptance or rejection of a material prop-
company using composite engineering analy-
erties database. There are cases where the
sis as discussed above.
customer will insist upon enough testing to
Although this discussion of typical proper-
develop a statistical property database. For
ties has mentioned only static strength and
that requirement the typical material database
modulus, typical properties are also available
would not be acceptable. However in any sta-
for fatigue, CTE and for elevated temperatures
tistical database, the applicability is confined
(Berg, 1993). Fatigue typical properties include
to the specific fiber, matrix and fabrication
all fatigue stress ratios. Fatigue statistical
process. The statistical testing process is
properties, of course, would be prohibitively
expensive and time consuming. In most cases,
expensive.
the statistical database would be limited to the
If users of typical composite material prop-
laminate orientations tested, which would also
erties are aware of their limitations, typical
be very limited.
properties can be a very useful database for
For applications where significant environ-
cost effective design and analysis.
mental conditions are present, the use of typical
material properties may not be applicable due
to large variations in the response of different
REFERENCES
resins to these environmental conditions. These
conditions were summarized above and Berg, K.R. 1967. The effect of fiber spacing on the
included impact, humidity other corrosive flu- strength of composites laminates, Paper pre-
ids, stress concentrations, temperature, etc. sented at AAWASME 8th Structures, Structural
Dynamics and Materials Conference, Palm
Springs, CA.
A.S SUMMARY Berg, K.R. 1982. The effect of matrix strain limita-
The design of composite material structures tions on composite design allowables. Proc. 27th
Nat. SAMPE Symp., May, 1992, San Diego, CA.
~quires a knowledge of the material proper-
Berg, K.R. 1993. Composite Material Properties Data
ties for all combinations of laminates. It is Books - Sample Data Sheets. RIGGS Corporation,
cost-prohibitive to test all combinations of Seattle, WA.
laminates, even to obtain typical properties. To CertainTeed Corporation. 1989. Sales Brochure, Test
obtain statistical design properties for a lim- Data, Wichita Falls, TX.
ited number of laminate configurations is also Hashin, Z. and Rotem. A. 1975. Failure modes of
expensive, but in some cases may be required angle ply laminates. J. Composite Mater., Volume 9.
Lubin, G. (ed.) 1987. Handbook of Composites. New
by a contract. Typical composite material York: Van Nostrand Reinhold.
properties can provide useful data and be cost Neal, D. and Spiridgliozzi, L. 1987. An efficient
effective for the design engineer. The data can method for determining the A: and 'B' design
I

be generated by utilizing typical unidirec- allowables. Army Materials and Mechanics

tional data for each class of materials, (as Research Center, Watertown, MA.
discussed above for high strength carbon com- Toray Industries, Inc. 1991. Torayca Sales Brochure,
posites). To generate all the laminate Typical Fiber Properties.
Tsai, Stephen w. 1988. Composites Design. Dayton,
configuration, both limit (first ply failure) and
OH: Think Composites.
ultimate, requires a comprehensive computer Tsai, Stephen w. and Hahn. H. 1980. Introduction to
program, including a failure criteria (Tsai, Composite Materials. Westport, CT: Technomic
1988; Hashin and Rotem, 1975) and utilizing Publication.
SPECIFICATIONS AND APPENDIXB
STANDARDS FOR POLYMER
COMPOSITES
Frank T. Traceski

limits. A list of typical composite material

B.l USES OF SPECIFICATIONS AND
qualification tests is provided in Table B.I.
STANDARDS
These examples are representative of the mate-
Material specifications and engineering stan- rial properties which are determined in a
dards for advanced composite materials have typical material qualification program in
a very broad applicability across the entire accordance with various test standards.
spectrum of defense and commercial applica- A concurrent engineering approach to com-
tions. From basic research, through engineering posites engineering requires that material
and manufacturing development, in produc- selection (as part of the design process) be inte-
tion, and for maintenance and field repair, grally linked with engineering and
material and process specifications establish manufacturing process development. In com-
requirements and procedures. posite manufacturing development, material
In research and development, testing stan- process specifications are defined for a given
dards are used for material characterization to composite material and manufacturing process.
determine physical, chemical, mechanical, ther- Once optimized, process specifications reduce
mal and electrical properties. In manufacturing manufacturing risk and enhance producibility.
development, material specifications are used Another aspect of composites manufactur-
to establish material quality and processability ing is the nondestructive testing and
requirements. In production, the material pro- inspection (NOTI) of composite structures to
ducer uses test standards and material verify structural integrity during production.
specifications for statistical process control Ultrasonic inspection, infrared thermography,
(SPC) to ensure batch-to-batch consistency. and a host of other NOTI methods are
Material testing standards are used exten- employed to locate voids, delaminations,
sively in design engineering to determine cracks, and other types of structural defects.
material design allowables (i.e. strength and Standards are also employed here to define
stiffness limits beyond which catastrophic fail- NDTI procedures and acceptance criteria.
ure occurs) at various temperatures and In summary, engineering test standards and
environments. Material testing to standards material and process specifications are
enables one to quantify material performance employed extensively in composites research,
development and production. Appendix B
Handbook of Composites. Edited by S.T. Peters. Published identifies specific material specifications and
in 1998 by Chapman & Hall, London. ISBN 041254020 7 test standards for polymer composites and the
1060 Specifications and standards for polymer composites

Table B.l Typical composite material qualification tests

Physical Mechanical Chemical

Resin content Tensile strength and modulus Infrared spectroscopy (IR)
Resin areal weight Compression strength and Liquid chromatography (HPLC)
Resin flow modulus Hydraulic fluid/solvent
Glass transition temperature Short beam shear (SBS) resistance
(T) ±45 In-plane shear Fuel GP-4) resistance
RhJological dynamic Open-hole tension
spectroscopy (RDS) Open-hole compression Thermal
Gel time Compression after impact
Volatiles content (CAl) Thermogravimetric analysis
Fiber density Flexural strength and modulus (TGA)
Fiber mass per unit length Fatigue strength Thermomechanical analysis
Fiber content Creep (TMA)
Prep reg tack Dynamic mechanical analysis Differential scanning
Laminate ply thickness (DMA) calorimetry (DSC)
Laminate void content Instrumented impact Thermal oxidative resistance
Laminate density (toughened resins) Thermal expansion (CTE)
Laminate fiber volume Fracture toughness (toughened Thermal cycling
Laminate flammability resins) Thermal conductivity
Solvent sensitivity compression
strength,SBS Electrical
Bolt bearing Dielectric constant
Dielectric strength
Dissipation factor
Volume resistivity
Dielectrometry
Some of these tests are specific to resin, fiber reinforcement, prepreg or laminate. There is no established universal set of
qualification test procedures which is widely adopted. See MIL-HDBK-17 for recommended guidelines.

organizations that develop them. It is limited several military specifications and standards
to consensus-type standards and does not for polymer composite materials. Other pri-
include private sector specifications which are vate sector organizations, such as the
generally not available. Aerospace Industries Association (AlA) and
Composite Materials Characterization (CMC),
Inc., are involved in the standardization of
B.2 STANDARDS-DEVELOPING
composite materials and tests in order to reap
ORGANIZATIONS
long-term economic savings.
The two principal organizations which Japan, Germany, France and UK are also
develop test methods for composites in the major players in composites technology. The
USA are the American Society for Testing and European Association of Aerospace
Materials (ASTM) and the Suppliers of Manufacturers (known as AECMA in Europe)
Advanced Composite Materials Association produces European Norm (EN) standards for
(SACMA). The Society of Automotive aerospace composites. Japanese Industrial
Engineers (SAE) Polymeric Materials Standards GIS) serve as a basis for standard-
Committee is the organization which pub- ization of composites in Japan. Germany has
lishes Aerospace Material Specifications issued DIN standards for composite materials.
(AMS) for advanced polymer composites. The France has AFNOR standards and the UK has
Department of Defense (DoD) has also issued British Standards.
 Standards used in USA 1061

Engineering standards for polymer com- standards for determining the physical and
posites also promote international commerce. mechanical properties of polymer composites.
In the global arena, the International AS1M standards are developed by a consensus
Standardization Organization (ISO) is the process and are widely used. The ASTM
body which develops international standards. Committee D20 on Plastics also has developed
The ISO fTC 61 Subcommittee (SC) 13 on standards which may be used for testing plastic
Composites and Reinforcements Fibers is cur- resins and reinforced plastics.
rently coordinating new standards for glass
and carbon fiber composites. USA participa-
B.3.2 SACMA METHODS
tion with ISO not only promotes international
commerce but also enhances USA global com- The SACMA has developed recommended test
petitiveness in the composites industry. methods for determining the physical, mechan-
ical, and chemical properties of composite
materials (Table B.3). Although SACMA is not a
B.3 STANDARDS USED IN USA
standards-setting body, it works actively with
AS1M, SAE, ISO, DoD, AlA and others towards
B.3.1 ASlM STANDARDS
standardization of composite test methods.
The AS1M Committee D30 on High-Modulus SACMA does not regard its SRMs to be 'stan-
Fibers and their Composites develops standard dards' in the truest sense because they have not
test methods for advanced polymer compos- been developed by a consensus process which
ites. Table B.2 lists the principal AS1M is typical of standard-developing bodies.
Table B.2 ASlM Standards for advanced polymer composites

ASlM C613 Resin Content of Carbon and Graphite Prepregs by Solvent Extraction
ASTM 0695 Compressive Properties of Rigid Plastics
ASTM D790 Flexural Properties of Unreinforced and Reinforced Plastics
ASTM D2290 Plastics, Ring or Tubular, Apparent Tensile Strength of, By Split Disk Method
ASTM 02344 Apparent Interlaminar Shear Strength of Parallel Fiber Composites by Short-Beam Method
ASTM 02734 Void Content of Reinforced Plastics
ASTM 03039 Tensile Properties of Polymer Matrix Composite Materials
ASTM 03171 Fiber Content of Resin-Matrix Composites by Matrix Digestion
ASTM 03379 Tensile Strength and Young's Modulus for High-Modulus Single-Filament Materials
ASTM 03410 Compressive Properties of Polymer Matrix Composite Materials with Unsupported Gage Section by Shear
Loading
ASTM 03479 Tension-Tension Fatigue of Oriented Fiber, Resin Matrix Composites
ASTM 03518 In-Plane Shear Response of Polymer Matrix Composite Materials by Tensile Test of a +/-45 Degree
Laminate
ASTM 03529 Resin Solids Content of Epoxy-Matrix Prepreg by Matrix Dissolution
ASTM 03530 Volatiles Content of Epoxy Matrix Prepreg
ASTM 03531 Resin Flow of Carbon Fiber-Epoxy Prepreg
ASTM 03532 Gel Tune of Carbon Fiber-Epoxy Prepreg
ASTM 03544 Reporting Test Methods and Results on High Modulus Fibers
ASTM 03800 Density of High-Modulus Fibers
ASTM 03878 Standard Terminology of High-Modulus Reinforcing Fibers and their Composites
ASTM 04018 Properties of Continuous Filament Carbon and Graphite Tows
ASTM 04102 Thermal Oxidative Resistance of Carbon Fibers
ASTM D4255 In-plane Shear Properties of Composite Laminates
ASTM 05229 Moisture Absorption Properties and Equilibrium Conditioning of Polymer Matrix Composite Materials
ASTM 05300 Measurement of Resin Content and other Related Properties of Polymer Matrix Thermoset Prepreg by
Combined Mechanical and Ultrasonic Methods
ASTM 05467 Compressive Properties of Unidirectional Polymer Matrix Composites Using a Sandwich Beam
ASTM 05528 Mode I Interlaminar Fracture Toughness of Unidirectional Fiber-Reinforced Polymer Matrix Composites
1062 Specifications and standards for polymer composites

Table B.3 SACMA Recommended methods (SRM)

SRMl Compressive Properties of Oriented Fiber-Resin Composites

SRM2 Compression After Impact Properties of Oriented Fiber-Resin Composites
SRM3 Open-Hole Compression Properties of Oriented Fiber-Resin Composites
SRM4 Tensile Properties of Oriented Fiber-Resin Composites
SRM5 Open-Hole Tensile Properties of Oriented Fiber-Resin Composites
SRM6 Compressive Properties of Oriented Cross-Plied Fiber-Resin Composites
SRM7 In-plane Shear Stress-Strain Properties of Oriented Fiber-Resin Composites
SRM8 Short Beam Shear Strength of Oriented Fiber-Resin Composites
SRM9 Tensile Properties of Oriented Cross-Plied Fiber-Resin Composites
SRMlO Fiber Volume, Percent Resin Volume and Calculated Average Cured Ply Thickness of Plied Laminates
SRMll Environmental Conditioning of Composite Test Laminates
SRM12 Lot Acceptance of Carbon Fibers
SRPl Printing and Applying Bar Code Labels
SRM13 Determination of Mass Per Unit Length of Carbon Fibers
SRM14 Determination of Sizing Content on Carbon Fibers
SRM15 Determination of Density of Carbon Fibers
SRM16 Tow Tensile Testing of Carbon Fibers
SRM17 Determination of Twist in Carbon Fibers
SRM18 Glass Transition Temperature Determination by DMA of Oriented Fiber-Resin Composites
SRM19 Viscosity Characteristics of Matrix Resins
SRM20 High Performance Uquid Chromatography of Thermoset Resins
SRM21 Fluid Resistance Evaluation of Composite Materials
SRM22 Determining the Resin Flow of Preimpregnated B-Staged Materials
SRM23 Determination of Resin Content and Fiber Areal Weight of Thermoset Prepreg with Destructive Techniques
SRM24 Determination of Resin Content, Fiber Areal Weight and Flow of Thermoset Prepreg by Combined
Mechanical and Ultrasonic Methods
SRM25 Heat of Reaction, Onset Temperature and Peak Temperature for Composite System Resins Using Differential
Scanning Calorimetry (DSC)
SRM26 Fiber/Matrix Adhesion of Carbon Fiber Reinforced Polymer Matrix Composites

However, SACMA recommended methods Table B.5lists military specifications for various
(SRMs) are being used as standards by various fiber reinforcements and composite materials.
organizations. The Military Handbook 17 effort is the most
widely recognized 000 standardization pro-
ject. The purpose of this handbook is to provide
B.3.3 AEROSPACE MATERIAL SPECIFICATIONS
a standard source of statistically based mechan-
The Society of Automotive Engineers (SAE) is ical property data for current and emerging
the primary organization in the USA which composite materials. The MIL-HDBK-17 gov-
issues material specifications for polymer ernment/industry coordination group meets
composites. Table B.4 lists some typical twice annually to develop this handbook.
Aerospace Material Specifications for polymer
composites. In general, the SAE has published
B.3.5 NASA STANDARDS
specifications for carbon, aramid, glass and
boron fiber composites. The National Aeronautics and Space
Administration (NASA) has developed five
standard tests and a material specification for
B.3.4 MILITARY SPECIFICATIONS
carbon (graphite) composites. Table B.6 lists
The Department of Defense (DoD) has issued the specific tests and the NASAl aircraft
several military specifications for polymer com- industry specification for toughened epoxy
posite materials used in aircraft applications. composite materials.
 Standards used in USA 1063

Table B4 Aerospace material specifications (AMS) for polymer composites

Carbon/Epoxy Composites
AMS 3892B Fibers, Carbon Tow and Yam, for Structural Composites
AMS 3894E Carbon Fiber Tape and Sheet, Epoxy Resin Impregnated
AMS 3895B Broadgoods and Tape, Multi-Ply Carbon Fiber /Epoxy, Resin Impregnated, Uniform Fiber

Aramid/Epoxy Composites
AMS 3901B Organic Fiber (Para-Aramid), Yam and Roving, High Modulus
AMS 3902C Cloth, Organic Fiber (Para-Aramid), High Modulus for Structural Composites
AMS 3903A Cloth, Organic Fiber (Para-Aramid), High Modulus, Epoxy Resin Impregnated

Glass/Epoxy Composites
AMS 3821B Cloth, Type 'E' Glass, 'B' Stage Epoxy-Resin-Impregnated, 7781 Style Fabric, Flame
Resistant
AMS 3828C Glass Roving, Epoxy-Resin-Impregnated, Type 'E' Glass
AMS 3831A Cloth, Type 'E' Glass, 'B' Stage Epoxy Resin Impregnated, 7781 Style Fabric, Flame
Resistant, Improved Strength
AMS 3832C Roving, Type '5-2' Glass, Epoxy Resin Impregnated
AMS 3906B Glass Tape and Flat Sheet, Non-Woven Cloth, Epoxy Resin Impregnated, For Hand and
Machine Layup

Boron/Epoxy Composites
AMS 3865C Filaments, Boron, Tungsten Substrate, Continuous
AMS 3867B Boron Filament Tape, Epoxy-Resin-Impregnated

B.3.6 FAA ADVISORY CIRCULARS (AC) Division is the Materials and Structures
Committee (AIA/MSC) which is responsible
The Federal Aviation Administration (FAA)
for the coordination and review of proposed
has issued Advisory Circulars (AC) to assist
specification requirements for materials,
commercial aircraft manufacturers in demon-
processes and structures. To promote stan-
strating compliance with the requirements of
dardization AlA has initiated Project 340-1
the Federal Aviation Regulations in the design
Standardization of Advanced Composite
and manufacture of composite material struc-
Materials and has issued two National
tures. Table B.7 lists two Advisory Circulars
Aerospace Standards (NAS) for composites
issued by the FAA. As of this writing, AC
manufacturing which are listed in Table B.8.
145-6 is a draft document undergoing coordi-
nation. AC 145-6 addresses requirements for
composite repairs, including materials, B.3.8 COMPOSITE MATERIALS
processes, and quality control tests. CHARACTERIZATION, INC. (CMC)
Composite Materials Characterization, Inc.
(CMC) is a joint enterprise funded by users of
B.3.7 AEROSPACE INDUSTRIES ASSOCIATION
advanced aerospace composite materials. CMC
(AIA)
was formed as a result of an Aerospace
The Aerospace Industries Association (AlA) Industries Association (AlA) initiative to pro-
plays a lead role in composites standardiza- mote industry research collaboration. CMC is a
tion. Within the AlA Engineering Standards Delaware corporation chartered in 1987 to
1064 Specifications and standards for polymer composites

Table B.S Military specifications and standards for polymer composite materials

MIL-Y-1140H Yarn, Cord, Sleeving, Cloth, and Tape-Glass

MIL-C-80730 Core Material, Plastic Honeycomb, Laminated Glass Fabric Base, For Aircraft Structural and
Electronic Applications
MIL-S-9041 B Sandwich Construction, Plastic Resin, Glass Fabric Base, Laminated Facings and Honeycomb
Core for J\ircraft Structural and Electronic J\pplications
MIL-C-9084C Cloth, Glass, Finished, For Resin Laminates
MIL-P-9400C Plastic Laminate and Sandwich Construction Parts and J\ssembly, J\ircraft Structural, Process
Specification Requirements
MIL-P-175490 Plastic Laminates, Fibrous Glass Reinforced, Marine Structural
MIL-C-19663D Cloth, Woven Roving, For Plastic Laminates
11IL-T-29586 (J\S) Thermosetting Polymer Composite, Unidirectional Carbon Fiber Reinforced Prepreg Tape
(Widths Up to 60 Inches), General Specification For
MIL-M-43248C Mat, Reinforcing Glass Fiber
MIL-J\-46103D J\rmor, Lightweight, Ceramic-Faced Composite
MIL-J\-46165 (MR) J\rmor, Woven Glass Roving Fabrics
MIL-P-46166 (MR) Plastic Laminates, Glass Reinforced (For Use in J\rmor Composites)
MIL-P-46169J\ Plastic, Sheet Molding Compound, Polyester, Glass Fiber Reinforced (For General Purpose
Applications)
MIL-P-46187 Prepreg, Unidirectional Tape, Carbon (Graphite) Fiber Polyimide (PMR-15) Resin
Impregnated, 316 C (600 F)
MIL-P-46190 Prepreg, Woven Fabric, Carbon Fiber Bismaleimide (BMI) Resin Impregnated
MIL-PRF-46197J\ Laminate: High-Strength Glass, Fabric-Reinforced, Polyester Resin Preimpregnated
MIL-R-60346C Roving, Glass, Fibrous (For Prepreg Tape & Roving, Filament Wmding & Pultrusion
J\pplications)
11IL-J\-62473B J\rmor: J\luminum-J\rarnid, Laminate Composite
MIL-F-64156 Fabric, Carbon (Graphite) Fiber, Nickel-Coated
MIL-Y-8337M (J\F) Yarn, Roving, and Cloth, High Modulus, Organic Fiber
MIL-Y-83371 (J\F) Yarns, Graphite, High Modulus, Continuous Filament
MIL-G-83410 (J\F) Graphite Fiber Resin Impregnated Tape and Sheet, For Hand Layup
MIL-R-8712M (J\F) Rods, Pultruded, Graphite Fiber Reinforced, Processing of
MIL-F-87121J\ (J\F) Fabric, Graphite Fiber
MIL-Y-87125A (AF) Graphite, 1000/3000 Filaments
MIL-STO-368 HPLC of PMR-15 Polyimide Resin and Prepregs
MIL-STO-373 Transverse Tensile Properties of Unidirectional Fiber/Resin Composite Cylinders
MIL-STO-374 Transverse Compressive Properties of Unidirectional Fiber /Resin Composite Cylinders
MIL-STO-375 In-Plane Shear Properties of Unidirectional Fiber/Resin Composite Cylinders
*MIL-STO-2031 (SH) Fire and Toxicity Test Methods and Qualification Procedure for Composite Material Systems
Used in Hull, Machinery, and Structural J\pplications Inside Naval Submarines
• Military Standard

Table B.6 NASA Standards for composites

NASA RP 1092: Standards Tests for Toughened Resin Composites, July 1983
NASA RP 1092 defines five standard tests (STs) for graphite/epoxy composite laminates:
ST-l: Compression after impact
ST-2: Edge delamination
ST-3: Open-hole tension
ST-4: Open-hole compression
ST-5: Hinged double cantilever beam
NASA RP 1142: NASA/ Aircraft Industry Standard Specification for Graphite Fiber Toughened
Thermoset Resin Composite Material, June 1985
 Non-US standards in use 1065

Table B.7 FAA Advisory circulars for composites compressive, and shear loadings at selective
hygrothermal test conditions. CMC also works
AC 20-10A Composite Aircraft Structures with other national organizations to promote
AC 21-26 Quality Control for the Manufacture of composites standardization.
Composite Structures
AC 145-6 Repair Stations for Composite and
Bonded Aircraft Structure B.4 NON-US STANDARDS IN USE

Table B.8 National aerospace standards for B.4.1 AECMA STANDARDS

composites
The European Association of Aerospace
NAS 990 Composite Filament Tape Laying Manufacturers (known as AECMA in Europe)
Machine - Numerically Controlled is developing various standards for carbon
NAS 999 Non-Destructive Inspection of fiber composites for aerospace applications.
Advanced Composite Structures Table B.1O lists proposed European Norm (EN)
standards for determining the physical and
mechanical properties of carbon fibers and
conduct research and development on emerging their composites.
composite materials including characterization,
screening, fabrication, and inspection of materi-
als or structures made from the materials. B.4.2 JAPANESE INDUSTRIAL STANDARDS (JIS)
The primary focus of CMC is on screening Standardization efforts in Japan are managed
testing of emerging composite materials. The by the Standards Department under the
CMC effort generates a standardized database Ministry of International Trade and Industry
of consistent properties for advanced compos- (MITI). The Japanese Standards Association is
ite materials which can be shared between in charge of publishing Japanese Industrial
member companies (Table B.9). CMC subcon- Standards OIS). Table B.lllists Japanese stan-
tracts to third parties the material dards for carbon fibers and their composites.
procurement, test specimen fabrication, inspec-
tion, test, data analysis, and documentation of
results. All tests and data are conducted in B.4.3 GERMAN (DIN) STANDARDS
accordance with CMC-approved procedures to German aerospace specifications for carbon,
produce a consistent and standardized data- aramid, and glass fiber polymer composites
base necessary for comparative assessment of are listed in Table B.12. The DIN specifications
material properties. Physical and mechanical and standards are published by the German
properties testing is performed, including lam- Institute for Standardization, which is the
ina and laminate evaluations under tensile, 'Deutsches Institut fiir Normung' (DIN) in
German.

Table B.9 CMC Member companies B.4.4 INTERNATIONAL (ISO) STANDARDS

The International Organization for
Dow Chemical
General Electric Standardization (ISO) Technical Committee 61
Northrop Grumman on Plastics, Subcommittee 13 on Composites
Lockheed Martin and Reinforcement Fibers is the body which is
UTC/Sikorsky developing international standards for poly-
Loral Vought mer composites. The USA is represented at
Rohr Industries international meetings through the American
1066 Specifications and standards for polymer composites

Table B.l0 AECMA Standards for carbon fibers and their composites

AECMA prEN2557 Carbon Fibre Preimpregnates, Test Method for the Determination of Mass Per
Unit Area
AECMA prEN2558 Carbon Fibre Preimpregnates, Test Method for the Determination of the
Percentage of Volatile Matter
AECMA prEN2559 Carbon Fibre Preimpregnates, Test Method for the Determination of the Resin
and Fibre Content and the Mass of Fibre Per Unit Area
AECMA prEN2560 Carbon Fibre Preimpregnates, Test Method for the Determination of the Resin
Flow
AECMA prEN2561 Unidirectional Laminates Carbon-Thermosetting Resin Tensile Test Parallel to the
Fibre Direction
AECMA prEN2562 Unidirectional Laminates Carbon-Thermosetting Resin Flexural Test
AECMA prEN2563 Unidirectional Laminates Carbon-Thermosetting Resin Test Method,
Determination of Apparent Interlaminar Shear Strength
AECMA prEN2564 Carbon Fibre Laminates, Test Method for the Determination of the Fibre and
Resin Fractions and Porosity Content

Table B.ll Japanese standards for carbon fiber Table B.12 German specifications for polymer
composites composites

JIS R 7601 Testing Methods for Carbon Fibers DIN 29965 Aerospace; Carbon Fibres, High
JIS R 7602 Testing Methods for Carbon Woven Performance Carbon Fibre Filament
Fabrics Yarns, Technical Specification
JIS K 7071 Testing Methods for Prepreg, DIN 29971 Aerospace; Unidirectional Carbon
Carbon Fiber and Epoxy Resins Fibre-Epoxy Sheet and Tape Prepreg,
JIS K 7073 Testing Method for Tensile Technical Specification
Properties of Carbon Fibre DIN 65090 Aerospace; Textile Glass,
Reinforced Plastics Preimpregnated Filament Glass Cloth
JIS K 7074 Testing Methods for Flexural for E-Glass (Prepreg), Technical
Properties of Carbon Fibre Specification
Reinforced Plastics DIN 65426 Aerospace; Aromatic Polyamide
Part 1: (Aramid) - Preimpregnated Woven
Fabric, High-Modulus Filament Yarn
(Prepreg);Innnensions, Masses
DIN 65426 Aerospace; Aromatic Polyamide
National Standards Institute (ANSI). Table Part 2 (Aramid) - Woven Filament Fabric
B.131ists draft and published (ISO) standards Prepreg from High-Modulus Filament
for composites. Yam and Epoxy Resin, Technical
Specification

B.5 PROPRIETARY SPECIFICATIONS

This appendix does not include the many
composite material and process specifications stitute a substantial data base which cannot be
which have been developed and are used by included herein. Please note that this appendix
individual private sector companies (Le. addresses only polymer matrix composites.
primes, fabricators and suppliers). Most, if not Specifications and standards for other types of
all, of the companies which either produce or composite materials (MMC, CMC, and C/q
use composites have some proprietary specifi- are still in early stages of development and
cations for composite materials and processes. may be either subject to export control or
It is known that proprietary specifications con- proprietary.
 Proprietary specifications 1067

Table B13 International (ISO) standards for polymer composites

CD 1268 Fiber Reinforced Plastics - Test Plates Manufacturing Methods -Part 1 - General Conditions
CD 3341 Textile Glass - Yams - Determination of Breaking Force and Breaking Elongation
CD 3374 Textile Glass - Mats - Determination of Mass Per Unit Area
CD 4605 Reinforced Products - Woven Fabrics - Determination of Mass Per Unit Area
CD 14127 Composites - Determination of Resin, Fiber and Void Content for Composites Reinforced
with Carbon Fiber
CD 15024 Determination of Mode I Delamination Resistance of Unidirectional Fiber Reinforced
Polymer Laminate Using the Double Cantilever Beam Specimin
CD 15034 Plastics - Prepregs - Resin Flow
CD 15040 Plastics - Prepregs - Gel Time
CD 15310 Reinforced Plastics - Determination of In-Plane Shear Modulus by Plate Twist Method
DIS 3374 Reinforcement Products - Mats and Fabrics - Determination of Mass Per Unit Area
DIS 5025 Textile Glass - Woven Fabrics - Determination of Width and Length
DIS 14126 Fiber Reinforced Plastic Composites - Determination of Compressive Properties in the In-
Plane Direction
FDIS 1889 Reinforcement Yarns - Determination of Linear Density
FDIS 1890 Reinforcement Yarns - Determination of Twist
FDIS 3344 Reinforcement Product - Determination of Moisture Content
FDIS 12114 Fiber-reinforced Plastics - Thermosetting Moulding Compounds and Prepregs-
Determination of Cure Characteristics
FDIS 11667 Fiber-Reinforced Plastics - Moulding Compounds and Prepregs - Determination of Resin,
Reinforced Fiber and Mineral Filler Content - Dissolution Methods
FDIS 12115 Fiber-Reinforced Plastics - Thermosetting Moulding Compounds and Prepregs -
Determination of Flowability, Maturation and Shelf Life
ISO 2559 Textile Glass - Mats (made from Chopped or Continuous Strands) Basis for a Specification
ISO 3605 Textile Glass - Rovings - Determination of Compressive Properties of Rod Composites
ISO 8515 Textile Glass - Reinforced Plastics - Determination of Compressive Properties in the
Direction Parallel to the Plane of Lamination
ISO 10119 Carbon Fiber - Determination of Density
ISO 10120 Carbon Fiber - Determination of Linear Density

CD - Committee Draft
DIS - Draft International Standard
FDIS - Final Draft International Standard
1068 Specifications and standards for polymer composites

REFERENCES 11. Documentation from Mr. Cecil W. Schneider,

President of Composite Materials
1. National Advanced Composites Strategic Plan, Characterization, Inc., 28 October 1991.
National Center for Advanced Technologies, 12. ISO/TC61/SCI3/WGI4 Letter Correspondence
September 1991. from Mr. Junichi Matsui, Toray Industries
2. Traceski, Frank T., Specifications and Standards for (Composite Materials Research Laboratories;
Plastics and Composites, ASM International Japan), 5 August 1992.
(Materials Park, Ohio), August 1990. 13. ISO/TC61/SCI3 Report to ASTM 020.61, Gary
3. Test Standards and Engineering Databases for Williams, July 1997.
Advanced Composites, Draft Position Paper by
Aerospace Industries Association, January 1992.
4. Advanced Composites Standardization, "White ACKNOWLEDGEMENTS
Paper," Prepared by the Committee for
Standardization of Advanced Composite The following individuals reviewed and com-
Materials, 4 November 1992. mented on this paper and are acknowledged
5. Annual Book of ASTM Standards, Vol. 15.03, for their constructive comments: Mr. Jerome
American Society for Testing and Materials Persh (retired), formerly Office of the Director
(Philadelphia, PA), 1991.
6. SACMA Recommended Methods, Suppliers of of Defense Research and Engineering; Dr.
Advanced Composite Materials Association, Gary L. Hagnauer, US Army Research
(Arlington, VA), 1997. Laboratory (ARL), Materials Directorate; Mr.
7. SAE AMS Index, Aerospace Material Specifications, Gary Hansen, formerly Hercules Advanced
Society of Automotive Engineers (Warrendale, Materials & Systems Company; Mr. Jerome R.
PA), July 1996. Jaeb, Chief Engineer, Structures Technology,
8. Department of Defense Index of Specifications and Boeing Defense & Space Group; Mr. Samuel J.
Standards (DoDISS), 1 January 1997.
9. Military Handbook 17, Polymer Matrix Dastin (retired), formerly Director, Advanced
Composites, Draft Volumes 1, 2 & 3, US Materials, Grumman Aircraft Systems.
Department of Defense, 17 June 1991.
10. Communication with Mr. Joseph R. Soderquist,
Federal Aviation Administration (FAA),
4 September 1997.
INDEX

Page numbers in bold type refer to figures; those in italics refer to tables

A-glass 134 bending failures 644 damage to 656

Abrading, surface preparation 669 bending stresses 644 ductile response 651--{" 652, 657
Abrasion composite 630, 660, 662-3 ductility 627-8,629-30,655--{'
mechanical 668 deformable 633, 634 elastic-plastic response 649
particulate fillers 252 deformations 630 environmental conditions 630
resistance, PET 224, 233 dissimilar 635, 660 epoxies 275
Abrasive water jet, see AWJ ductility 630 failure in 628
Abrasives equal thickness 650 fillet forming 271
machining 606 failures 627-9,628,657 flow, sandwich 288
in slurry 607 fully tapered 646, 647 heavy liquids 275-6
types of 607 identical 639 high temperature 658
ABS layered 646 joints 61Q--{,3
compressive strength 258 loading stress 642 bonded doubler 628
density 258 modulus 643 bonding techniques 611
maximum service temperature moisture absorbancy 630 double-lap 628
258 rigid 632,633,634 double-strap 628
shear modulus 259 stiffness 636 durability 631-2
shear strength 259 unbalance 629,648 inspection methods 627
tensile strength 258 stretching elongations 633 manufacturing deficiencies
thermal conductivity 259 tapering 627,629,644-51,645, 611
Accelerators 59--{'1, 60-1 647, 648, 649, 650, 653, 656, poor bonding 611
anhydride-cured system 72 658 scarf 628, 629
Acceptance values 724 thermal expansion 630, 661 single-strap 628
Accumulation, damage 797 thickness 627-9,628,628,656-7, single-lap 628
Acids, Lewis 60-1 658 stepped-lap 628,629
Acoustic transverse shear 630 stresses in 632
emission 849 weaker matrix 630 surface preparation 611
honeycomb assemblies 849 Adhesion tapered single-lap 628
fatigue testing 811 aramids 26 tapered-strap 628
pressure vessel 472, 472 failure 674-5 unsupported single-lap 628
Acoustical properties, fiberglass 135 Adhesives 255,271,274-6,517-18 light liquids 275-6
Acousto-ultrasonics 849 adaptability 274 mechanical 518
Acrylic acid, and polypropylene 251 aerostructural 651, 652 nylon 670
ACTS, reflectors 1017, 1017 blisters 271 pastes 275-6
Acute toxicity 823-4, 824 bond line control 274 peel
measuring 824 bondWng 120255,727 strengths 274-5
Adaptability, adhesives 274 aluminum alloy 683 stresses 628
Additives pressure 271 phenolics 255
low profile 382 brittle 651 blended 275
pultrusion 515-18 cell-edge 276 Plycosite 255
wetting agents 515-16 co-curing 868, 870 putties 275--{'
Adducts, thermoplastic polymers cohesive failure 674-5 Redux 255
526 core 271 repairs 869
Adherends cure temperature 661 reticulating films 276
axial stress 644 Cydeweld 255 scrim-supported 873
1070 Index

self-adhesive skins 276 Agglomerates 313, 314 thermal cycling 818

shear stresses 632-7, 633, 634, Aging through-the-face, sheet
635,636 air 85,90 evacuation 865
strain energy 629-30, 653 aircraft 860 water evacuation 863
supported films 276 thermoplastics 127 water ingestion 859
syntactic foams 275-6 Air wet lay-up repairs 865-7
test methods 630 aging 85,90 Aircraft Energy Efficiency (ACEE)
thixotropic liquids 275-6 3F /36F polyimides 85 1035
toughness 274-5 AFR700B/S2Iaminates 94,95 Alignment, graphitization 191
unsupported films 276 AvimidN 85 Aliphatic
urea-formaldehyde 255 Skybond 85 amine-cured (TETA) 65
urethanes 275 release agents 516 amines, health effects 832, 833
water absorption 668 Airbus A320, composite applications content 75
Adjusting, lamina properties 763 1036,1038 polyether triamine (APTA),
Advanced Communications Aircraft viscosity 54
Technology Satellite, see ACTS aging 860 system 70
Advanced composites aluminum core removal 862 Alkaline
applications 23 applications 1022-42 cleaning, metal surfaces 871
military 23 certification requirements 1022-3 degradation 1000
Advanced fibers, repairs 877 components 1029-30, 1029 All purpose vehicles 908, 908
Advantages current prices 1024, 1026 Alleo 59
AWJ 604 damaged protective coatings Alternative, tooling 449-50, 449, 450
correlators 834 859~ Altex 309, 309
hand lay-up 352 design 878 Alumina 249-50
injection pultrusion 497 process 1024-5, 1026 composites 325
machining 596 requirements 713, 1022-3 grit, light grit-blasting 675, 676
laser 605 effect of contaminants 863 silicate
ultrasonic 608 environmental exposure 860-1 boria modified 309,310
preform 438 face sheet materials removal 864 strength 309
pull winding 497 face sheets removal 862 Young's modulus 309
RTM 434-5 fixed wing, current prices 1024, trihydrate 245
tapering 646 1026 Aluminosilicate, staple 160
turning 597 fluid contamination 859 Aluminum
twin-screw extruders 539 flushing contaminants 864 alloy
vertical pultrusion 495 honeycomb core removal 862 adhesive bonding 683
AECMA standards 1065, 1066 impact damage 858-9 linear thermal expansion 705
Aerospace interim repairs 859 core removal 862
applications, materials selection leak paths 859-60 CTE 558,589
1009 maintainability 1024 density 170, 558
design 709 maintenance 878 elastic modulus 170
goal 712, 713 metal face sheets removal 862 fiberglass cloth repairs 877
equipment 1004-21 metal repairs 857 foil 820
Industries Association (AIA) minor damage 859 fracture elongation 170
1063 moisture barrier removal 863-4 oxide
industry non-metallic core removal 862 fibers 27
composite tools 593 open core evacuation 865 particulate fillers 250
facing material 255 production 840-1, 841 rivets 668
prepregs 885 costs 1024 specific
material specifications 1062, 1063 reinforced plastics 1023 modulus 1049
Aerostructural, adhesives 651, 652 repairs 857-8,857-80,1024 strength 1049
AFR7ooB, chemistry of 81 planning 861 surface preparation 871, 876
AFR700B/S2Iaminates skin damage removal 861, 861 tensile strength 170
air aging 94, 95 skin penetration 858, 859 tension-tension ratio 1049
compression strength 93 speed tape repairs 859 thermal conductivity 558
flex strength 93, 95 stabilizer America's Cup yachts 922
shear strength 93 design requirements 710 Amine curing 49-50,51
tensile strength 93 preliminary materials 711 epoxy resins, aliphatic 70
weight loss 94 structural applications 1027-8, Amines, aliphatic 53, 54
AFRP 605, 1024 1027,1028 Amino resins, health effects 834
 Index 1071

Aminolysis, degradation 888 biomedical, thermoplastics 127--8 uranium enrichment centrifuges

Amoco 58 boron 163,165,166 953-4
anhydride curing agents 56 bridges 933-4, 933 valves 954
Analysis business equipment 934, 934 wind turbine blades 954
composites 736-57 cable cars 934-5 Aramid 23, 26, 425
cost of 736 carbon fibers 196-7 adhesion 26
equilibrium equations 737 casting 38 composites 210, 210
Ancadride MHHPA (Ciba-Geigy) 57 ceramic fibers 163,165,166 compression 26
Anchorage 987 CMCs 935 costs 717
grounded 986 consumer products 946 density 26
prestressed concrete 985-6 cooling towers 936-7 fibers 202-22
Angle of impact, meteoroids 815 drive shafts 936-7 applications 216,217-20,
Angle ply 741 electrical 937, 937 222-3
Anhydride electronics 937, 937 athletic shoes 222
curing 50 firefighter breathing apparatus attributes 358
agents 56-9,56-9 937--8 availability 216,217
health effects 832, 833 flywheel mechanical battery by-products 222
high temperatures 50 systems 938, 938 chemical properties 213-15,
Anhydride-cured system 72 fuel tanks 939-40 214
accelerator 72 glulam beams 940-1 choice of resin 222
cure cycle 72 grating 936 composites 207
curing agent 72 health 943-4, 944 compressive properties 221
heat distortion temperature 72 high silica 163, 165, 166 costs 223
maximum stress 72 high-pressure tubing 941 creep resistance 216, 222
modulus of elasticity 72 high-speed train brakes 941 cutting 222
pot life 72 hydrogen fuel storage 941 definition 202
solvent absorption 72 industrial equipment 951 design of 216, 221-2
tensile properties 72 ladders 936 electrical insulators 215
viscosity 72 laptop computers 942 electrical properties 215
Anisotropic lighting poles 948-9, 948 environmental properties
bodies, thermoelastic behavior loom components 942 213-15,214
470-1 maglev train guideways 942-3 equilibrium moisture content
composites 797 marine 916-28 215,241
materials 687,688 medical 943-4, 944 failure 210
material constants 688 metal surfaces 871 fatigue resistance 221, 222
pitches, stabilization 178 MMCs 944-5 flame resistance 205,207
Anisotropicity, PAN 185 mobile storage 945 forms 216
Anisotropy molding compounds 945-6 handling 222
crystal 334, 335 musical instruments 946 health effects 835
GFRP 1006 non-aerospace 935 length-to-diameter ratio 204
PET 224 O-rings 953 limitations 207
Annealing, and coating thickness oil and gas 946-7, 947 machining 222
297 POO fiber 238 manufacture 203-4
Anodizing 667 PET 230 mechanical properties 207-13,
aluminum alloys 668 pick-up trucks 936-7 208-9, 209, 210, 211, 212
low voltage 871 pipeline rehabilitation 947--8 melting 205
non-tank 871 piping 936 point 204
Antennas 932-3 PMR-1596 microbuckling 207
Antimony oxide 244 ports and harbors 949 moisture absorption 221
high cost 244 power poles 948-9,948 opaque 221
synergism 244 quartz 163, 165, 166 optical properties 215
Antistatic agents, fiberglass 146-7 railways 950 physical properties 205-7, 206
Appliances, reuse 901 rolling stock 949-50 pricing 216
Applications 3,270,931-56 rolls and air shafts 950-1, 951 properties 205-15
aircraft 11 safety 943-4, 944 reinforced plastic, see AFRP
antennas 932-3 seals 953 safety 222
aramid fibers 216,217-20,222-3 SMEs 952 sandwich construction 222
AS-4/PEKK 127--8 structural 709 self-screening 215
battery casings 941 tunnels 950 solvents 204
1072 Index

specific ligaments, PET 234 Kevlar 217-18

stiffness 207 AS-4 carbon tiber 123 PET 231
strengths 207 in-plane shear modulus 124 Spectra 231
strength 205 tensile modulus 123 SVM 220
to weight ratio 223 tensile strength 123 Technora 220
structure 204-5 AS-4/PEEK(APC) Tekmilon 231
temperature resistance 205 observed life 806, 807 Twaron 219-20
tensile predicted life 806,807 Vectran 236
modulus 715, 716 residual strength 806, 807 Aviation, price of 1024
stiffness 209 strength 125 Avimid N, air aging 85
strength 209 AS-4/PEKK, applications 127-8 AWJ
thermal Asbestos, health hazards 252 advantages 604
properties 205-7,206 Assembly bowing, GFRP 1014,1014 disadvantages 604
resistance 205 ASTM drilling 604
thermoplastic matrices 222 methods, cured epoxy resin linear cutting 603,603
toughness 221 systems 72 machining 600, 602
treatments 215-16 numbers 66, 66 milling 604
twisted 215-16 standards 1061, 1061 process 602
UV absorption 215, 221 Atmospheric electricity 860 traverse rates 603
water absorption 215 Atomic oxygen 813--14,814 turning with 603--4
weave 222 destructiveness 813 without abrasives 600
weight 222 erosion 814 AXAF orbiting observatory 970
mechanical properties 718 fluence 813 Axial
moisture absorption 26 reaction efficiency 814 compressive properties, Spectra
paper honeycomb 269-70 Atomite 247, 247 226,226
applications 270 ATOTech 61 loading 745--7
price 716 ATR72 orientation, PET 224
strength retention 230 CFRP wing box 1038 stress 613
strengths 718 composite applications 1038, adherends 644
tensile strength 26 1937 tensile properties, Spectra 226,
Aramid/ epoxy, C1E 589 Autoclave 22, 116,461 226
Aramid/fiberglass 359 consolidation 298,577 Axisymmetric deformation
Aramid/graphite 359 cure cycle 84 circular plates 752-3
hybrids 354 definition 577 cylindrical shells 753-4
Arc jet tests, carbon-carbon repairs 868 disks 752-3
composites 347 tooling 589-91, 590 AZ epoxy N (AZS Corp) 63
Archimedean screw 538 Autoclave/oven, molding 361
Architecture, microstructural 795, Automated Real-time Inspection Bag molding 352-77, 361, 362,
796 System (ARIS) 842 366-72,368,369,370,371
Areas of transition 731-2,732 Automated Tool Manufacture for thermal cures 367
Armos Composite Structures 566 Bagging 577
availability 220 Automation, cost savings 16 Balance 9,10-11
elongation at break 208 Automobiles, reuse 900-1 Balanced laminates 693
initial tensile modulus 208 Automotive Balls, tooling 565
sources of information 221 body applications 909--10, 913 Barcol hardness
specific gravity 208 SMC 909 recycled PET 894
tensile strength 208 industry testing 38
Aromatic design requirements 713 Barium ferrite, IDlers 527
amine-cured MPDA 65 recycling 901-2 Barium sulfate 249
amines, health effects 832, 833 surface preparation 669 fillers 527
polyamide fiber, see aramid fibers interiors 912, 913 Barrier
polyester fibers 235--7,235,235 market, GMT 127 creams 829-30,830
system 71 suspension 911 layers, silicon carbide 292
Aromatization, and carbonization Availability materials 863-4
183 aramid fibers 216, 217 Bars
Articulation components 962-3 Armos 220 bending 785
Artificial Dyneema 231 testing 784
blood vessels 963 glass fibers 24 Bases, Lewis 60--1
joints 810 Hoechst Celanese 231 Basic structural units, see BSUs
 Index 1073

Basketvveave 140,145 Dlelting point 60 cOffiDlercial 380

Bathtub manufacture, Japan 889 Besfight 1054 cODlpression molding 492
Battery casings 941 properties of 170 definition 379
BeaDls 745-7,746,995-7,996 BF3MEA 59 fatigue properties 386, 387
bending stiffness 997 Biaxial flexural Dlodulus 386, 386
curved 791 fabrics, non-vvoven 504 flexural strength 386,386
deflection 995,996 plain-vveave 399 injection Dlolded 380
displaceDlents along 747 stretching 546 IZOD impact 386, 386
Dlodes of failure 997 Bicycles 1050 recycling 890-3
properties of 283 carbon fiber fraDle 1050 specific gravity expansion 386,
shear deforDlation 748-9,995 fraDle tubes 1051 386
strength analysis 308-9 handlebars 1051 tensile Dlodulus 386, 386
stresses in 747 tube and lug designs 1050-1 tensile strength 386,386
thin-vvalled 747-9, 748 Binders therDlal coefficient 386, 386
types 283 fiberglass 147 BMI 30,99-114
sandvvich 281 organic 314 CAl perforDlance 105
Bearing solutions, availability 82 cODlpressive Dlodulus 111
failure 616, 729 Bioabsorbable cODlposites 961 dielectric properties 106-7
load 620-1,620 biocoDlpatibility 962 edge delaDlination 111
contact region 620-1 Biological attack 810-11 F-22 fighter 113, 113
edge distances 621 Biomaterials 957--64 filDl adhesives 114
joints 626 Bisaleimide resins 30 galvanic corrosion 108-9
strength 624-5, 625 see also BMIs hODlopolYDlers of 99
joints 618-19 BisphenolA hydrolysis 112
tests 625 fuDlarates 36,39 Dlatrices 106
stress 619, 619 hODlopolYDler property 102 DlechaniCal strength 111
at fasteners 620 structure precursor 102 Dlolding cODlpounds 114
Beech Starship 472,627,1022,1039, supplier 102 DlonODlers 100,103
1040 trade naDle 102 suppliers 112
Bell DlOUth 500-1 Bisphenol A-based, health effects therDlal properties 106,109
Bend radius 308 832 toughening, diepoxides 105
Bending 785,791-2 Bisphenol E BMI-DAB 100-1
bars 785 hODlopolYDler property 102 definition 105
deflections 637 structure precursor 102 BMI-MDA, definition 105
fasteners 626 supplier 102 Boat hulls 1045
single lap joint 641 trade naDle 102 Boeing 777
failures BisphenolM cODlposite structure 1035, 1036
adherends 644 hODlopolYDler property 102 Bolt
joints 617, 617 structure precursor 102 hole 621
fasteners 625, 625 supplier 102 tensions, torque levels 626
longitudinal 747 trade name 102 Bolted
nonlinear 747 Blast pressure 675--6, 676, 677 joints 728,728, 729, 1015
rings 785, 792 Bleed-out 367 edge distance 617, 617
skin fracture 751 Bleeder 577 Bolting and riveting 517
stiffness 745 perDleability 581 Bond
strength 791 ply, Dlolding 584-6, 585 defects 630-1
Kulon 305 Blistering failures 627-9,628
testing 446 CFRP 918 layer 656-8,657
VMN-4305 isophthalic polyesters 918 line control, adhesives 274
stresses Dlarine applications 918, 918 stresses 640
adherends 644 orthophthalic polyesters 918 lateral deflections 640
fOrDlulas 284 perDleation barrier 918 thickness 631
three-point 785, 791 Blisters, adhesives 271 to adherend, thickness 641
and tvvisting 3 Blovv van der Waals 224, 226
vvhole rings 785 Dlolding 529,532,549 Bonded
3,3',4,4'-Benzophenone- accuDlulator 549 doubler, adhesive joints 628
tetracarboxylic dianhydride polYDler properties 550 joints 728,728,1015
(BTDA), Dlelting point 59, 76 Blovvn core, NDE 854 cracking 659,659
BenzyldiDlethylaDline (BDMA), BMC 379-80,491-2 creep failure 658
1074 Index

design 728 interlacing patterns 415, 415 Butt tensile test 638, 638
durability 658-9 Maypole machines 413, 414 Butyl glycidyl ether (BGE), viscosity
environmental extremes 658 multidirectional 473 61
viscoelasticity 631 prepreg 418 By-products
step lap joints 627 processing technology 413-15 aramid fibers 222
Bonding rackets 1050 volatile 48, 78
adhesives 127, 255 sporting goods 1045-6 Bypassload 616,619,620
agent 190-1 structural geometry 415-16 definition 620
fiber and matrix 801-2 tape 418 joints 626
fusion 127 track and column 414, 414, 416 pure 620,621
hydrogen PPTA 204 tubular 413 ratio 620
inspection 668 Break-out 597
interfacial 149 Breakage 302,319 C-glass 134
and joints 374 Breaking strength C-scan
metals 668 fiberglass 139-40, 141-4 pulse-echo testing 842, 842
pressure 872-4 Kevlar 214 ultrasonic through-transmission
adhesives 271 Technora 214 testing 841, 841
skin/core 917-18 Breather 288 Cable cars 934-5
techniques, adhesive joints 611 Bremsstrahlung radiation 816 Cables 927
Bondline Bridges 933-4, 933 Calcium
adhesive voids, NDE 854 deck 986,986 aluminoborosilicate 24
thickness control 872-3 enclosures 992-3, 994 carbonate 246--7, 247
Bone cement 962 rehabilitation 988 abundance 243
Boron 23 Brittle fillers 527
applications 163,165,166 adhesives 651 low shrinkage 243
continuous fibers 163 composites 797 Mohs hardness scale 243
fibers 27 fiber breakage 291 particulate fillers 243, 247
reinforced plastics (BFRP) materials, ultrasonic machining stiffness 243
1023 605 Calcium-sodium metaphosphate 957
forms 161-3 solids 188-9 Calendering 532
manufacture 156--7 Broken fibers 803 Carbide, coating 296
properties 161 buildup 803 Carbon
trifluoride (BF:J 59 Bromine, halogens 43 black 245
trifluoride-monoethylene amine BSUs 169 content 306
(BFaMEA), melting point Buckling 547,747,750-2,752 fiber 25-6,234,335-8,336,337
61 asymmetric 755 see also CF
Boron-aluminum, MMC 299 column-type 755 aluminum interaction 297
Boron-epoxy columns 997 applications 196--7
CTE 589 cylindrical shells 755-7 in carbon, matrix 337
stabilizer covers 1030-1 flange 998 compatibility of 295, 296
Bosses 392,392 local 997-8 competitive prices 25
mold making 390 pressure 756 composites 190-1
Boundary conditions 588, 750 sandwich 282 ISO standards 1067
laminates 689 Buffalo Color 58 typical compression
Bowing, GFRP 1013 Building strength 1055
Braiding 18, 33, 164, 402, 402, construction, recycling 901-2 cost of 196, 917
413-18,413,414,415,417,436, systems 989,991 CVD 335
437 Buildup, broken fibers 803 discontinuous 337-8
2-D 413 Bulk molding compound, see BMC elastic modulus 169
definition 415 Bumpers 815 electrical
2-step 416 Bundle theory 189-90 conductivity 169
3-D 413 Buoys and floats 927-8 properties 184-5,185,186
geometric parameters 416 Bushings ffiament directionality 335
angle, and fiber volume fraction fiberglass 138 from PAN 171, 335, 336
417 platinum 133 from pitch 171
cross section 417,417 tooling 562 health effects 835
definition 413 Business equipment heat treatment 335
dry tows 418 applications 934, 934 high modulus 335
horizontal 413,413 shielding 936 high strength 335
 Index 1075

honeycomb 270-1 melting point 333 Carving bits, honeycomb 289

laminate, tensile strength modulus of elasticity 333 Cast
1056 multiformity 333, 333 aluminwn, tooling 442
MPa 425 oxidation resistance 333 boron, aluminum cable 299
nanoporous 183 pistons 349 ceramic, CTE 589
noncircular 336 powders 343 lamina, composites 299
PAN-based 169 pyrolysis 341, 342 Casting, applications 38
attributes 358 reproducible strength levels 333 Catalysts 49,832
pitch-based 25, 353 stiffness 333 homopolymerization 49
attributes 358 strength efficiency 344 metal coordination 104
production 25 tensile strength 333 Catalytic curing 50-I, 51
pyrolysis 335 TEOS 349 Caul plates 14, 288-9
reinforced thermal design 591
composites 358-9 conductivity 333, 349 tooling design 590
plastics, see CFRP expansion 333, 349-50 Cavity pressure 443
reuse 884 gradients 333 CDS 798,798
rovings 502 oxidation 345, 347-8 CE 99-114,104
selection 90 thermophysical properties 348-9, adhesives 114
shape 335,336,337 349 composition 108
shear strength 295 Carbon-carbon part compression 111
tensile 2-D 340 dielectric properties 106-7
modulus 335 in oxidizing atmosphere 340 edge delamination 111
strength 169 Carbon-epoxy 1031 galvanic corrosion 108-9
thermal composites 27 hot-wet performance 108
conductivity 169 laminates 669 hydrolysis 112, 112
decomposition 25, 25 prepreg 558 of cyanurate linkages 107
properties 184-5,185,186 Carbon-fiberglass 359 mechanical strength 111
torsional strength 295 Carbon-graphite, quasi-isotropic moisture absorption 107
volwne fraction 335 laminates 13, 13 monomers 105
Young's modulus 295 Carbon-matrix composites 333 prepreg reinforced 113
fiber-epoxy, drilling 599 Carbonic, properties of 170 property of casting 108
fillers 527 Carbonization reinforcement 111
specific modulus 1049 and aromatization 183 suppliers 112
specific strength 1049 chemical changes 181-3 tensile strength 108
tension-tension ratio 1049 and effluent loss 183 thermal properties 106, 109
type T300 503 emissions during 181-2 Young's modulus 108
Carbon-aluminwn microstructural changes 183-4, Celion 1054
infiltration 303 184 compressive strength 87
layer reinforced 304 stabilized PAN 181 flexural strength 87
magnesium evaporation 303-4 Carbonized organic composites fracture toughness 87
MMC, tensile strength 300 339-41,340 shear strength 87
plate reinforced 304 curing temperatures 339-40 interlaminar 91
producing 303 impregnation 340 tensile
rolling 303 pyrolysis 340 modulus 87
soldering 303 starting material 339 strength 87
Carbon-carbon, rocket nozzles 712 Carbonyls, thermal decomposition weight loss 91, 93
Carbon-carbon composites 333--50 293-4,294 Cell
advanced 344 Carcinogenicity 824 configurations, honeycomb 262
applications 341, 349-50 Cards combs 495 size, honeycomb 263
arc jet tests 347 Carpet plots 718-20, 719 Cell-edge, adhesives 276
coating 343-4 construction 719 Celluloid acetate
cure cycle 342 laminate selection 720 compressive strength 258
density 333 lay-up sequence 719 density 258
high temperatures 340, 344 shear modulus 720, 721 maximum service temperature
impact damage 334 strength values 719 258
lay-up 342 tensile shear modulus 259
linear thermal expansion 333 modulus 719 shear strength 259
manufacturing 341-4 strength 720,720 tensile strength 258
mechanical properties 344-5,345 Cartilage 958 thermal conductivity 259
1076 Index

Cellulosic, fillers 527 PET 121 applications 1035-9, 1036, 1037,

Central cylinder, satellites 375-6 polyamide 6, 6 121 1038, 1039
Centred injection, ports 451, 451 polyamide 12 121 material weights 1027
Centreless polybutylene-terephthalate 121 Civil engineering, pultrusion 519-20
grinder 429 polyesters 121 Clamping 501-2
sander 429 polyolefin 121 levels 625
Ceramics 307-29 polypropylene 121 pressure 625
applications 328-9,328 Charge joints 618
brittle 361 pattern 384-5 loss of 626
continuous unidirectional 317-18 preparation 384-5 Day
cutting tool inserts 328, 328 Chassis application 910--11,913 grades of 244
densities 307 Chemical particle sizes 244
density 312 composition, fiberglass 147--8 particulate fillers 244
electrical conductivity 307 contamination 860 Clearances, fasteners 624, 626
fibers 27 degradation 299,888-9 Cleavage 50
applications 163,165, 166 name 121 asymmetric 50
continuous 163 oxidation, direct wet 191 thermolytic 81
forms 161-3 precipitation 292-3 Climb, milling 596
health effects 835 properties, aramid fibers 213-15, Close packing 403, 403
manufacture 156-7 214 Closure, of mold 442
properties 161 resistance 38 CMCs 935
fracture toughness 312 and crystallinity 120 ultrasonic machining 605
hardness 307 feldspar 243-4 CNC, lathes 597
matrices, designing 317-23, 318, fiberglass 134 CO2, gas laser 605
319,320,321 Kevlar 214 Co-poly-p-phenylene/3,4'-
matrix nepheline 243-4 oxydiphenylene
composites, see CMCs PBI 237 terephthalamide 203, 203, 205
lamina, fiber fracture 799 PET 224,230 Coarse fillers
materials 311-12 Technora 214 silica 516
melting thermoplastics 126 talc 516
point 312 stability 214 Coating
temperatures 307, 311 vapor deposition, see CVD aluminum foil 820
modulus of rupture 312 vapor infiltration, see CVI by precipitation 293, 294
Poisson's ratio 312 Chemically resistant, gloves 829 carbide 296
powder processing 312-14,313, Chimneys 983 carbon fibers 292-4,294,295
314 rehabilitation 988 chemical precipitation 292-3
reinforcing 307-11,308,308,309, Chlorendic 39 electroless deposition 292-3
310,311 anhydride (CA), melting point in gas phase 294--8
strengthening 323-4, 323 58 in-mold 385
thermal conductivity 307 resins 36, 36, 43 medical applications 957
thermal expansion 312 Chlorinated solvents, health effects metal, heating temperature 294
tolerance to flaws 307 836 nickel 292
unreinforced 328-9 Choice of resin, aramid fibers 222 nickel vapor-deposited 819-20
Young's modulus 307,312 Chopped strand, mat 155, 917 nitride 296
Certification requirements Chopped-fiber, reinforced polyurethane 813
aircraft 1022-3 composites 355 protective 819-20,819
military aircraft 1023 Chopper gun 353 silicon carbide 294
CF/PEEK 115 Chord, winding 462 thickness 293,294,297
CF/PPS 115 Chromic acid anodized, aluminum and annealing 297
CF/PSU 115 foil 820 and deposition time 294
CFRP 1023-4 Chronic toxicity 824 Kulon 306
blistering 918 testing 824 VMN-4306
cathodic 918 CIC 381 two-layer 300
prepregs 1-23 modified 381 Cockroft-Walton accelerator 848, 848
Chain folding, PET 224 Circuit boards, fiberglass 136 COD 124
Channeling 443 Circular plates, axisymmetric Coefficient of thermal expansion
Char yield, NR-I50 82 deformation 752-3 seealsoCTE
Characteristic damage state, see CDS Circumferential, reinforcement 461 aluminum 558
Characteristic temperatures Civil aircraft carbon-epoxy prepreg 558
 Index 1077

glass-epoxy 558 fiber dominated 795 hot-wet 108

high carbon cast steel 558 flywheels 474-5 measuring in 195
Invar 558 glass-reinforced 716,717 molding 22,43,374,384,529,
laminates 322 graphite-reinforced 716, 717 532,545
mahogany 558 health and safety 880 composite parts 549
monolithic graphite 558 history 353-4 fiber orientation 548
steel 558 hybrids 353 GMT 117
whiskers 308 inhomogeneous 797 one-dimensional flow 583-4,
Cohesive fracture 674, 675 inspection 880 584
Collision avoidance 484, 485 layer, permeability 581 polyimides 545
Color, stability 43 Materials Characterization, Inc PTFE 545
Columns 997-9,999 (CMC) 1065, 1965 and pultrusion 490
buckling 997, 999 panels 909 rackets 1049
crushing 997 preventive maintenance 880 thermoplastics 116,544-9
failure mechanisms 997 processes two-dimensional flow 584
properties 998,999 health hazards 831 UHMWPE 545
slenderness 997 safety hazards 831 molds 394,395,395
testing 999 processing 879 properties 360
Commercial preforms 397 strength
resins 101, 102, 103 times 353 AFR700B/S21aminates 93
stabilization 178 protective coatings 879 marine laminates 920
Comminuted polymers, fillers 527 pultruded, outdoor use 504 normalization 721
Compacted plies, Springer's model quality controls 880 strength after impact 125, see also
588 reasons for using 3 impact energy
Compaction 470 repair methods 880 and tension 786-9
pressure 457 repairs, aircraft 857-8 TEOS 346
Compliance 195 sealants 879 testing 781
Compliant ring method 788-9 sealing 302 buckling 787
Components standard tests 880 flat specimens 787
of composites 879 talc 246 rings 782, 788-9
design 709 testing 194-6 transverse cracking 787
sizing 712 thermal properties 661 unidirectional composites 787
Composites 166 Ti-Al-C 303 Compressive
adherends 630, 660 tools 566-75 fiber stress, and fiber volume
AI-B-C 304 aerospace industry 593 fraction 581
analysis 736-57 ease of preparation 592 modulus
anisotropic 797 low density 592 BMl 111
aramid prepreg 566,567-71 PMR-15 laminates 88
fibers 207 thermal expansion 592 properties
reinforced 716, 717 thermal stability 592 aramid fibers 221
beam wet lay-up 566 testing 38
cost 723 toxicological properties 831-7 strength
design example 722-3, 722, tubes ABS 258
723 curing 428-9 Celion 87
design values 723 fishing rods 427 cellulois acetate 258
bolting and riveting 516-17 unidirectional 191 epoxies 258
brittle 797 whisker-alumina 320-1 fiberglass epoxies 153
cast lamina 299 Composition honeycomb 264-5,266-7
components of 879 CE 108 Kulon 305
coupling agents 250 continuous ceramic fibers 309 laminate 616
damage fiberglass 138-9 Nomex 272, 273
assessment 880 fibers 158 PBO fiber 238
sources 838 whiskers 158 phenolics 258
definition 378 Compound structures 456 plied-yam 151
density, thermoplastics 534 Compression PMR-15 82,88
environmental effects 879-80 aerospace applications 26 polyurethane 258
fabrication 879 after impact (CAl) 105 polyvinyl chloride 258
failure 193 ararnids 26 recycled phenolic 894
analysis 880 CE 111 S-glass epoxies 152
1078 Index

single-yam 151 areas 184 splices 276,278,278

skinned molded foam 258 inverse method 611 Comers
thermoplastics 124 region, bearing load 620-1 flexure, TEOS 346
thermosets 124 Continuity condition, resin flow injection, ports 451,451
VMN-4305 582-3 preform 438
yam distribution 154 Continuous Corporate Average Fuel Economy
stress, surface 321 casting 29&-9 (CAFE) 907
Compton scattering 845 ceramic fibers 163,309 Correction method, tooling 590-1
Computed tomography 846-7, 848 dry jet wet spinning 204 Corrective Optics Space Telescope
advantages 846-7 fiber-reinforcement 22 Axial Replacement, see
density map 846-7 fibers 156, 338, 338 COSTAR
Computer codes, joints 624, 630 alumina-based 156-7 Correlation, ultrasonic 843--4, 844
Computer numerical controlled, see density 162 Correlators
CNC description 162 advantages 834
Concrete diameter 162 block diagram 844
dilation 988 elastic modulus 162 P1FE 834
fonns 925 filament Corrosion
linear thermal expansion 705 directionality 338 barrier 47
polymer 984, 984 winding 338 galvanic 10&-9
post-tensioning 986, 987 layer interlocking 338 resffitance 38,42,303
prestressed 985-6 manufacturers 162 construction 982,983
rehabilitation 988 multilayer locking 338 high-temperature 322
reinforcing 999-1000 reinforced composites 355 reinforced composites 387
Condensed core, NDE 854 specific strength 162 Corrugated configuration, reinforced
Conductivity, and modulus of thermal expansion coefficient composites 389
elasticity 186 162 Corrugation process, honeycomb
Consolidation 314,576-94 trade names 162 257
autoclave 298, 577 filament, fiber architecture 401-2 Cortical bones 958
definition 576 glass, rovings 492 COSTAR 101&-19,1019
equation 582 impregnated compound, see CIC Costs
one-dimensional 584 reinforcement 502 aramid 717
equipment selection 577 strand mat 436, 503 carbon fiber 917
fiber deformation 576 whisker 156 E-glass 717
history 578-86 Contour, variation 372-3 glass fibers 24
mechanisms involved in 576 Contoured tape, lay-up 16 graphite 717
mode~ 578-86,581-6 Control surfaces, marine meta-aramids 205
pressure 577 applications 924 metals 717
resin flow 576 Coolants, drilling 600 particulate fillers 242
techniques 576 Cooling towers 936-7, 993-4, 995 PBOfiber 238
temperature 577 Coordinating partial plies, tooling 52-glass 917
thick composites 585 design 590 Count
thin composites 585 Copolymerization 34 fiberglass 141-4
time 585 Core plied-yam 151
window 588 corrosion single-yam 151
tooling materia~ 577 adhesives 271 Countersinking, and drilling 59&-9
Constituent NDE854 Countersunk
materi~ 22 edge treatment 277 fasteners 624,624,625,625
properties 766-7 fatigue, NDE 854 head 729
Constitutive equations 739,741 materi~ 256-7 Coupling
stiffness coefficients 743-4 movement, adhesives 271 agents 671
Construction node disbonds, NDE 854 chemical functionability 147
applications 982-1001 plug repair 866 commercial 148
materials selection 982 selection, sandwich 284 composites 250
carpet plots 719 shaping 289,289 effect on mechanical
methods 1014-15 shear stress, formulas 284 properties 149
mold 447 size, sandwich 288 fiberglass 146-7
pultrusion 519-20 splitting, adhesives 271 polypropylene and acrylic
Consumer products 946 Cores acid 251
Contact density 276 resin interaction 149
 Index 1079

silanes 147, 250--1 795 and resin viscosity 354, 355

titanates 251 Crooked chain, meta-aramids 205 R1M 433
viscosity 250 Cross section selecting 585,586
zirconates 251 constant 489 times, tools 559
with diepoxides 105 shape of 461 degree of 587, 587
eliminating 193-4 Cross sections, strength 613 monitoring 67-9,67,68,69
and stiffness 702-3 Crosslinking 31, 48, 100 in situ 69
Coupon tests 195, 687, 688 matrix 460 infrared spectroscopy 68
Coupons reversible 887-8 off-line 67-9
configuration 721 rotational molding 551 part slippage 432
standard 721 Crowfoot satin, weave 145, 145 rate 509-10,509,510
Crack Crushed core epoxy resins 509
bridgUng 320,323-4,323 NDE 854 polyesters 509
deflection 320,323-4,323 sandwich 282 shrinkage 48
formation 469 Crystal, anisotropy 334 temperature 105
growth, fatigue 811 Crystalline adhesives 661
initiation 124, 632 polymers 118, 120 Cured epoxy resin systems, AS1M
opening displacement, see COD silica 250 methods 72
propagation 124 Crystallinity Curing 13,16,37,385
Cracking advantages of 126 agents 49,513,813,832
bonded joints 659, 659 and chemical resistance 120 amine 53-6, 54-6
matrices 319 PET 224 anhydride 56-9,56-9
Crashworthiness 3 Crystallite melting point, Spectra 26 anhydride-cured system 72
Cratering 814 Crystallization kinetics epoxy resins 509
Crazing 343 injection molding 540 amine 49-50,51
Creel rotational molding 551 anhydride 50
bookshelf 492 CTE 698-9, 704 autoclave 33
horizontal feed 494 aluminum 589 composite tubes 428-9
loaded multiple spindle 494 aramid4epoxy 589 cyclotrimerization 99-101
mat/fabric 494 boron-epoxy 589 dicyanates 100
package positions 493 cast ceramic 589 low temperature 567
pultrusion 492-5 compatibility 557 non-autoclave 589
Creep electroformed nickel 558 ovens 22, 428-9
effects 126 fiberglass-epoxy 589 press 22
environmental 1000 graphite-epoxy 589 reaction 49
failure high-temperature cast epoxy 589 R1M 445-6, 445
bonded joints 658 iron (electroformed) 589 shrinkage 41
hot-wet conditions 659 M401/F854 704 temperatures, carbonized organic
fatigue 631 nickel (electroformed) 589 composites 339-40
homopolymers 212 shrink factors 557-8 thermal 499
low temperatures BOO silicone rubber 589 time 385
para-aramids 212 tool steel 589 Curved
resistance tooling 556-7, 557-8 beams 791
aramid fibers 216, 222 urethane board stock 558 pultrusion 489-90
PET 227 Cupping 541, 541 Cut-off
Spectra 227,228-9,230,232 polyoxymethylene 541 high pressure water jet 502
rupture 797, BOO, 806 Cure saw 502,502
time-dependent BOO,800 catalysts 104 station 502
viscoelastic 797 control 586-9, 586, 587, 588 wastage 502
Cresyl glycidyl ether (CGE), objective 588 Cut-outs, in cylinders, GFRP 1013
viscosity 62 cycle Cutting
Crimp 416 anhydride-cured 72 aramid fibers 222
Critical aromatic system 71 Kevlar 600
bend radius, continuous ceramic autoclave 83-4 plies, surface preparation 870
fibers 309 stepped 355 speed 596-7,598,605
energy, failure criteria 795 carbon-carbon composites tools 329, 329
load, orthotropic plates 751 342 (;VI) 27,157,294,296,311,338-9,
processing temperature 326 elements 67 339
shear resistance, failure criteria no-bleed 673 carbon fibers 335
1080 Index

drawbacks 317 Debulking 567,569,592 textile preforming 397

isothennal 338-9, 339 periodic 574 thermosetting reaction 49
silicon carbide 157 Deburring 565 tows 476
temperatures 316 Decking 925 toxic 822
thermal gradient technique 339, Decomposition temperature viscosity 527
339 Kevlar 206 weaving 404
CVI 316,316,401 Nomex 206 Deflection
forced flow-thermal gradient PBO fiber 238 effect 640
processing 317,317 Technora 206 equations 750
isothermal processing 317 Teijinconex 206 fasteners 624, 626
Cyanate Twaron 206 formulas 284-5
ester, see CE Decontamination, PMC 863 lateral 644
resins 29-30 Deep delamination, NDE 854 sandwich 284
Cyanates 358 Deep submergence devices 474,475 Defonnable, adherends 633, 634
Cyanurate trimer 99,100 Defective fibers 303 Deformation 398
CYCAP 93 Defects, inspection methods 733--4, adherends 630
chemistry of 81 734 degree of 448
Cycle times, RTM 433 Defibrillation 215 laminates 691
Cycles to failure, Kevlar 213 Definitions point 302
Cycleweld 255 ararnid fibers 202 Degradation 215, 796
Cyclic loading, microcracking 801 autoclave 577 aminolysis 889
Cycloaliphatic amines, health effects BMC 379 chemical 888-9
832, 832, 833 BMI-DAB 105 factor 697
Cylinders BMI-MDA 105 fiber length 536
filament winding 467 braiding 413 glycolysis 889
Cylindrical shells 466, 752-3, 753 bypass load 620 hydrolysis 889
axisymmetric deformation 753--4 composites 378 methanolysis 889
buckling 755-7 consolidation 576 NDE 838
pressure 756 damage tolerance 794, 794 PET 889
nonsymmetric deformation denier 241 products obtained 889
754-5 detailed design 710 strength 796
shear defonnation 756 durability 794, 794 Degreasing, metal surfaces 871
stiffnesses 756 electroforming 591 DEH 20 (Dow) 54
Czochralski method 160 epoxide 48 DEH 24 (Dow) 54
equilibrium moisture content 241 DEH 26 (Dow) 54
D-glass 134 extrusion 534 de Havilland Mosquito 686, 686
Dagger drill 599-{)(}(), 599 fiber 242 Delamination 548, 597-8, 598, 599,
Damage placement 476 615,798-9,799,801
accumulation 797 hazard 823 atedge 798
assessment heat capacity 532-3 helical winding 789
composites 880 hygrothermal 694 Delta Clipper experimental launch
wet lay-up repairs 866 knitting 408 vehicle 974-5,974
drivers 801-3 lamina 687 Demolding 443,446-7
inspection, NDE 854 material properties 709 precautions 446
low velocity impact 839-40 matrix 378 DEN 431 (Dow) 52
modes 797-800, 798, 799, SOO particulate fillers 242 DEN 438 (Dow) 52
protective coatings, aircraft preliminary design 710 DEN 439 (Dow) 52
859-60 prepreg 425 Denier, definition 241
resistance 801-3 pultrusion 488 Denitrogenation 183
tolerance 107, 111, 794, 794-808 pyrolysis 888 Densification 314
definition 794,794 reinforcements 378 shuttle parts 343
Darcy's law 447,581 rheology 527 Density
resin flow 578 risk 823 ABS 258
OCB, tests 124 specific heat 532-3 aluminum 170,558
OCPD 37, 37, 39 specific modulus 715 aromatic system 71
low cost 37 steering 482 carbon (Type T300) 503
maleate half ester 37 strength retention 241 carbon-carbon composites 333
Debonding, moisture 811 tenacity 241 carboni epoxy prepreg 558
Debris impacts 814-15,814,815 Tex 136,241 cellulois acetate 258
 Index 1081

ceramic composites 312 continuous fiber 162 Dicyandiamide (DICY), melting

ceramics 307 staple 162 point 56
continuous ceramic fibers 309 whiskers 162 Dicyclopentadiene, see DCPD
continuous fiber 162 Design 32 Dicyclopentadienyl bisphenol
E-glass 503 aerospace 709 homopolymer property 102
electroformed nickel 558 allowables 709, 758-77 structure precursor 102
epoxies 258, 511 nomenclature 759,759 supplier 102
glass/epoxy 558 processing variables 762-3 trade name 102
high carbon cast steel 558 testing 758 Dielectric
Invar 558 checklist 712, 714 constant 95
Kevlar 206, 503 components 709 E-glass fibers 231
Kulon 305, 306 of composites 18-20, 19 PET 224
mahogany 558 considerations 994-1001 Spectra 231
meteoroids 815 data, using 724-7 thermoplastics 107
and modulus 23 detailed 710 thermosetting 107
monolithic graphite 558 laminates 686-708, 697-706 properties 106-7
Nomex 206 low strain 838 strength, Skybond 95
NR-150 82 methodology Dielectrometry 68,69
PAN-based fibers 170 knitting 410-13 bulk 68,69
phenolics 258 nonwoven textiles 419 Diels-Alder 101
pitch-based fibers 170 preliminary 710 reverse 80
PMR-1582 process 710 Dies 497
polyester 511 requirements chrome plated 501
polyurethane 258 aircraft 713, 1022-3 forming 497
polyvinyl chloride 258 stabilizer 710 heating and curing 498-501
recycled NBC 897 automotive industry 713 inspecting 501
recycled phenolic 894 helicopters 713 internal profile 499-500
regrind RIM 895 industrial pressure vessels manufacturing 500
resin, aliphatic system 70 713 multi-cavity 491, 500
S-glass 503 marine submersibles 713 positioning 498
SiC/ AIP3 315 recreational 713 purging 514
skinned molded foam 258 rocket motors 713 steel 500
Spectra 503 sailboats 713 temperature
staple 162 satellites 713 control 511-13,512,513
steel 170, 558 team 709, 712, 714, 715 profile 511, 513
and strength 23 materials supplier 715 stability 514
Technora 206 values 720-2 Diethyl ester diacid derivative 80-1
Teijinconex 206 verification 733-4, 733 Diethylaminepropylamine (DEAPA),
titanium 170 Designations, continuous ceramic viscosity 54
Twaron 206 fibers 309 Diethylenetriamine (DETA),
urethane board stock 558 Despooling 479 viscosity 54
vinylester 511 Detailed design 710, 723-34 Different thicknesses, sandwich
VMN-4 305, 306 definition 710 structures 744
whiskers 162, 308 Determinable values 779 Diffusion welding, solid stage
Dental applications, hydroxyapatite Diacids, and glycols 34-5 production 291
(HA) 958 Diameter Diglycidyl ether of 1, 4-butanediol
Dentin 958 continuous fibers 162 (BDE), viscosity 62
Deoxidizing, metal surfaces 871 ceramic 309 Diglycidyl ether of bisphenol A
Deposition low 160 (DGEBA)
electrolytic method 293 staple 162 crystallization 51
time, and coating thickness 294 whiskers 162, 308 viscosity 51
DER 330 (Dow) 51 3,3' -Diaminodiphenylsulfone, Diglycidyl ether of bisphenol F
DER 331 (Dow) 51 melting point 55 (DGEBF) 52
DER 332 (Dow) 51 4,4' -Diaminodiphenylsulfone (DDS), crystallization 52
DER 661 (Dow) 51 melting point 55 Diglycidyl ether of neopentyl glycol,
DER 732 (OOW) 63 Diaphragr.nforming 117 viscosity 63
DER 736 (OOW) 63 thermoplastics 116 Diglycidyl ether of polypropylene
Derivatives, phenolic 382 Dibromoneopentyl glycol 43 glycol 63
Description Dicyanates, curing 100 Diluents 61-3,61-3
1082 Index

reactive 66, 66 cantilever beam Duramite 247,247

Dimensional see also DCB Dust particles, milling 597
stability, reinforced composites specimens 124 Dy 023 (Ciba-Geigy) 62
387 head, pull winding 496-7 Dy 027 (Ciba-Geigy) 61
tolerances 560-1 lap DY 062 (Ciba-Geigy) 60
drawings 731 adhesive joints 628 DY 064 (Ciba-Geigy) 60
Dimensions, joints 612, 649 balanced 655 Dyeing, PET 233
DIMOX 315-16,315,316 joints Dyes 516
Dimpling, sandwich 282, 284 stress distribution 654 Dynamic mechanical analysis
Direct wet, chemical oxidation 191 thermal stress 662 (DMA) 816
Direction dependence 698 shea~joints 624,625 Dynamic mechanical thermal
Directional metal oxidation, see strap analysis, see DMTA
DIMOX adhesive joints 628 Dyneema 223
Directionally reinforced molding joints, tapering 646, 647, 648 availability 231
compotuld, see XMC Doublers 279-80, 280 elongation at break 225
Disadvantages Draft 389, 390 fiber type 225
AWJ 604 zero 389 pricing 232
hand lay-up 352 Drapability, preform 438 specific gravity 225
injection pultrusion 497 Draping simulation 448, 448 tensile modulus 225
laser machining 605 Drawings 730-3 tensile strength 225
preform 438 construction 730-3
RTM 434-5 dimensions 731 E-glass 24, 134, 425
turning 597 tolerances 731 composite tanks 985
twin-screw extruders 539 material description 731, 731 compression properties 360
ultrasonic machining 608 thickness 731 constituent properties 766
Disbonds Drill costs 717
inspection 839 cutting parameters 600 density 503
NDE 854 dagger 599-600,599 dielectric constant 231
surfaces 677 fixtures, fabrication 573 elastic properties 764-7
Discontinuous fully fluted 600 elongation at break 503
fiber 156 geometry 599-600 fiber modulus 360
representative volume 804 performance 599 Kevlar, fiber modulus 360
whisker 156 templates 573-4 polyester rods 985
Discrete, fiber architecture 401,401 twist 600 references 774--7
Disks, axisymmetric deformation Drilling 597-600,598,599 reinforced concrete 984
752-3 carbon fiber-epoxy 599 reinforcement 383
Displacements, measuring 779 coolants 600 specific
Disposal and cotultersinking 598-9 gravity 383
see also recycling, reuse damage 729 modulus 1049
nonrecyclables 899-900 glass fiber-epoxy 599 strength 1049
and reuse 883-904 machining 602 strength properties 764-7
Dissimilar adherends 635 secondary 598 tensile
Dissipation factor 95 Drive shafts 911,936-7 failure 383
fiberglass 135 Dry spots, eliminating 453, 453 modulus 360,383,503,716
thermoplastics 107 Dry tows, braiding 418 properties 360
thermosetting 107 Drying, of composite 811 strength 24,383,503
Distortion temperature, testing 38 Dual-shell reflector, Kevlar 1004, tension-tension ratio 1049
Diving equipment 926-7, 926, 927 1004 thermal expansion 383
Dixie Chemical 57 Ductile response 655 typical properties 1055
DMP 30 (Rohm & Haas) 60 adhesives 651-6,652,657 uniaxial strength 194
DMTA 120, 120 Ductility, adhesives 627-8,629-30, YOtulg's modulus 24
Dodecenyl succinic anhydride 655-6 Ears 371
(DDSA), melting point 57 DuPont FP fiber 309 Earthquakes 197
Domes, contours 467 elastic modulus 309 Ease of preparation, composite tools
Dornier 328 Durability 794-808 592
composite applications 1037, bonded joints 658-9 ECN 1273 (Ciba-Geigy) 52
1039 definition 794,794 ECN 1280 (Ciba-Geigy) 52
rear fuselage and fin 1039 PET 233 ECN 1299 (Ciba-Geigy) 52
Double predicting 796 Economics, land transportation
 Index 1083

905--6,906 conductivity Teijinconex 209

Eddy current testing 849-50 carbon black 245 Tekmilon 225
Edge carbon fibers 169 Twaron 208
definition, prefonn 438 ceramics 307 Vectran 235
delamination 111 equipment, pultrusion 519 vinylester 511
distances insulators, aramid fibers 215 VMN-4306
bearing load 621 properties EMI
bolted joints 617, 617 aramid fibers 215 testing 982
effects 780,789 carbon fibers 184-5, 185, 186 buildings 989, 991, 991
turning, reinforced composites fiberglass 135 EMI-24 (Air Products) 60
390 resistance, reinforced composites End effects 780
Edge-defined film-fed growth, see 387 End grain balsa 917
EFG resistivity, and microtexture 189 End-capping 81
Effect of contaminants, aircraft 863 Electrically conductive, particulate Energy
Effluent loss, and carbonization 183 fillers 250 equation 586-7
EFG 160 Electroformed nickel fracture surface 320
Egg crate structure 569, 570, 572 CTE 558 release rates 632
Eight-harness satin, weave 145, 145, density 558 Engineering
436,436 mandrel 591 constants
Einstein coefficient 531-2 thennal conductivity 558 IM6/epoxy 699,700,701
fillers 533 tooling 591-2, 592 laminates 693
Ejection, part 385 Electroformmg 591-2 parameters, preforms 421
Elastic definition 591 properties 22
constants Electroless deposition, coating 292-3 Environmental
graphite/epoxies 4 Electromagnetic interference, see EMI aspects, recycling 902-3
Kevlar 210 Electron diffraction 186 conditions, adhesives 630
preform 449 Electronics, applications 937, 937 effects 81G-20, 100G-l
displacement 195 Electroplating 820 composites 879-80
modulus Element tapering 624 creep 1000
aluminum 170 Eliminating history 810
carbon fibers 169 dry spots 453,453 relaxation 1000
continuous fiber 162 microcracking 1012 exposure 860-1
duPont FP fiber 309 Elongation 26 extremes, bonded joints 658
Kevlar 205 epoxy resins 29 properties, aramid fibers 213-15,
Kulon 305 NR-150 82 214
PAN-based fibers 170 nylon 6/6 PCI-glass 899 resistance 3,49,65,564
pitch-based fibers 170 PMR-1582 Epi-Rez 5014 (Hi-Tek Polymers) 62
staple 162 PP and granulated SMC 897 Epon 826 (Shell) 51
steel 170 recycled NBC 897 Epon 1001 (Shell) 51
titanium 170 recycled PP 895, 896 Epotuf 37-053 (Reichold) 62
VMN-4305 regrind RIM 895 Epotuf 37-057 (Reichold) 61
whiskers 162 Skybond 86 Epoxides 51-3,51-3
properties Elongation at break curing agent ratio 64
E-glass 764-7 Armos 208 definition 48
graphite epoxies 768-71 carbon (Type T300) 503 molecules 49
Kevlar 764-7 Dyneema 225 Epoxies
laminate 764-7 E-glass 503 adhesives 275
52-glass 764-7 epoxy 511 compressive strength 258
Spectra 764-7 Hoechst Celanese 225 density 258, 505
response 319,655 Kevlar 205,208,503 elongation at break 505
stress-strain, microcracking 798 Kulon 306 flexural modulus 505
symmetry, axis of 779 Nomex 209 flexural strength 505
zone length 658 particulate fillers 252-3 heat distortion 505
Elastomer, tooling 14 PHO 235 laminates, tooling 562
Elbows 464 polyester 511 maximum service temperature
Electric discharge machining (EDM) S-glass 503 258
605 Spectra 225, 503 nitrile rubber modified 275
Electrical SVM 208 pultrusion 511
applications 937,937 Technora 209 reinforced 196, 442
1084 Index

resills 28-9,29,30,382,504,832 routes 826 lap joillts 643

bismaleimides 6 to moisture, PET 227 criteria 193, 795, 802-3, 802
cure rate 509 Extended critical energy 795
curing agents 509 chaill critical shear resistance 795
cyanate ester 6 PET fibers 202, 223-34 laminates 695-7
elongation 29 manufacture 223-4 envelopes 696
flexibility 63 exposure,mechanicalproperties load, joillts 615
formulation 63-4, 64 215 mechanisms, columns 997
gel times 510, 510 Extruder die modes 621, 622, 797-800, 798,
glass content 510 extrusion 535, 535 799,800,804
modifiers 105 shape 535 properties 194-5
moisture absorption 29 Extruders stresses, joillts 626
phenolic triazine (PT) 6 sillgle screw 536 Fairings 1024
polyirnides 6 twill-screw 536 marille applications 923-4, 923,
pull loads 510 Extrusion 529 924
rigidity 63 cf. pultrusion 488-9 Fasteners
selection of 5 definition 534 bendillg 625,625,729
viscosity vs. time 510, 510 extruder die 535, 535 deflections 626
shear modulus 259 melt pumping 534-5,535 clearances 624, 626
shear strength 259 melting 534 countersunk 624, 624, 625, 625
tensile modulus 505 orientation 538 deflections 623,624,625,626
tensile strength 258,505 plasticating 534,535 design issues 729
thermal conductivity 259 solid stage production 291 diameter 623-4, 625
Equations solids conveyillg 534, 535 illcreasillg 614
of continuity 582 thermoplastics 526, 534-8 effects 612,624-6,624,625
lamina material 760 Eye protection 830 joillt around 612, 612
of motion 738-9,740-1 loads 622
illertia terms 739 3F dianhydride 77 mechanical 517
straill-displacment 739 3F/36F polyimides, chemistry of 77 multi-rowed 620
Equilibrium FAA advisory circulars 1063,1065 multiple arrays 612
crack spacillg 798 Fabric formation 402 parameters 517
moisture content, definition 241 Fabrication 31, 33 protrudillghead 624,624
Equipment composites 879 selection 729, 729
selection, consolidation 577 drill fixtures 573 tension head 625
wet lay-up repairs 867 first article 372 Fatigue 811
ERL 4206 (Union Carbide) 63 low cost 33 crack growth 811
Erosion 813 marille applications 919-20 low temperature 811
atomic oxygen 814 methods 762 para-aramids 213
Ester formation 34 quality 33 properties 386, 387
Esterification 35, 38 techniques 13-18,14,15,16 resistance
EtchUng 667 Fabrics, orientation 504 aramid fibers 221, 222
Ether linkages, Technora 210 Face PET 230
2-Ethyl-4-methylimidazole (EMI), dimpling, formulas 285 Technora 213
melting poillt 60 sheet thermoplastics 115
Europe, recydillg 884 materials removal 864 Feedrates 598
European retrievable carrier 971 repair 866-7 Feldspar 243-4
Evacuation,P~C 863 wrUnklillg chemical resistance 243-4
Evaluation, strength 737 formulas 285 particulate fillers 243-4
Exothermic reactions 827-8 sandwich 282, 284 refractive illdex 243-4
Exotherms 68 Facing Female toolillg, large power yachts
Expansion process, honeycomb 257 failure, sandwich 282 921
Expendable materials 359--60 material 255-6 Femoral components 960-1
Exposed surface, voids 432 Failure 193 design 960
Exposure adherends 627-9, 628, 657 development 961
alkaline 46 analysis, composites 880 FEP 362
assessment 826 aramid fibers 210 mold releases 362
limits 824-5,826 bond 627-9,628 Fiber
terminology 826 characteristics alignment 494
preventing 827 joillts 659,660 aluminum based 159
 Index 1085

architecture 398 separation 615 rebars 999

continuous filament 401-2 shape, carbon fiber 335,336,337 reinforced plastics 23
discrete 401,401 strength, degradation 806 reinforced polyesters, see FRP
infiltration 401 tensioning 457 reinforcing 38
integrated 401, 401 to resin ratio 509 rovings 136,146-7,353
interlooped 402 trade names 158,225 specific gravity 135
laminar 401, 401 type 225 spire 991
linear 401,401 volume 721 staple fibers 131
planar interlaced 402 fraction 21,404,412,416,578 strand 131
areal weight 425 3-D fabric 416 suppliers 136
breakage,~E 838 and braiding angle 417 surface resistivity 135
breakout 599 carbon fiber 335 tapes 146
breaks, NDE 854 and compressive fiber, Te glass 134
bundles 403 stress 581 tensile strength 134
coating 166, 802 Gutowski's model 588 thermal properties 135
composition 158 high 585--6 thickness 141-4
continuous 156 and permeability 399-400, three-dimensional fabrics 146
definition 242 400 volume resistivity 135
deformation 580-1 processing window 412, warp yam 141-4
consolidation 576 412 water resistance 136
curve 580-1,581 waviness 780 weave 141-4
diameter 156 Fiberfrax 160 weight 141-4
structural hierarchy 778 Fiberglass yarn 138
direction, tensile strength 804 acoustical properties 135 designation 139-40
discontinuous 156 aircraft industry 136 yield 139-40
distribution 154, 154 antistatic agents 146-7 Fiberglass-epoxy, CTE 589
preform 438 binders 147 Fiberite 99
dominated, composites 795 breaking strength 139-40,141-4 Fibrous
flow 527-31 bushing 138 dust 835
d. particulate flow 530 chemical composition 147-8 reinforcement types 795,796
fracture 799, 799 chemical resistance 134 Filament
fragments, inhaling 835 circuit boards 136 angle of 456
length, degradation 536 composition 133-4, 134, 138--9 cross-section, PET 225
low risk 835 continuous 131 diameter 139,148--9
manufacturers 158 strands 147 Kevlar 206
manufacturing processes 158 count 141-4 Kulon 306
modulus 360 coupling agent 146-7 Nomex 206
orientation 527-31 dissipation factor 135 PET 225
compression molding 548 electrical properties 135, 136 Technora 206
injection molding 543, 543 epoxies 153 Teijinconex 206
preforms 421 fabric count 140 Twaron 206
packing fraction 412 fibers VMN-4306
placement 17, 476--87, 477 attributes 358 directionality, carbon fibers 335
definition 476 hollow 134 equilibrium,equation 459
head 476 milled 146 equilibrium 459-61
inspection 486-7,487 filament designations 133 lay-up 457
machines 478,482 filament diameter 133 shape 206
materials 478--9 filling yarn 141-4 strength, S-glass 24
steering 482 fire resistance 134 on surface 459-61
surface geometry 484--6 fluted core fabrics 146 tension 457
tooling 479-80 heat resistance 134 winding 17,17,66,116,117,
properties 400,400 laminates 356 456-75,762
pull-out 320, 323-4, 323 lubricants 146-7 applications 471-5
rayon-based 173 manufacturing 132 continuous fiber 338
reinforced polymers, see FRP mat 137--8 cylinders 467
reinforcement 22,23,216,435-6 mechanical properties 135 low costs 463
for reinforcing 166 optical properties 135 machine 463,464
and resin 425 platforms 989 marine applications 920
rovings 492, 503 proof testing 849 pressure vessels 471
1086 Index

rackets 1050 impact strength 249 plied-yam 151

spherical shapes 466-7 mica 249 polyester 511
sporting goods 1046, 1047 silver 249, 250 recycled PET 894
wet 762-3 Flame recycled phenolic 894
Filed inspections, neutron polishing 157 recycled SMC 890, 892, 893
radiography 848 spread 44 5-glass epoxies 152
Fill,yarns 407,407 Flammability, evaluating 44 short fiber mat 401
Fillers 382, 516 Flank wear 599 single-yam 151
barium ferrite 527 Flat SMC 386,386
barium sulfate 527 cylinders, filament winding 467 unidirectional tape 401
calcium carbonate 527 specimens vinylester 511
carbon 527 compression testing 787 woven laminates 401
cellulosic 527 testing 781, 786-7 yam distribution 154
clay 516 tension 786-7 ZMC 386,386
coarse, disadvantages 516 tape, lay-up 16 stress
comminuted polymers 527 Flat-layer, microstructures 177 PP and granulated SMC 897
effect on processing 552 Flaws, planar 840 recycled PP 895, 896
Einstein coefficients 533 Flex modulus Flexure
fine 516 NR-150 82 TEOS 346
functional 382 PMR-15 laminates 88 testing 195
glass 527 Skybond 86 Float 150
ground limestone 382 Flex strength Floc 216
ground petroleum coke 527 AFR700B/S21aminates 93,95 Flocking lay-up, short fibers 337, 338
inorganic phosphate-based 957 Cellon 87 Flow
medical applications 957 NR-150 82 measurements 357
metallic oxides 527 PMR-15 laminates 88 and rigidity 67
mica 527 Skybond 86 Fluid contamination, aircraft 859
molybdenum disulfide 527 Flexibility 161 Fluids
non-functional 382 Flexible aircraft 812
pultrusion 516 backbone polyester resin 381 automotive 812
shape of 531 mold wall 444 methylene chloride 812
silica products 527 Flexural Newtonian 527, 528
silicon carbide 527 modulus Fluorine, toxicity 362
specific heat 533 BMC 386,386 Fluoroethylene propylene, see FEP
Fillet forming, adhesives 271 epoxy 511 Flushing contaminants 864
Filling 301,301,302 LPMC 386, 386 Fluted core fabrics, fiberglass 146
mold 442 nylon 6/6 PCI-glass 899 Flywheel mechanical battery systems
single-yam 151 polyester 511 938,938
yam 140 polyetherirnide 545, 546 Flywheels 474-5
fiberglass 141-4 polysulfone 546 Foam
Film impregnation 8 PP and granulated SMC 897 adhesive voids, NDE 854
Film-stacking 117-18,118 recycled NBC 897 core materials 256-7
Filter cloths, PET 233 recycled PET 894 cores for fiberglass 256
Finish 215 recycled PP 895, 896 injection 256
5-glass epoxies 152 recycled SMC 890, 892, 893 insulation 256
Fire regrind RIM 895 in place system 257
prevention 367 SMC 386,386 polystyrene 256-7
resistance, fiberglass 134 vinylester 511 PVC 257
retardance 36,42-5 ZMC 386,386 radar transparency 256
halogen inclusion 43 properties shear strength 257
retardants, particulate fillers 249 Spectra 226, 226 Folding 547
Firefighter breathing apparatus testing 38 low-cost 117
937-8 strength thermoplastics 116
Fishing poles 427, 1045, 1051 BMC 386,386 Forced flow-thermal gradient
FIT-technology 117 epoxy 511 processing, CVl 317, 317
Flags fibreglass epoxies 153 Foreign materials
bias 425,426 LPMC 386, 386 NDE 854
longitudinal 425 marine laminates 920 X-ray imaging 844
Flakes 248 nylon 6/6 PCI-glass 899 Formability
 Index 1087

jamming angles 399 fire retardant 42 cut-outs in cylinders 1013

preforming 398-9 flame resistant 44 development 1016
weft knitted fabrics 399 low bending stiffness 960 diffusivity 1008
woven glass fabrics 399 orthopedic applications 959 economics 1020
yarn slippage 399 permeability 959 hygroscopic nature 1006
Forming dies 497 pins 959 impact damage 1006
fabrication 497 properties of 46 joints 1012, 1012
materials used for 497 pultrusion 47 material cost 1006
Forms strength of 45 microcracking 1006
aramid fibers 216 thermal performance 45 moisture effects 1007-10
boron 161-3 total hip arthroplasty 959 new materials 1019-20
ceramic fibers 161-3 Fuel tanks 939-40 peel strength 1006
high silica 161-3 Fully fluted drill 600 predictions 1019
quartz 161-3 Functional fillers 382 properties 1005
Vectran 236 Fungal growth 810-11 undesirable 1006
Formulas Fusion 157 springback 1013
bending stress 284 bonding 127 temperature extremes 1004-21
core shear stress 284 PEEK 127 warping 1013, 1013
deflection 284-5 Future directions, transportation 915 woven broadgoods 1024
face dimpling 285 Gibbsite 245
face wrinkling 285 Galvanic, corrosion 108-9 Glass
moment of inertia 285 Gap 487 bottle industry 97
safety factor 285 Gas laser content
Formulation applications 605 epoxy resins 510
diglycidyl ether of bisphenol A C02 605 nylon 6/6 PCl 899
hexahydrophdlalic Gas-spargmg 174-5 pP/PCI 898
anhydride 64 Gate 541, 542 fabrics, knitted 917
diglycidyl ether of bisphenol A fiber orientation 543 fiber-epoxy, drilling 599
triethylene diamine 64 Gating 450-1 fibers 24-5,25
diglycidyl ether of bisphenol A Gel alkaline environment 1000
triethylene tetramine 64 coat 567 applications 24
Fortafil 1054 spinning 223-4 availability 24
properties of 170 time 37,66,357,446 cost 24
Fountain flow, injection molding 542 aromatic system 71 handling 24
Fractionizing plant 887 zone 50S health effects 835
Fracture 597 Gelation history 131
effects, isolating 799 pressure during 508 processing 24
elongation 170 resins 499 reinforced plastics (GFRP)
fixation devices 961-2 Gelstar Thermal Analyzer 515 1024
mechanics, joints 631-2 General reuse 884
path 324, 324 aviation, applications 1039-40 S-glass 24
strength 325 equations 738-45 silane coupling 24
toughness Geodesic toxicity 24
alumina composites 325 curvature 459 fillers 527
Celion 87 deviation, angle of 460 history 131
ceramic composites 312 line 457 length, recycled PP 895, 896
high silica 156 Geometry mat thermoplastics, see GMT
moisture 811 joints 638-51 reinforced plastic, see GRP
NR-150 82 single lap joint 641 rovings 502
PMR-1582 Germany to resin, ratio 150-4, 153
Sliar 325 recycling 902 transition temperature
Tateho 325 standards 1065,1066 see also T
thermoplastics 122, 124 GFRP 1004-21 aromaticgsystem 71
Free radical generators 104 aerospace applications 1007 moisture 811
FRP 38,41 anisotropic behavior 1012-14 RIM 440
composites, pultrusion 517 anisotropy 1006 Glass/epoxy
corrosion resistant 44 applications 1016 coefficient of thermal expansion
corrosive attack 45 assembly bowing 1013, 1014, 558
durability 959 1014 density 558
1088 Index
thermal conductivity 558 thermal data 818 fibrous dust 832
Gloves 828-9, 828 weathering 813 HDPE
resistant 829 Graphite/polysulfone, weathering glass-fiber reinforced 549
types 829 813 graphite filled 549
Glued laminated timber 991 Graphite/wood, hybrids 354 mica flake reinforced 549
beams 940-1 Graphitization 169,183 HDT 126,126
Glycidyl alignment 191 recycled PET 894
amines, health effects 832 large regions 170 Health
compounds 831-2,832 Grating 936 applications 943-4, 944
ethers, health effects 832 Green form 314 effects
Glycol Green strength 446 aliphatic amines 832, 833
and diacids 34-5 Grill opening panels, SMC 907 amino resins 834
propylene 35 Grinding 600 anhydride curing agents 832,
selection 39 accuracies 600 833
Glycolysis, degradation 888 cryogenic 886 aramid fibers 835
GMT 115 polymer matrix composites 600 aromatic amines 832, 833
automotive market 127 silicon carbide wheels 600 bisphenol A-based 832
compression molding 117 surface speeds 600 carbon fibers 835
extrusion compounded 118 Grit blasting, surface preparation ceramic fibers 835
semi-finished 118, 118 870-1 chlorinated solvents 836
Goland-Reissner 641 Ground limestone 382 cycloaliphatics 832, 832, 833
Golf shafts 429,463,465,698,1051 Ground petroleum coke, fillers 527 glass fibers 835
filament wound 1051 Growth factors 558 glycidyl amines 832
test methods 1051 GRP 839 glycidyl ethers 832
Grafil 1054 Grumman F-14 1030 graphite fibers 835
Granulation Grumman X-29 aircraft 698, 698 imides 834
knife 886 Guide pins 442 ketones 836
recycling 886-7 compression molds 395 phenolics 834
Graphite 245, 334, 334 Gutowski's model 578,580,581,582, polyamides 832, 833
3-D lattice 334 583,585 polyaminoamides 832, 833
composites, machining 599 fiber volume fraction 588 polyurethanes 834
compression properties 360 numerical schemes 586 thermoplastic resins 834
costs 717 GY 281 (Ciba-Geigy) 52 hazards
epoxies GY 6010 (Ciba-Geigy) 51 asbestos 252
elastic properties 768-71 Gypsum 564 beryllium oxide 252
references 774-7 composite processes 831
strength properties 768-71 Half-discs, testing 782 particulate fillers 252
fibers 25-6 Halogens 44 and safety, composites 880
competitive prices 25 bromine 43 Heat
health effects 835 Hammer handles, pultrusion 491 application
modulus 360 Hand lay-up 352-77, 762 heating blankets 874
price 716 advantages 352 repairs 874-7
production 25 applications, aerospace 375 risks of 875,877
reinforced plastics, see GFRP disadvantages 352 capacity
reinforcement 353 large power yachts 921 definition 532-3
specific gravity 383 marine applications 376, 919-20 thermoplastics 532-4
tensile precautions 363 cleaned, reinforcements 898-9
failure 383 Handling deflection temperatures, see HDT
modulus 360,383,715,716 aramid fibers 222 distortion 41
properties 360 glass fibers 24 epoxy 511
strength 383 PET 233 polyester 511
thermal expansion 383 Hardeners 49, 832 vinylester 511
turbostratic layers 184 Hardness distortion temperature 38
Graphite/epoxies 4 ceramics 307 aliphatic 70
CIE 589 regrind RIM 895 anhydride-cured 72
elastic constants 4 tooling materials 577 aromatic 71
physical properties 4 HarrierVTOL 1031-2 lamps, wet lay-up repairs 865,
strains 4 Hazard 823 866
strength properties 4 definition 823 resistance, fiberglass 134
 Index 1089

treatment, carbon fibers 335 carbon-carbon composites 269

Heated curing dies 500-1 344 plate shear modulus 266-7,
design 500 cast epoxy, CTE 589 269
Heating resins 818 plate shear strength 266-7,
blankets sandwich 288 269
repairs 869 High-pressure tubing 941 Kevlar 271, 272, 273
wet lay-up repairs 865 High-speed train brakes 941 paper 271
radio frequency 499 Higher temperatures, and tensile metal, roll-forming 289
rate 181, 499 strength 182 non-metallic, thermal resistance
single zone 511, 513 History 260
Heavy liquids, adhesives 275-6 consolidation 578-86 panels, repairs 871-7
Heel blocks, compression molds 394 pultrusion 488 paper 263, 268
Helical winding 460 recycling 883-4 plate shear modulus 264-5,
delamination 789 HMC 381 266-7
reinforcement 461 Hoechst Celanese plate shear strength 264-5, 266-7
Helicopters availability 231 specimen geometry 263
applications 1040 elongation at break 225 stainless steel based 257
current prices 1026 fiber type 225 test method 263, 263
design requirements 713 pricing 232 thickness 263
rotor blades 1024 specific gravity 225 titanium based 257
Heloxy WC-63 (Wilmington tensile modulus 225 Hoop
Chemical) 62 tensile strength 225 stresses 621
Hercules Hole fittings, tooling 562 winding, reinforcement 461
3501-6 epoxy resin 579,579 Hollow fiber, fiberglass 134 wound, rings 789
IM7 fiber, thickness 478 Holography 851-3,852 Horizontal
HETacid 36 laser interferometric 852 drawing 299
Hexachlorocyclopentadiene, see HET phase locked loop 852 tape wrapper 428, 428
Hexafluorobisphenol A Homopolymerization 40,50-1,51, Hot
homopolymer property 102 511 air blowers, wet lay-up repairs
structure precursor 102 catalysts 49,59, 60-1 865
supplier 102 Homopolymerized BF3MEA 65 extrusion
trade name 102 Homopolymers, creep 212 compression 298
Hexafluoroisopropylidene, bridging Honeycomb 257, 260-71 MMC 291
77 aluminum 268 oil jackets 499
Hexahydrophthalic anhydride alloy pressing 314
(HHPA), melting point 57 compressive strength spots, particulate fillers 242
High 264-5 Hot-wet service 99
carbon cast steel plate shear modulus HRDI optical bench 1017, 1018
CTE 558 264-5 HS carbon composites
density 558 plate shear strength 264-5 shear modulus 1057
thermal conductivity 558 alloys 268 tension 1057
local stresses 729 thermal resistance 260 HT 972 (Ciba-Geigy) 55
modulus, carbon fibers 335 aramid paper 260, 269-70 HT 976 (Ciba-Geigy) 55
pressure water jet, cut-off 502 assemblies, acoustic emission 849 HT 9720 (Ciba-Geigy) 55
silica carbon fiber 270-1 Hubble Space Telescope 967, 969
applications 163, 165, 166 carving bits 289 Humidity 3
continuous fibers 163 cell configurations 262 Humphrey Chemical 57
forms 161-3 cell shape 261-3,262 HY 906 (Ciba-Giegy) 57
fracture toughness 156 cell size 263 Hybridization, PET 233
manufacture 156-7 compressive strength 264-5, Hybrids 795
properties 161 266-7 aramid/graphite 354
strength core shear strength 261 composites, reuse 883
carbon fibers 335 corrugation process 257 graphite/wood 354
molding compound, see HMC defects, X-ray imaging 845 materials 359
PET 233 density 261, 261, 262 Spectral graphite 354
reinforced composites 387 expansion process 257 Hydraulic
temperature glass-reinforced 268 ejection, part removal 446
adhesives 658 applications 268 system, testing 782
applications 818 compressive strength 266-7, test technique 782, 788
1090 Index

liydrocarbons 37 strength Nomex 209

liydrocodes 815 aliphatic system 70 SVM 208
liydrogen flakes 249 Technora 209
fuel storage 941 testing 38 Teijinconex 209
reduction 161 TEOS 346 Twaron 208
liydrogenation 179 Imperfections 780 Initiators, and inhibitors 382
liydrolic stability 38 Implants, biologic response 960 Injection
liydrolysis Impregnation cycle 442
BMI 112 carbonized organic composites gate 452
CE 112 340 molded 380
degradation 888 powder 118 molding 22,529,532
resistance 41 prepolymer 118 cavity filling 542
liydrophobic, PET 230 Impregnators, marine applications clamp 539-44, 540
liydrous aluminosilicate 244 919-20 crystallization kinetics 540
liydroxyapatite (lIA) 958 In-mold coating (IMC) 385 fiber orientation 543,543
dental applications 958 In-plane shear fountain flow 542
glass-reinforced 958 Kevlar 212 particulate fillers 242
liygroscopic nature, GFRP 1006 methods, testing 783 plasticating 538, 538
Hygrothermal modulus, AS-4 carbon fiber 124 polymer 540
definition 694 Incineration 888 pumping section 538
effects 694-5 nonrecyclables 899-900 rackets 1050
load 695 Inclination angle, yarns 407 resins 122
properties 96 Inclusion of particulates, polymers screws 539
Kevlar 212 528 thermoplastics 116,526
Hypersonic vehicles, materials for Inconel329 ports 442-3
973 Industrial centred 451, 451
equipment 951 comer 451,451
ICBM equipment 375 hygiene 825-30, 827, 828, 829 pressure 444
Identical adherends 639 order of priorities 828 RTM 433
IM6/epoxy pressure vessels, design pultrusion 497, 498
engineering constants 699, 700, requirements 713 sensor controlled 452-3
701 Indm.try, pulp and paper 45 Inplane shear 374
longitudinal Inert, PET 230 Inspection
CTE 699 Inertia 3 bonding 668
extension 701 Infiltration composites 880
strain 701 carboni aluminum 303 criteria 487, 838
tensile modulus 699 fiber architecture 401 disbonds 839
Poisson's ratio 699 improving 300 fiber placement 486-7, 487
shear modulus 699 isothermal gradient 338 first article 486
stiffness coefficients 702, 703 of preforms 299 in-service 839
strength ratio 701 pressure gradient 338 methods
transverse pressure pulsation 338 adhesive joints 627
CTE 699 spontaneous 300 defects 733-4, 734
extension 701 technology 300 and nondestructive testing 1059
modulus 699 under pressure 299-300 on-aircraft 849
strain 701 vacuum 299-300 paint surfaces 839
volume fraction 699 Inflatable bladders 874 part 733
IM-7/PEEK, strength 125 Infrared thermography 850--1, 850, post cure 572-3
Imidazoles 104 851, 1059 speeds 842
Imides, health effects 834 aerospace 851 subsurface damage 839
Impact marine applications 851 times 839
damage Inhalation, measure by 824 ultrasonic 1059
aircraft 858-9 Inhibitors, and initiators 382 visual 839, 839-40
carbon-carbon composites Inhomogeneities, NDE 838 Insulation
334 Inhomogeneous, composites 797 foam 256
GFRP 1006 Initial rise 646,647 resistance, Skybond 95
tap testing 858 Initial tensile modulus Integrated, fiber architecture 401,
energy, thermoplastics 125 Armos 208 401
resistance, moisture 811 Kevlar 208 Inter-yam slip 448
 Index 1091

Interaction, constituents 801-2 NBC 897 single lap 626, 627, 729
Intercontinental ballistic missile PET 894 single shear 624
program 353 phenolic 894 bending moments 624, 625
Interface, fiber/matrix 123 PP 895,896 step lap 645-6,645,649,650,651,
Interlacing patterns, braiding 415, SMC 892,893 652,659
415 SMC 386,386 strength
Interlaminar LPMC 386, 386 experimental 616
fracture 682 recycled, SMC 890 predicting 615
shear 790-2 ZMC 386,386 tests 626
TEOS 346 stress analysis 611
shear strength Jamming 412, 417 structural performance 613
Celion 91 angles 399 test methods 626
PMR-15 laminates 88 Japan two fastener 623
stresses 738 bathtub manufacture 889 untapered 624
eliminating 801 recycling 887,888,903
tension, TEOS 346 Japanese Industrial Standards 1065, Kaolin
Internal 1066 calcined, hardness 244
bone fixation 957 Jeffamine T 403 (Texaco) 54 particulate fillers 244
friction background 296, 296 Joining 727-30,728,729 Kapton 820
mold release 383 selection process 727, 728 Kardos' model 578, 583
International Organization for thermoplastics 127 Kerf
Standardization 1066, 1067 thermosets 127 cutting 604, 606
Interphase, toughening 802 Joints width 606, 695
Invar adhesive 517-18,610-63,627-63 Kerimid 99
CTE 558 bearing load 626 Ketones, health effects 836
density 558 bearing strengths 618-19 Kevlar 108,472,598,761,1024
linear thermal expansion 705 bending failures 617, 617 availability 217-18
thermal conductivity 558 bolted 611, 617, 617, 1015 breaking strength 214
Inverse method, contact 611 bonded 658-9, 1015 charring 815
Ionic spraying 300 step lap 627 chemical
Ionizing radiation 816 and bonding 374 resistance 214
Iosipescu shear test 790 bypass load 626 stability 214
Iron (electroformed), CTE 589 clamping pressure 618 colored 216
ISO standards, carbon fiber computer codes 624, 627, 630 constituent properties 766-7
composites 1067 design 728 cutting 600
Isophthalics 39 dimensions 612, 649 cycles to failure 213
polyester disassembly 611,728 decomposition temperature 206
blistering 918 double shear 624 density 206,503
marine applications 916 failure 615,618,618,626,659,660 dual-shell reflector 1004, 1004
resins 504 fracture mechanics 631-2 E-glass 360
resins 35,36 geometry 613-17,613,628,629, elastic
Isothermal 638-51 constants 210
processing, CVI 317 GFRP 1012, 1012 modulus 205
thermal gradient, infiltration 338 load magnitude 728 properties 764-7
Isotropic materials 618-19,618,619 elongation at break 205, 208, 503
casting, short fibers 337, 338 mechanical response 630 filament
layers, different moduli of mechanically fastened 610-63 diameter 206
elasticity 744 advantages 517 shape 206
materials 687 d. adhesive 610-11 hygrothermal properties 212
metallic tape, winding 468 metallic 613 in-plane shear 212
nonwoven fabric 164 multi-fastener 612, 613, 619-24, initial tensile modulus 208
plates 614 620, 621, 622, 623 linear thermal expansion
Italian, talc 248-9,248 multi-row 616 coefficient 206
Izod open hole coupon strength 626 longitudinal compression 212
impact predicting peak stresses 613 longitudinal tension 212
BMC 386,386 pure bearing load 626 machining 600
nylon 6/6 PCI-glass 899 scarf 611,623 marine applications 917
PP and granulated SMC 897 selecting design 627 melting 815
recycled single fastener 612,612,613 temperature 206
1092 Index

milling 600 Kulon 300 fiberglass 356

modulus and temperature 210, bending strength 305 D\16/epoxy 699,700,701
211 carbon content 306 kinematics 690-1
moisture content 206 coating thickness 306 linear bending 749-50
properties 205 compressive strength 305 load carrying capability 696
references 774-7 density 305, 306 loads on 689, 693
refractive index 215 elastic modulus 305 macromechanics of 193-4
reinforcement 383 elongation at break 306 mechanical properties 717
sources of information 221 filament diameter 306 nonsymmetric 752
specific longitudinal CTE 306 off-axis stiffness 689-90
gravity 208, 383 specific modulus 305 with plies 697
heat 206 strand properties 306 ply angle 689,689
modulus 1049 tensile modulus 306 ply stacking sequence 692
strength 1049 tensile strength 305, 306 quasi-isotropic 9-11,697
strength retention 205, 215 reinforced 321-3,321,619
stress rupture 212-13,213 Labour requirements, pultrusion 489 resulting strain state 693-4
tensile Ladder polymer 178-9,178 selection, carpet plots 720
failure 383 Ladders skin 694
modulus 383,503 applications 936 stacking sequence 616,630
strength 208,383,503 pultruded 518 stiffness matrix 693
and temperature 210, 211 Lamina 687-9,688 strength 695,696,764-7
tension-tension ratio 1049 allowables 759-62 ratio 696
thermal definition 687 stress
coefficient 383 macromechanics 192-3 patterns 322
conductivity 206 material resultants 691-3
transverse equations 760 symmetric 194,692,749
compression 212 properties 760 thermal stresses 322,589
tension 212 properties void free 588
trimming 600 adjusting 763 Land transportation 905-15
turning 600 estimating 761-2 economics 905-6,906
twist 215-16 references 772-3, 774-7 history 906-8, 907
ultimates 210 three-dimensional 761 market growth 905-6,906
weight loss 205 two-dimensional 760-1 Landfill, nonrecycIables 899-900
Kevlar / epoxy thickness, structural hierarchy Lap joints 638-44
thermal data 818 778 abrasive cleaning 872
weathering 813 Laminar failure characteristics 643
Kinel99 discontinuities 843 peel stresses 637-8,637,638
Kinematics, laminates 690-1 fiber architecture 401,401 single 640, 641
Kinking 302 Laminated symmetric 639
Kneaded molding compound 381 plate theory 689, 690-2 Laptop computers 942
Knitted structures 585 Large Area Composite Inspection
glass fabrics 917 Laminates System (LACIS) 842, 842
reinforcements 916-17 advantages of 322 Large diameter, PAN 180,181
Knitting 402,402,408-13,409,410, allowables 762 Large power yachts
411,412 analysis of 11 blister protection 921
3-D fabrics 409 balanced 194, 693 displacement hulls 922
definition 408 bending 730 female tooling 921
design methodology 410-13 boundary conditions 689 hand lay-up 921
machines 409 carbon-epoxy 669 marine applications 921-2
multiaxial warp knit, see MWK coefficient of thermal expansion variable mold 921
stitch formation 409 322 Laser
unit cell geometry 411 compressive strength 616 beam 601
warp 408, 409 cross-ply 322 heated floating zone, see LHFZ
weft 408,409 deformation 691 machining 605
Kortex 108, 271 design 686-708 shearography 853, 853
Kozeny constant 579 elastic properties 322, 764-7 immunity to vibration 853
Kozeny-Carman equation 400, 422, engineering constants 693 Lateral
422,579 failure criteria 695-7 cohesion 184
Kraft paper 268 failure in tension 696 compression tests 449
 Index 1093

deflections 642 lamillates 749-50 loadillg 213

bond stresses 640 cutting, machllllllg 602 microcrackillg 801
Lathes, CNC 597 elastic response to failure 615 use temperature 206
Lattice fiber architecture 401,401 Longitudlllal
fringe imaging 186 thermal expansion bendillg 747
structures 743,744 aliphatic system 70 compression, Kevlar 212
Lay-up 15, 16 alumillum alloy 705 CTE
carbon-carbon composites 342 aromatic system 71 IM6 I epoxy 699
contoured tape 16 concrete 705 Kulon 306
cornertechrUques 373 Invar 705 M401/F854 699
filament 457 steel 705 PBO fiber 238
flat tape 16 titanium alloy 705 extension, IM6/epoxy 701
manual 16 thermal expansion coefficient ply wavilless 432
molds 566-7 carbon--carbon composites straill, IM6/epoxy 701
operation 576 333 tensile modulus 699
sequence, carpet plots 719 Kevlar 206 tension
wet 353,355 Nomex 206 Kevlar 212
Layer reinforced, carboni aluminum Technora 206 test 688
304 Teijillconex 206 Loomcomponents 942
Layered, adherends 646 Twaron 206 Lot-to-Iot variation
Layers yarns,lllsertion 410,410 tensile modulus 725
coordillates 740,746 Llllkages, flexible 75 tensile strength 725
orthotropic 742 Liquid Low
LOEF 970-2, 970, 971, 972 illjection moldlllg simulation, see bendillg stiffness, FRP 960
composite specimen testbed LIMS cost parts, RTM 433
970-1 processillg 314-16 density
orientation ill orbit 971 stage, MMC 291 composite tools 592
Lead powders 249 waste 889 polyethylene, viscosity 528
Leaf springs 910,910 as fuel 889 K-glass 134
Liteflex 910 Lithium alumino silicate (LAS) 318 pressure
Leak paths, aircraft 859--60 LMOs 179, 182 grit blasting 631
Leakage 456 size of 179--80 moldlllg compound, see
Legal aspects, recycllllg 902-3 Loading LPMC
Length conditions 736 temperature, fatigue 811
to diameter ratio, aramid fibers history, remallllllg strength 804 toxicity; thermoplastics 115
204 levels, particulates 528 viscosity, ZMC 380
whiskers 308 methods 779 void, composites 79
LEO 813-17 selecting 787 voltage anodizillg, metal surfaces
Lewis sections, test specimen 786 871
acids 50-1,60--1 slow cyclic 627 Low-earth-orbit, see LEO
bases 50-1,60--1 stress, adherends 642 LPMC 381
LHFZ, directional solidification 160 Loads flexural modulus 386, 386
Life expectancy, tools 559 bypass vs. bearillg 620,620 flexural strength 386, 386
Light carryillg capability, lamillates 697 formulation 384
aircraft 1024 direction of 779 IZOO 386, 386
grit-blasting 668, 670, 671 distribution 621 shelf life 381
alumllla grit 675, 676 enhancement 655 specific gravity expansion 386,
machllle 677, 677 fastener, variable 622 386
liquids, adhesives 275-6 hygrothermal 695 tensile modulus 386, 386
weight, reinforced composites measuring 779 tensile strength 386,386
387 path eccentricity 637, 637 thermal coefficient 386, 386
Lighting poles 948-9, 948 sillgle lap joillt 637, 637 Lubricants, fiberglass 146-7
Lightweight fillers, particulate 249 transfer 610
Limitations, aramid fibers 207 Local molecular order, see LMOs M60Jtape
LIMS 447, 453 Localised, microbuckllllg 800 mechanical properties 718
Lilldau 59 Logging 3 strengths 718
Lille source 453 Long Duration Exposure Facility; see M401/F854
Linear LOEF CTE 704
bendlllg 745-7 Long-term longitudillal 699
1094 Index

transverse 699 graphite 464 920-1

longitudinal tensile modulus 699 instrumented 468 oil platforms 924-5, 925
Poisson's ratio 699 low cost materials 464 phthalic anhydride 916
shear modulus 699 metallurgy 429 piping systems 926
transverse modulus 699 pressure on 468 pressure hulls 922
volume fraction 699 puller 428 propulsion shafting 925
m-phenylenediamine (MPD) 55, 76 removable 464-5 reinforcements 916-17
McDonnell Douglas removal 466, 467 resins 916
Bea AV-8B, material usage 1032 selection 465 transfer molding 919-20
C-17A, composite applications spider/plaster 466 shipboard armor 925
1033 surface 591 small boats 921
F/ A-18E/F, material usage 1031 table rolling 429 sonar domes 922
MD-11, composite structure 1036, tensometric 468 Spectra 917
1037 water-soluble sand 465-6 vacuum bag processing
Machinery, marine applications Manufacturers 919-20
925-8 continuous fibers 162 vinyl esters 916
Machines ceramic 309 bacteria 919
caterpillar type 502 fibers 158 construction 994
dual mandrel stations 477, 478 PAN-based tow 198 fouling 810-11
fiber placement 478 pitch-based tow 199 laminates 920
light grit-blasting 677, 677 staple 162 submersibles, design
tools 363-5 whiskers 158, 162 requirements 713
MachWning 363-5,364,596-608 Manufacturing 6 Market growth, land transportation
abrasives 606 aramid fibers 203-4 905-6,906
advantages 596 boron 156-7 Mass transit applications 914
aramid fibers 222 ceramic fibers 156-7 Master models 563-6, 592
characteristics 596 deficiencies, adhesive joints 611 fabricating materials 563
closed loop 364--5, 364 extended chain PET fibers 223-4 storing 563
drilling 602 fibreglass 132 Mat 164
electric discharge 605 high silica 156-7 chopped strand 138, 155
graphite composites 599 options 22 continuous strand 138
Kevlar 600 quartz 156-7 fiberglass 137-8
linear cutting 602 SMC 381 surface 138
milling 602 Marine Matched die, molding 361
PET 233 applications 916-28 Materials
requirements 602 adhesives, paste 918 anisotropic 687, 688
turning 602 America's Cup yachts 922 cost, GFRP 1006
ultrasonic 605-8 blistering 918, 918 definition 731
Macromechanics, of laminae 192-4 buoys and floats 927--8 density, X-ray backscattering 846
Maglev 979 cables 927 description 731, 731
train guideways 943 concrete forms 925 difficult to roll 430-2
Magnamite 1054 control surfaces 924 fabrication 786
Magnesium oxide, particulate fillers cores 917-18 isotropic 687
250 decking 925 monolithic 460
Magnetic transparency 983 diving equipment 926-7, 926, preliminary design 715-22
Mahogany 927 procurement costs 722
CTE 558 environmental effects 918-19 properties 736
density 558 fabrication 919-20 definition 709
thermal conductivity 558 fallings 923-4, 923, 924 T50 graphite 724, 724
Maleic anhydride 35 filament winding 920 T50/F584 epoxy 726, 726
isomerization 35 future developments 928 property, equations 760
melting point 58 hand lay-up 919-20 quality 780
Mandrels 429,459,464--6, 46S impregnators 919-20 quasi-isotropic 687
basic requirements 465 isophthalic polyester 916 selection 45, 712
collapSible 466 Kevlar 917 applications
cone-shaped 482, 483 large power yachts 921-2 aerospace 1009
dissolvable materials 465 machinery 925--a construction 982
electroformed nickel 591 microbial degradation 919 tools 559
fusible materials 465 mine counter measure vessels specification 722
 Index 1095

table rolling 429-30 resistant, gloves 829 density 815

weights 1027 Mechanics 18,18 impacts
wet lay-up repairs 867 Median lethal dose 824 angle of 815
for winding 458-9 Medical applications 943-4, 944, space 814-15,814,815
Matrix 28 957--M Whipple-type shield 814-15,
adhesion 215 coatings 957 814
cracking 319,798 fillers 957 size of 815
rate of 806 polylactids 957 velocity 815
cracks polyorthoesters 957 Methacrylic acid 40
NDE 854 strength retention 957 Methanolysis, degradation 888
X-ray imaging 844 Melt Methylene chloride 812
crosslinking 460 extrusion, PET 224 4,4'-methylenedianiline (MDA) 55,
definition 378 flow index, see MFI 76
formulation 101-5 impregnation 117,118 Methyltetrahydrophthalic
materials 356--8 infiltration 315 anhydride, melting point 59
ceramic composites 311-12 pumping, extrusion 534-5,535 MFI 529
polyester 28 spun, Vectran 235 Mica 245--6
systems 4--6 stretching 529-30 fillers 527
moisture absorption 5 Melting Microbial degradation, marine
thermosetting 28 aramid fibers 205 applications 919
transfer molding 315 extrusion 534 Microbuckling 799-800, 800
vinyl ester resins 28 point aramid fibers 207
Maximum carbon-carbon composites localised 800
bond 333 Microcracking 323-4,323,797-8,
peel stress 642,642-3,644 ceramic composites 312 798,801,1011-12,1012
shear stress 642-3, 644 4,4'-diaminodiphenylsulfone cyclic loading 801
service temperature 258 (DDS) 55 elastic stress-strain 798
stress PET 225 eliminating 1012
anhydride-cured system 72 recycled NBC 897 GFRP 1006
aromatic system 71 S-glass 24 long term loading 801
Maypole machines, braiding 413, temperature minimizing 1012
414 see also Tm stiffness changes 796
Measuring ceramics 307, 311 Micromechanics 191-2,796-7
displacements 779 Kevlar 206 Microspheres 246, 248
loads 779 Nomex 206 hollow 246
strains 779 Technora 206 particulate fillers 246
Mechanical Teijinconex 206 types of 246
fastening 727 Twaron 206 Microstructures 803
joints, design process 729 Vectran 235 flat-layer 177
pressure 874 Mesophase onion-skin 177
properties injection 176 radial 177
aramid 207-13,208-9,209, pitch, precursors 335,336 random 177
210, 211, 212, 718 Meta-aramids Microtexture
carbon-carbon composites cost of 205 and electrical resistivity 189
344-5,345 crooked chain 205 and tensile strength 188
extended exposure 215 manufacture 203-4 and Young's modulus 189
fiberglass 135 Metal Microwaves
laminates 717 matrix composites, see MMC radome wall 113
M60J tape 718 molten 812 testing 853
P100 tape 718 repairs, aircraft 857 transparency 109
T300 fabric 718 surfaces 871 Migration 577
and weave pattern 152 Metallic Military
strength 111 devices, total hip arthroplasty aircraft
tests 778-92 960 certification requirements
Mechanically joints 613 1023
fastened oxides, fillers 527 components 1030-5
joints 610--63 Metallurgy, mandrels 429 applications 914-15
disadvantages 517 Metals, costs 717 specifications 1062, 1064
repairs 877 Meteoroids 814-15 Milling 596-7
1096 Index

AWJ 604 CE 107 cavity design 442

ball or hammer 886 epoxy resins 29 closing 385,442
climb 596 PET 224 speed 385
disadvantages 597 PMR-1582 construction 447
dust particles 597 resins 440 design 362
Kevlar 600 tooling 561 dual steel 361
machining 602 barrier removal, aircraft 863-4 elements of 541
with polycrystalline diamond contamination, X-ray imaging filling 385,442,444-5
596 858 pressure during 452, 452
recycling 886-7 content 206 simulation 450-1
square pockets 604 debonding 811 graphite-epoxy 590
Mine counter measure vessels 920-1 desorption 817 maintenance 361
Minimizin~ microcracking 1012 detectors 858 making 390,390
Minor damage diffusion, GFRP 1007-8 materials 441-2
aircraft 859 effects 96 metal 441
dent fillers 859 fracture toughness 811 polymeric composite 441
Minor impact resistance, reinforced glass transition temperature 811 preparation 442
composites 387,388 impact resistance 811 release 361,362,428,429-30,514
Mix viscosity, resin 441 prebond 670 problems and solutions 568
~C 27,291-306,944-5 regain, PBO fiber 238 secondary 430
alumina/ carbon 291 resistance, PBO fiber 238 RTM design 441
applications 1005 transverse strength 811 sealing 443
boron/aluminum 299 vs. time, P755/cyanate 1010 shrinkage
by hot extrusion 291 Moisturizers 829-30, 830 nylon 6/6 PCI-glass 899
carbon reinforced 291 Molded surfaces 392 recycled NBC 897
continuous casting 298-9 Molded-in steel 442
cross sections 298, 298 color, reinforced composites 388 selection 393
liquid stage 291 threads 392 stops, compression molds 395
production 291 Molding stresses 392-3
recyclability 944 autoclave 367 temperature control 541
solid stage production 291 autoclave/oven 361 Molecular
solution sedimentation 291 bag 366-72,368,369,370,371 arrangement 133
Mobile storage 945 bleeder ply 584--6, 585 characteristics, rotational
Modulus blovv 529-30,529,532 molding 551
adherends 643 compounds orientation 27
and density 23 applications 945-6 vveight, thermoplastics 116
of elasticity BMI 114 Molybdenum disulfide, fillers 527
aliphatic system 70 compression 127, 365, 384, 545 Moment of inertia, formulas 285
anhydride-cured system 72 enclosed 366 Monobands, sealed 304
aromatic system 71 matched die 361 Monofilaments 164, 307
carbon-carbon composites matched metal compression reinforcements 311, 311
333 378-88 Monolithic
and conductivity 186 negative draft 366 graphite 559,561,562-3
fiberglass epoxies 153 oven/press cure 366 coefficient of thermal
plied-yam 151 preparations 361-2 expansion 558
S-glass epoxies 152 pressure 385, 545 density 558
single-yam 151 bag 367 low CTE567
yam distribution 154 resin transfer 374,492 thermal conductivity 558
quasi-isotropic 12 see also RTM material 460
of rupture, ceramics 312 sheet 374 Monolithicity
5-glass 24 thermal expansion 365-6, 365, loss of 469
transverse 12 366,593 thick-walled structures 456
Mohs ratings 251-2 vacuum 385 Monomers
Moisture 811-12 bag 361,366,577 BMI100
absorbancy waste 374 ratios 79
adherends 630 Moldless construction 257 reactive 34, 37
RTM 440 Molds 441 Montana talc 248-9, 248
absorption 695,811 aluminum 361 Morphology, resins 505
aramids 26, 221 backing up 442 MPDA/MDA, viscosity 56
 Index 1097

~ullite 313,314,318 Neat polymers, shrinkage 541 Nonlinear, bending 747

heat treatment 310 Nepheline 243-4 Nonrecyclables
~ulti-fastener joints 619-24, 620, chemical resistance 243-4 disposal of 899-900
621, 622, 623 oil absorption 244 incineration 899-900
Multi-rowed, fasteners 620 particulate fillers 243-4 landfill 899-900
Multifilament, continuous 28 refractive index 243-4 Nonwoven
~ultilayer locking, continuous fiber syenite 244 fabrics 402, 402
338 Neutral hydrolysis 887 textiles 418-20
~ultiple Neutron radiography 847-9, 848 3-D technology 418
gates 444 field inspections 848 design methodology 419
ply angles 19-20 military aircraft 849 orthogonal 3-D 418,419, 420
stage drawing, PET 224 New York, talc 248-9, 248 processing 418
~ultiwarp, weaving 406 Nextel 309, 309, 310, 316, 403 structural geometry 419
~usical instruments 946 312 Normalization 721
~utagenicity 824 fiber 159 compression strength 721
~WK Nibbling 601 Notation, stacking sequence 619
3-D 409,410 Nicalon 310-11,311,316,318 Notched fatigue behavior 802
LmA system 411 Nickel 249 Novolac resin
MY 0510 (Ciba-Geigy) 53 coating 292 homopolymer property 102
MY 720 (Ciba-Geigy) 53 electroformed, CTE 589 structure precursor 102
Mylar powder 250 supplier 102
film sheet 574 vapor-deposited coating 819 trade name 102
template 487 Nitride, coating 296 Novoltex 418, 418
Nitrogen release, and tensile NR-150
N, N, N', N'- tetraglycidyl strength 182 char yield 82
methylenedianiline 53 Nomex 203-4,269 chemistry of 77, 78
N/C ratio 182 compressive strength 272, 273 density 82
Nadic methyl anhydride (NMA) 57 cores 270 elongation 82
Nanoporous, carbon fiber 183 decomposition temperature 206 flex modulus 82
NASA standards 1063, 1064 density 206 flex strength 82
NASP 972, 972, 974 elongation at break 209 fracture toughness 82
National Aerospace Plane, see NASP filament 206 neat cured 82
Natural composites initial tensile modulus 209 Rockwell hardness 82
cartilage 958 linear thermal expansion tensile strength 82
cortical bones 958 coefficient 206 T 82
dentin 958 long-term use temperature 206 t6ermal expansion 82
wood 958 melting temperature 206 Nuclear magnetic resonance (~R)
Nd-YAG moisture content 206 982
applications 605 plate shear Nylon 6/6 PCI-glass
solid state laser 605 modulus 272, 273 elongation 899
NDE 733,838-55 strength 272, 273 flexural modulus 899
blown core 854 specific gravity 209 flexural strength 899
bondline adhesive voids 854 specific heat 206 glass content 899
condensed core 854 tensile strength 209 izod impact 899
core 854 thermal conductivity 206 mold shrinkage 899
damage inspection 854 Nominal thickness, reinforced tensile strength 899
deep delamination 854 composites 390, 391 Nylon
degradation 838 Non-aerospace, applications 935 modulus and temperature 210,
disbonds 854 Non-functional, fillers 382 211
fiber breaks 838,854 Non-metallic core removal 862 peel plies 673
foam adhesive voids 854 Non-symmetric imprint 682
foreign material 838, 854 bending, testing 783 tensile strength and temperature
inhomogeneities 838 deformation, cylindrical shells 210,211
matrix cracks 854 754-5 weld line strength 544
porosity 854 laminates 752
records 733 Noncomposites 166 O-rings 443, 953
shallow delamination 854 Nondestructive nitrile rubber 443
test methods 733-4, 734 evaluation, see NDE Observed life, AS-4/PEEK(APC)
water intrusion 854 testing, and inspection 1059 806,807
1098 Index

Octyl, decyl glycidyl ether blend, 4,4'-oxydialinine (ODA) 75,76 shape, and tooling 589-94
viscosity 61 slippage, during cure 432
Off-axis tension, testing 781 P75S/cyanate, moisture vs. time Particle
Oil 1010 accelerators 976-9,977,978
absorption Pl00 tape size
nepheline 243 mechanical properties 718 clay 244
number, particulate fillers 251 strengths 718 distribution, rotational
containment, Spectra 233-4 p-t-butyl phenyl glycidyl ether, molding 551
and gas viscosity 62 Particulate fillers
applications 946-7,947 PA6, 6, water absorption 127 abrasion 252
industry, pultrusion 521 Packages alumina trihydrate 245
platlo~ 924-5,925 style of 493-4, 493 aluminum oxide 250
One-dimensional weight 494 antimony oxide 244-5
consolidation equation 584 wound 493 calcium carbonate 243, 247
flow, compression molding Packing carbon black 245
583-4,584 fraction 531-2 clay 244
stress analysis 627 interfiber 403 common 243-6
One-sided, pulse-echo testing 842 powder processing 313 cost 242
Onion-skin, microstructures 177 pressure 539 definition 242
Opaque, aramid fibers 221 Paintsurface~Unspection 839 electrically conductive 250
Open core evacuation, aircraft 865 PAN 25,169,171-3,172 elongation at failure 252-3
Open hole coupon strength, joints anisotropicity 185 end uses 253
626 cyclization 172 feldspar 243-4
Open packing 403, 403 cyclized 179, 181 fire retardants 249
Open unloaded holes 614 dry spinning 175 flakes 248
Openings 373,373 large diameter 180, 181 orientation 252
Operational requirements 737 modified polymers 175 graphite 245
Optical properties molecule structure 172 hardness 251
aramid fibers 215 spinning 175 health hazards 252
fiberglass 135 stabilization 172, 177-80, 178, 179 high
Organic solvents, thinning 65 stabilized, thermal degradation density 249
Orientation angle, yarns 412 181 hardness 249-50
Orifice plates 495 wet spinning 175 thermal conductivity 250
Orthogonal nonwoven fabrics 163, PAN-based fibers hot spots 242
164 density 170 injection molding 242
Orthopedic applications 95~ elastic modulus 170 kaolin 244
FRP 959 fracture elongation 170 lightweight fillers 249
hydroxyapatite (HA) 958 oxidation 818 low density 243
PMMA 962 tensile strength 169, 170 low friction 250
Orthophthalics 35,39 PAN-based tow magnesium oxide 250
polyester, blistering 918 manufacturers 198 mica 245-6
Orthotropic trade names 198 microspheres 246
layers 742 Para-aramids natural 246
nonzero stiffness 742,743 available 203 nepheline 243-4
thermal coefficients 742 creep 212 oil absorption number 251
plates, critical load 751 fatigue 213 organic 243
Outgassing stress rupture 212 packing 252
measuring 817 Para-phenylene terephthalamide 26 production 246-8,247
and vacuum 817 see also aramids shrinkage 242, 252
Oven/press cure, molding 366 Parallel axis theorem 694 silica 244
Ovens, with thermocouples 429 Part specific 243
Overheat, environmental exposure consolidation 386 surface
860-1 ejection 385 properties 250-1
Overshoots 531 geometry 526,530 treatments 250
Oxidation inspection 733 synthetic 246
PAN-based fibers 818 methods used 733-4 talc 246
resistance 818 removal 446 thermal conductivity 242
carbon-carbon composites hydraulic ejection 446 toxicity 252
333 system 541 types of 242-3
 Index 1099

use of 242 Penetration 814 scrap 36

Particulates 307 Percolation 547 seawater resistance 233
flow, d. fiber flow 530 matrix 548 shear modulus 120, 120
inhaling 827 Perlite 249 sources of information 232
rrregularshaped 249 low cost 249 specific
loading levels 528 Permeability 399-400, 1000 gravity 224
radiation 816 axial 579 modulus 224
Parts, RTM process 762 bleeder 581 strength 224
Pastes, adhesives 275--6 composite layer 581 stiffness-to-weight ratio 223
Patterns 362-3 and fiber volume fraction strength retention 227
control 463-4 399-400,400 strength-ta-weight ratio 223
PBI 237,237 preforms 438, 439, 449, 578-9 structure 224,224
chemical resistance 237 Permeation barrier 918 surface treatments 231-2
chemical warfare applications Permissible exposure limits, see PEL sutures 234
237 Peroxides 382 use temperature 225
solvent resistance 237 Personal protective equipment 828 UV
structural formula 237 PET 35,120 resistance 224,230,233
PBu\ 203,203,205 ability to float 225 stability 233
PBU\-based fibers, SVM 210 abrasion resistance 224, 226, 233 Phase locked loop, holography 852
PBO 26-7 anisotropy 224 Phenolics 504-5
elongation at break 235 applications 230,233-4 adhesives 255
fiber 236,237-8,237 artificial ligaments 234 compressive strength 258
specific gravity 235 availability 231 density 258
structural formula 237 axial orientation 224 health effects 834
tensile modulus 235 chain folding 224 maximum service temperature
tensile strength 235 characteristic temperatures 121 258
PDI fibers, sources of information chemical resistance 224, 230 polypropylene, density 258
236 cleaning 230 pulling 504-5
Pedestrian bridges 991-2, 992, 993 creep resistance 227 pultrusion grades 505
PEEK 31,120 crystallinity 223, 224 resins 268, 504
glass transition temperature 812 degradation 889 resistance to fire 505
Peel design 232-3 shear modulus 259
plies 668, 669 dielectric constant 224 shear strength 259
bonded joints 669 durability 233 SMC 382
dry 672 dyeing 233 spall liners 914
evaluation 670 energy to break 227 tensile strength 258
imprint 678 exposure to moisture 227 thermal conductivity 259
nylon 682 fatigue resistance 230 Phenoxides
sanding 680 fibers, extended chain 202 allyl functional 101
nylon 673 filament 225 propenyl functional 101
polyester 681, 681 filter cloths 233 Phenyl glycidyl ether (PGE),
preimpregnated 672 forms 231 viscosity 62
removing 683 gel-spun 225--6 Phosphoric acid anodized,
silicon transfer 670 handling 233 aluminum foil 820
surface preparation 683, 870 high strength 233 Phthalic anhydride 35
strengths hybridization 233 chemical resistance 35
adhesives 274-5 hydrophobic 230 marine applications 916
GFRP 1006 inert 230 melting point 57
stress 629,631,637-8,637,638, machining 233 thermal stability 35
647,655,662 melt extrusion 224 Physical properties
adhesives 628 melting point 225 aramid fibers 205-7,206
eliminating 644-5 moisture absorption 224 graphite/epoxies 4
lap joints 637-8, 637, 638 multiple stage drawing 224 Pick-up trucks 936-7
linear variation 643 non-conductive 231 Pigments 516
maximum bond 642 non-woven 234 inorganic 516
reducing 644-5 processing temperatures 233 titanium oxide 516
tapering 646 recycling 883 zinc sulfide 516
PEKK 31 for reinforcing 234 Pins, FRP 959
PEL 825,826 resin penetration 233 Pipelines 466
1100 Index

rehabilitation 948 Nomex 272,273 PMR-ll

Piping 936 shear strength d. PMR-15 90, 93
systems 926 honeycomb 264-5,266-7 chemistry of 80-1,80
Pitch 175-7, 176 Nomex 272, 273 processing conditions 84-5
fibers 169,352 stiffness 622 PMR-ll-50
graphitized 183 thickness 790 flex strength 92
flow through spinneret 177 structural hierarchy 778 PMR-ll-30 90
forming coke 173 Platelets 307 reinforced 90
mesophase 176 Platens, heated 499 Poisson's ratio 19,698
precursors 173 Platforms 989 ceramic composites 312
carbon yield 197 fiberglass 989 IM6/epoxy 699
treatments 173-5,174 well bay 989,990 M401/F854 699
pretreatments 173, 197 Pleating 547 Polar backscatter, ultrasonic 843
softening 197 Plied-yam Polar winding 461-2,462
spinning conditions 175-7, 176 compressive strength 151 Poly (p-phenylene benzobisoxazole)
stabilization 177 count 151 seePBO
stirring 177 flexural strength 151 Poly (p-phenylene benzobisthiazole)
Pitch-based fibers 170 modulus of elasticity 151 see PBT
density 170 tensile strength 151 Poly-2, 2' -m-phenylene-5, 5'-
elastic modulus 170 warp 151 benzimidazole, see PBI
fracture elongation 170 weave 151 Poly-m-phenylene isophthalamids
tensile strength 170 Plugs 362-3 203,203
Pitch-based tow Ply Poly-p-phenylene terephthalamide,
manufacturers 199 alignment 689 seePPTA
trade names 199 angle 699-700 Poly-p-phenylene-benzimidazole-
Placement, on mold 384-5 boundaries 481-2,481 terephthalamide, see PBIA
Plain locating, templates 574--5 Polyacrylonitrile, see PAN
strain 801 orientation 9,10-11,374 Polyaluminoxane 159
weave 140,145,405,406 interspersing 19 Polyamide 6,6, characteristic
biaxial 399 properties 8,737 temperatures 121
fiber volume fraction 408, 408 reinforcement 730,730 Polyamide 12, characteristic
triaxial 399 sequence, table 732 temperatures 121
Planar flaws 840 shape 481-2,481,482 Polyamide-imides
Plane generating 481 chemical name 121
source 453 stacking sequence 692 processing temperature 121
stress, stiffness 689 strength 745 structure 121
Plasma assisted CVD (PACVD) 316 Plycosite 255 suppliers 121
Plasma spraying 300
Plaster masters 564-{;, 564
follow board method 565, 565
PMC
decontamination 863
evacuation 863
r:.
T 121
121
trade name 121
NC machining 566 PMDA/ODA polyimide 76 Polyamides
sweep method 565-6 PMMA characteristic temperatures 121
template method 564-5 fatigue resistance 959 health effects 832, 833
Plasters, breakout/washout 465 orthopedic applications 962 Polyaminoamides, health effects
Plastic faced plaster 567, 573 PMR-1580 832,833
preparing 573 air aging 89 Polyarylene ether
Plastic zones 654 chemistry 79-80, 79 chemical name 121
Plasticating 538 compressive strength 82 processing temperature 121
extrusion 534, 535 density 82 structure 121
injection molding 538, 538 elongation 82 supplier 121
Plasticization 107
moisture 112
Plastics, recycling 883
fracture toughness 82
laminates 88, 89
moisture absorption 82
r:.
T 121
121
trade name 121
Plate neat cured 82 Polybutylene-terephthalate,
elements 621 processing 84 characteristic temperatures
reinforced, carbon/aluminum tensile modulus 82 121
304 tensile strength 82 Polycarbonate, weld line strength
shear modulus T 82 544
honeycomb 264-5,266-7 ~ermal expansion 82 Polycrystalline diamond, milling
 Index 1101

596 Polymers shear strength 259

Polyester 28, 34--9 composites 378 thermal conductivity 259
characteristic temperatures 121 properties 385-6,386 viscosity 530, 531
chemical name 121 resin 382 weld line strength 544
cure rate 509 compound grades 525 Polysulfone
density 505 crystalline 118, 120 see also PSU
elongation at break 505 economics 525 chemical name 121
flexural modulus 505 extrudate swell, injection flexural modulus 546
flexural strength 505 molding 540 processing temperature 121
glass fiber reinforced 28 fire retardant 42-5 structure 121
heat distortion 505 frictional coefficients, injection supplier 121
linear 34, 34 molding 540 T 121
modulus and temperature 210, inclusion of particulates 528 r:, 121
211 major processes 526 trade name 121
peel plies 681, 681 matched metal dies 546 weld line strength 544
polyarylate 235 matrix composites Polytetrafluoroethylene, see PTFE
processing temperature 121 see PMC Polyurethane
reinforced 442 grinding 600 coatings 813
resins, shut down 515 matrix lamina, fiber fracture 799 compressive strength 258
shrinkage 507 melt compressibility, injection density 258
structure 121 molding 540 health effects 834
supplier 121 melt fracture potential, injection maximum service temperature
tensile modulus 505 molding 540 258
tensile strength 505 polymerization shear modulus 259
and temperature 210, 211 condensation 75-6 shear strength 259
T 121 shrinkage at 515 tensile strength 258
r:, 121 pressure dependent shrinkage, thermal conductivity 259
trade name 121 injection molding 540 Polyvinyl chloride
unsaturated 35 processes 529, 532 compressive strength 258
Polyether ether ketone, see PEEK unit costs 532 density 258
Polyether ketone ketone, see PEKK processing 525 maximum service temperature
Polyetherimide, flexural modulus properties, blow molding 550 258
545,546 rheological concerns 525 shear modulus 259
Polyethylene selection 525 shear strength 259
recycling 883 semi-crystalline 534 tensile strength 258
terephthalate, see PET processing temperature 120 thermal conductivity 259
Polyglycidyl ether of o-cresol- shear sensitivity, injection Pontoon pier 995
formaldehyde Novolac molding 540 Poor bonding, adhesive joints 611
melting point 52 thermal properties, injection Porosity 90,300,630-1
viscosity 52 molding 540 bond layer 656
Polyglycidyl ether of phenol- thermoplastic, adducts in 526 NDE 854
formaldehyde Novolac thermoset 34 preform 439
viscosity 52 viscosity, injection molding 540 X-ray imaging 844
Polyimide Polymethylmethacrylate, see PMMA Ports and harbors 949
addition chemistry 78-81 Polyolefin 549 Post-curing 84--5,385,446-7,589
aromatic 75 characteristic temperatures 121 Post-fabrication 732-3
binder solutions 76 Polyorthoesters 957 Post-processing 446-7
chemical name 121 Polyoxymethylene Post-tensioning 986, 987
compression molding 545 cupping 541 anchors 987
condensation 78 shrinkage 541 Pot life 66,510,510
epoxy resins 6 warping 541 anhydride-cured system 72
precursor solutions, see binder Polyphenylene sulfide, see PPS NMA 66
solutions Polypropylene resins 440,513
processing temperature 121 see also PP Potassium
structure 121 and acrylic acid 251 alurninosilicate 245-6
supplier 121 characteristic temperatures 121 metaphosphate 957
T 121 mica-reinforced 253 Powder
r:,121 phenolics 258 alloys 300
trade name 121 shear modulus 259 consolidation of 343
1102 Index

impregnation 118 disadvantages 438 tooling 562

processing drapability 438 unidirectional 118
ceramic composites 312-14, edge definition 438 waste 885
313,314 elastic constant 449 hazardous 885
packing 313 engineering parameters 421 wet 355
sintering 343 fabricated 438 Pressing
Powdered amorphous fiber distribution 438 temperature, SiC I A403 315
thermoplastics (TPs) 105 fiber orientation 421 time, SiC I AlP3 315
Power heating 458 Pressure
poles 948-9,948 infiltration 299 compliant ring, testing 782
shears 426 permeability 438, 439, 449, 578-9, consolidation 577
trowels 461 578 damage 874,875,876
Powertrain applications 911-12, 911, porosity 439 gradient, infiltration 338
913 prefabricated 438 hulls, marine applications 922
PP and granulated SMC processing parameters 421 pulsation, infiltration 338
elongation 897 Rlld 436-9,437,438 vessel, acoustic 472, 472
flexural modulus 897 textile 401-2 vessels 471
flexural stress 897 thickness 438-9,438 X-ray backscattering 846
izod impact 897 uncured 456 PRESTEK 992
tensile modulus 897 winding 458 Prestressed concrete 985-6
tensile stress 897 Preforming 397-422,448 anchorage 985-6
PP IPCI-glass 3-D 398 bridge deck 986, 986
glass content 898 formability 398-9 bridges 987
strain at break 898 Preliminary steel tendons 985
tensile modulus 898 design 710-11,715-23 Preventive maintenance, composites
tensile strength 898 definition 710 880
yield stress at break 898 materials 715-22 Pricing
PPD 81 support data 711 aramid 216, 716
PPS 31 materials, aircraft stabilizer 711 Dyneema 232
weld line strength 544 Preload relaxation 729 graphite fiber 716
PPTA 203,205 Prepegging 6-7,7,32-3 Hoechst Celanese 232
chemical stability 213 advantages of 7 Spectra 232
hydrogen bonding 204 hot melt 7 Primary Adhesively Bonded
structural formula 203 roving 7 Structures Technology
Prebaking 871 thermoset matrices 7 (PABST) 667-8,669
Precarbonization 180 Prepolymers 105 Principle strengths 803
Preceramic polymer chemistry 157 Prepregs Processing 64-9
Precured patching 867-8 aerospace industry 885 composites 879
limitations 867-8 availability 82 glass fibers 24
Precursors 335 braiding 418 parameters, preforms 421
fiber, pyrolysis 157 co-curing 868, 870 polymers 525
materials 173 combining 117 speeds
mesophase pitch 335, 336 composite tools 566, 567-71 d. pull loads 506, 506
PAN 335,336 consolidation 117 pultrusion 514
pitch 173 definition 425 technology, braiding 413-15
polymeric 157 dry 425,430 temperature
rayon 335 facing material 255 PET 233
treatments, pitch 173-5, 174 lay-up 576 polyamide-imides 121
Predicted life, AS-4/PEEK(APC) low tack 425 polyarylene ether 121
806,807 over metal substrates 870 polyester 121
Predicting ply kits 374 polyimides 121
durability 796 repairs 868-70,869 polysulfones 121
joint strength 615 resins 352 thermoplastics 115,525-52
peak stresses, joints 613 matrix 577 time, terephthalic resins 36
Preform reuse 884-5 window 176
advantages 438 shredding 885 Procurement costs, materials 722
ceramic 314 suppliers 109 Producing, carboni aluminum 303
comers 438 tapes 430 Production
designing 438 thermoplastics 357 costs, aircraft 1024
 Index 1103

MMC 291 additives 515-18 definition 888

particulate fillers 246-S, 247 advantages 489 precursor fiber 157, 159
Profile extrusion 532 applications 518-21 recycling 896--8
Programmed winding 469-70 basic process 489 tooling 341
Propeller shafts 911 cf. extrusion 488-9 yields from 895
Properties cf. pull forming 490, 491 Pyromellitic dianhydride (PMDA)
aramid fibers 205-15 civil engineering 519-20 75,76
boron 161 and compression molding 490
casting, CE 108 construction 519-20 Quadratic failure 695-7
ceramic fibers 161 continuous production 489 criteria 707-8
elastic 194-5 creel 492-5 Qualification tests 1060
high silica 161 curved 489-90 Quality controls, composites 880
Kevlar 205 definition 488 Quartz
lamina material 760 dies 497 applications 163, 165, 166
polymer composites 385-6, 386 electrical equipment 518-19 drawing 157
quartz 161 epoxy 511 fibers 24-5
Technora 205 fibers used 503 fused 163
thermoplastics 115 fillers 516 forms 161-3
Twaron 205 FRP 47,517 fused rods 157
using recyclate 892 hammer handles 491 manufacture 156--7
Vectran 235-6 history 488 properties 161
Proprietary specifications 1066--7 injection 497, 498 Quasi-~tropic
Propulsion labour requirements 489 laminates 9-11,697
shafting 925 machine 488 materials 687
systems 1041,1041 oil and gas industry 521 modulus 12
Prostheses 963 pigments 516 Quaternary ammonium salts 382
Protective coatings 292 processing speeds 514
composites 879 products 489
R-glass 134
Protruding head, fasteners 624, 624 pull-winding 496--7
Race tracking 443
Pseudoisotropic laminates, CMEs resins 504
pattern 461-2
1011 impregnation 495
winder 462
PTFE 362 shapes 989
Racing yachts 353-4
see also Teflon shut down 515
Rackets 1049-SO, 1050
compression molding 545 sloughing 514
braiding 1050
velocity cf. temperature 843-4, sporting goods 1045-6
compression molding 1049
844 start-up 514-15
filament winding 1050
Pull surface finish 514
injection molding 1050
forming 490-2,490 thermoplastics 116
resin transfer molding 1050
automation 492 three zone model 505, 505
Radar
cf. pultrusion 490, 491 transportation 520-1
signals, trans~sion 472
curved 490,491,491 tubular structures 496--7
transparency, foam 256
straight 490,491-2 vertical 495-6
Radial
loads 512 wastage 489
microstructures
cf' processing speed 506, S06 and winding 496--7
folded 177
epoxy resins 510 windmills 521
with wedge 177
winding 489,490,496--7 Pure bearing load, joints 626
Radiation 3,816
Pullers Purging 515
Bremsstrahlung 816
continuous belt SOl Putties, adhesives 275-6
ionizing 816
intermittent 501 PVC
particulate 816
Pulling SOl-2,501 foam 257
solar 816
continuous SOl crosslinked 917
thermal 816
Pulp molding, short fibers 337, 338 linear 917
ultraviolet 817, 817
PuIse-echo testing Pyrex 318
Radio frequency
advantages 842-3 Pyrolys~ 159,182,888
C-scan 842, 842 carbon fibers 335
seealsoRF
heating 499
one-sided 842 carbon-carbon composites 341,
Radius of curvature 459
ultrasonic 842-3, 842 342
sheets 186, 187
Pultrusion 17-18,33,117,488-521, carbonized organic composites
Radius of tw~t 459
529,532 340
1104 Index

Radomes 1024 polyurethane composites 893 recycling 388

composites, quartz reinforced pyrolysis 896-8 ribs 389,391,391
110 reinforced composites 388 shell and plate construction
Rail shear, testing 783 5~C 890-3 389,389
Railroad 5weden 887 size and shape 388
applications 912-14 thermoplastics 893-5 spray-up 355
rolling stock 949-50 Redux 255 structural requirements 388
Railways 950 bonding 667, 669 surface quality 387
Random, microstructures 177 Re-entry temperatures, spacecraft concrete 983-8
Rapid cure, resins 440 816 beams 1000
Rate of relaxation, water absorption References cracking 1000
811 E-glass 774-7 E-glass 984
Ratio, glass-to-resin 150-4, 153 graphite epoxies 774-7 steel reinforced 983
Raw materials, falling cost 838 Kevlar 774-7 epoxies 442
Rayon 25 lamina property 772-3,774-7 polyester 442
RD-1 (Ciba-Geigy) 61 52-glass 774-7 reaction injection molding, see
RD-2 (Ciba-Geigy) 62 Spectra 774-7 RRIM
Reaction efficiency, atomic oxygen Reflectors 1015 Reinforcement
814 ACTS 1017 CE 111
Reactive rubber, tougheners 106 Refractive index circumferential 461
Reactivity, resins 440 feldspar 243-4 composites
Reaming 598 Kevlar 215 continuous unidirectional
Rebars 984--5, 984 nepheline 243-4 318--20,318,319
fiberglass 999 Regrind RIM discontinuous 320-1, 320
placement 985 density 895 configurations 461-4
pultruded 999 elongation 895 continuous 308--11,309,310,502,
Recreation, design requirements 713 flexural modulus 895 762
Rectangular plates 749-53,749,751 hardness 895 definition 378
in-pane loading 750-1, 751 tensile modulus 895 E-glass 383,916
Recyclability, MMCs 944 tensile strength 895 geometry 797
Recycled Rehabilitation glass fibers 378,383
materials 889-99 bridges 988 graphite 353
NBC 897 building construction 988 heat cleaned 898--9
PET 894 chimneys 988 helical winding 461
phenolic 894 concrete 988 hoop winding 461
PP 895,896 steel 988 Kevlar 383
S~C 890,892,893 wood 988 knitted 916-17
Recycling Reinforced local 730,730
see also reuse, disposal composites 712, 713 marine applications 916-17
automotive industry 901-2 applications 386-8 materials 832, 835, 835
B~C 890-3 carbon fiber 358--9 RN 435
building construction 901-2 chopped-fiber 355 monofilament 311, 311
electrical parts 902 continuous-fiber 355 multi-directional 718
environmental aspects 902-3 continuously 358 plies 730, 730
Europe 884 corrosion resistance 387 polar winding 461-2,462
Germany 887, 902 corrugated configuration 389 52-glass 383
glass-filled PP 893 dimensional stability 387 spatial 471
granulation 886-7 edge stiffening 390,391 specific heat 533
heat cleaning 889 edge turning 390 thermoplastic polymers 526
history 883-4 electrical resistance 387 three-dimensional 456, 801
Japan 887,888,903 high strength 387 two-dimensional 456
legal aspects 902-3 inserts 392 types
mechanical 886-7 light weight 387 fibrous 795, 7%
milling 886-7 minor impact resistance 387, unidirectional 438
organizations 903 388 volume 510-11,510
PET 883 molded-in color 388 woven 361,916-17
phenolic composites 893 nominal thickness 390,391 Reinforcing
plastics 883 non-structural requirements bars, see also rebars
polyethylene 883 388 ceramics 307-11,308,308,309,
 Index 1105

310,311 marine laminates 920 viscosity 64,440,578,579,579

load carrying capabilities 307 minimizing 673 and cure cycle 354, 355
Relaxation, environmental effects epoxies 382, 504 volume fraction 478
1000 and fibers 425 weather resistant 813
Release agents 362 flow 578-80 Young's modulus 440
Releases 429-30 consolidation 576 Resistance
fluorocarbon 361,362 continuity condition 582-3 chemical 39
molds 361, 362 Darcy's law 578 to fire, phenolics 505
silicon 429-30 gelation 499 water 39
Remaining high performance 122 Resonance, ultrasonic 843
life 804-7,804,805,806,807 high temperature 818 Respirators, filter 830
strength 797,804-7,804,805,806, impregnation 495 Respiratory protection 830
807 ingredients 513 Restoration of coatings 867
estimating 797 injection 117 Resulting strain state, laminates
loading history 805 molding 122 693-4
normalized 805,805 system 433,434,440-1 Reticulating films, adhesives 276
predictions 806 thermoplastics 116 Reusable bags, silicone rubber 371,
Removal of peel ply 631 isophthalic polyester 504 371
Repairs matrix 504 Reuse
adhesives 869 prepreg 577 see also recycling, disposal
advanced fibers 877 mixing 513-14 appliances 901
aircraft 857-80 temperature 441 automobiles 900-1
autoclave 868 viscosity 441 carbon fiber 884
heat application 874-7 moisture absorption 440 cutting 885
heating blankets 869 morphology 505 and disposal 883-904
honeycomb panels 871-7 penetration, PET 233 glass fiber 884
joint preparation 871 phenolic 504 hybrid composites 883
mechanically fastened 877 polymerization 499 prepregs 884-5
metal bondments 871-7 potlue 440,496,513 sheet molding compound 883
methods 880 pultrusion 504 shredding 885
non-autoclave 868-9 rapid cure 440 technologies 885-9
prepreg 868-70,869 reactivity 440 RF preheating 513, 513,515
resins 869 recycling 515 Rheology 527-31
skill requirements 878 removing excess 576, 673 definition 527
standardization 878 repairs 869 Ribs
surface preparation 870-1 replenishing 513 designing 391
technical training 878 safety 831-2 geometry 391
vacuum pressure 869 selection 45,440, 504 mold making 390
Repeatability 8 solvent based 431 parameters 743
turning 603 T point 440 reinforced composites 391, 391
Representative volume 796, 804, 804 ~ermoplastic Rigid
discontinuous 804 matrix 496 adherends 633,633,634
Reproducibility 3 mechanical properties 122 fiber 802
Reproductive toxicity 824 toughness 440 tool, thickness 450,450
Residual transfer Rigidity, and flow 67
strength, AS-4/PEEK(APC) 806, molding 17,374,492 Ring specimens
807 see also RIM bending 785
stresses, control of 470, 472 marine applications testing 782, 784
Resins 382 919-20 Rings
bath life 496 rackets 1050 compression testing 782, 788-9
bleed sequence 367 transverse hoop wound 789
catalyzed 507 modulus 1056 thick-walled 789
changing 497 tensile strength 1056 thin-walled 789
chemistry 99-1,382 uncatalyzed 506 Risk 823
comparing 505-7, 505, 506 unsaturated polyesters 382 definition 823
consumption 513 used Rivets, aluminum 668
content 425 RIM 440 RK, properties of 170
by volume 503 selecting 440 Rocket
fibreglass epoxies 153 vinyl esters 382, 504 motor
1106 Index

cases 23,465 heating/ cooling 441 resins 831-2

X-ray backscattering 846 injection pressure 433 Saffil 159
design requirements 713 low cost parts 433 grades of 155Hi0
nozzles, carbon-carbon 712 moisture absorbance 440 Sailboats, design requirements 713
Rockwell hardness, NR-150 82 mold sealing 441 Same materials, sandwich structures
Rods, testing 784 optimum viscosity 440 744
Roll reinforcement materials 435 SAN, weld line strength 544
fomting 117, 289 resin 440 Sand bags 874
wrap process, sporting goods Tl!.446 Sanding 672,672,678-80,679,683
1045 sporting goods 1046 peel-ply imprint 680
Roller press 426, 427 tensile elongation 440 scuff 679
Rollers 461,495 tensile modulus 440 Sandwich
Rolling vacuum assisted 433, 453 adhesive flow 288
carboni aluminum 303 viscosity 440 beam type 281
compaction 477 Rubbery sheet deformation 526 buckling 282
rate of 302 Rule of mixtures 150, 318, 588, 697 construction 694
solid phase 302 Runner system 444 aramidfibers 221
solid stage production 291 conversion to 255
table 427,427 52-glass 134 core
in vacuum 300-6, 301, 302, 305 compression properties 360 crushing 282
temperature conditions 300 constituent properties 766 selection 276, 278, 279, 284
Rolls and air shafts 950-1,951 cost of 917 size 288
Rotational molding 529, 532, 550, elastic properties 764-7 deflection 284
551 references 774-7 limitations 281, 284
crosslinking 551 specific gravity 383 design 281, 284
crystallization kinetics 551 strength properties 764-7 compressive modulus 281
molecular characteristics 551 tensile failure 383 core 281
particle size distribution 551 tensile modulus 383 facings 281
polymer properties 551 tensile strength 383 notation 280-1
thermal properties 551 thermal expansion 383 designing 276-80,277,278,279
zero-shear viscosity 551 5 twist, yarn 138 dimpling 282, 284
Router speeds 289 5-glass 23, 24, 425 fabrication 276
Routing 601 based, rovings 147 face wrinkling 282, 284
Rovings 163, 164 density 503 facing failure 282
carbon fiber 502 elongation at break 503 high temperatures 288
for chopping 146 epoxies 152 manufacturing 287-9,289,290
continuous glass 492 fibers, stress rupture 212-13,213 modes of failure 282, 282
fiber 492 filament strength 24 shear crimping 282
fiberglass 136 melting point 24 skin materials 284
glass 502 modulus 24 structures 467--8
harder 146 tensile modulus 503 design 467--8
nonwoven 164 tensile strength 503 different thicknesses 744
5-glass based 147 5-S2-glass elements of 254, 254
single strand 136,146 reinforcement 383 same materials 744
woven 136-7,137,164,917 specific modulus 1049 space vehicles 254
reinforced 137 specific strength 1049 thickness 284
yields 136 tensile modulus 360 transverse shear failure 282
RRlM 907 tension-tension ratio 1049 wall 742,743
RTM 433-54, 762 5-S2-giass, fiber modulus 360 transverse shear defonnation
advanced technology 447 5-triazine ring 99,100 743
advantages 434-5 SAAB JAS39 Gripen, composite Satellites
cavity design 441 applications 1034 central cylinder 375--6
compounds, rheology 101 SACMAstandards 1061-2,1062 design requirements 713
cure cycle 433 Safety Satin weave 405, 406
curing 445--6,445 applications 943-4, 944 Saturated acids 37
cycle times 433 aramid fibers 222 Sawing 601
disadvantages 434-5 epoxy resins 831-2,831 Scarf
gate and vent 441 factor, fonnulas 285 adhesive joints 628, 629
glass transition temperature 440 hazards, composite processes 831 joints 611, 623, 645
 Index 1107

shear stress 649 rail test 789 particulate fillers 242, 252
distributions 647-8, 648, strength polyester 507
649 ABS 259 polyoxymethylene 541
Scarfing 623-4 AFR700B/S2laminates 93 rate 509
Scientific applications 967-80 carbon fibers 295 recycled SMC 892, 893
Screws, injection molding 538 Celion 87 volumetric 507
SCRIMP 919 cellulois acetate 259 Shut down, pultrusion 515
Scuff sanding 679 epoxies 259 Sialon 329
Sealants, composites 879 interlamina 196 SiC/A1P3
SeaIing577 phenolics 259 density 315
composites 302 polypropylene 259 pressing
Seals 953 polyurethane 259 temperature 315
Seawater resistance, PET 233 polyvinyl chloride 259 time 315
Secondary, drilling 598 skinned molded foams 259 SiC/SiC composites 401, 401
Section stress 636,636 Signal-to-noise ratio, ultrasonic 834
failure curves 615-16,615 5-step 649, 650 Silanes 250-1
load, last fastener 620 10-step 649,651 coupling agent 147
stress 615 adhesive 632-7, 633, 634, 635, organofunctional 147
Secured Modular Automotive Rail 636 Silar
Transport (SMART) 914 distribution 636 fracture
Selection process, joining 727, 728 scarf joints 647-8, 648, 649 strength 325
Self-adhesive skins 276 maximum bond 642 toughness 325
Self-screening, aramid fibers 215 peak-to-average 648 work of 325
Semifabricated composite, winding scarf joints 649 Young's modulus 325
468 TEOS 346 Silica
Sensitization 824 test 688 crystalline 249
cross 824, 825 in-plane 196 fillers 527
equations 825 Shearing, weave 439 flocculated varieties 244
Sensors Shearoutfailures 618,619 fumed 247-8
pressure 499--500 Sheet fused 244
temperature 499-500 extrusion 532 natural 244
Separation 597 molding 374 particulate fillers 244
Shallow delamination, NDE 854 molding compound surfaces 244
Shear 394, 395 see also SMC thixotropic effect 244
coupling, coefficient 193 reuse 883 Silicates, fillers 527
crimping, sandwich 282 radius of curvature 186, 187 Silicon
deformation 448 Shelf life, before molding 31 carbide 27, 319
cylindrical shells 756 Shell and plate construction, coating 294
edges 394 reinforced composites 389, CVD 157
failure 195 389 fillers 527
in-plane 783,789-90 Shells, netted-ribbed 473 wheels, grinding 600
interlaminar 790-2 Shims 570,572 nitride, whiskers 161
knife edge 394 Shipboard armor 925 releases 429-30
lag analysis 633-4 Shipments 955 Silicone rubber 365, 366
modulus Short CTE 589
ABS 259 beam shear properties, Spectra reusable bags 371, 371
cellulois acetate 259 226,226 Silver flake 250
epoxies 259 fiber mat, flexural strength 401 Simple twist 138
IM6/epoxy 699 fibers 337,338 Single
M401/F854 699 term exposure limit, see STEL head, pull winding· 496-7
PET 120 Shredding lap
phenolics 259 mobile 887 joints 626, 627, 628, 729
polypropylene 259 prepregs 885 bending deflections 641
polyurethane 259 Shrink factors 557-8 load path eccentricity 637,
polyvinyl chloride 259 Shrinkage 5,179,506,507-9,507, 637
skinned molded foams 259 508 shear, joints 624
properties fillers 508 strand, rovings 146
aliphatic system 70 following gelation 508 strap, adhesive joints 628
aromatic system 71 neat polymers 541 yam
1108 Index

compressive strength 151 grill opening panels 907 Vectran 236

count 151 IZOD impact 386, 386 Sources of pressure 460
filling 151 lower density 381 Space 813-17
flexural strength 151 manufacturing 381 atomic oxygen 813-14
modulus of elasticity 151 molding 384 composite systems 814
tensile strength 151 phenolics 382 debris impacts 814-15,814,815
warp 151 recycling 890-3 meteoroid impacts 814-15,814,
weave 151 regrinding 890,891 815
Size, meteoroids 815 specific gravity expansion 386, shuttle 344,345,347,472
Sizing 802 386 tooling development 1015-16
agent 495 tensile modulus 386, 386 Spacecraft 968-72,969,970,971
fiber reinforcement 435 tensile strength 386, 386 heat sources 816
Skill requirements, repairs 878 thermal expansion 386, 386 LDEF 970-2,970,971,972
Skin SMEs 952 re-entry temperatures 816
core, bonding 917-18 Snowflake P.E. 247, 247 thermal cycling 818
creams 829-30,830 Sodium-calcium-aluminum- truss structure 968-70
grafts 963 polyphosphate 957 Spall liners, phenolic 914
materials, sandwich 284 Soft gauge, machine tools 364 Spallation 814
penetration Soft tissues, composites for %3-4 Sparging 174-5
aircraft 858, 859 Sol-gel technology 156-7 gas 174-5
foil tape repairs 858 Solar Spatial reinforcement 471
protection 828-30 absorptance 816 Spatially sewn structures 456
Skinned molded foam array Special tests 792
compressive strength 258 backing 820 Specific gravity
density 258 panels 353 Armos 208
maximum service temperature radiation 816 Dyneema 225
258 Soldering, carboni aluminum 303 E-gIass 383
shear modulus 259 Solid expansion 386, 386
shear strength 259 conveying, extrusion 534,535 fiberglass 135
tensile strength 258 phase, rolling 302 graphite 383
thermal conductivity 259 rubber, tooling 670 Hoechst Celanese 225
Skybond 76-7, 77 stage production 291 Kevlar 208,383
dielectric strength 95 state laser, Nd-YAG 605 Nomex 209
electrical properties 95 Solution sedimentation, MMC 291 PBO 235
elongation 86 Solvents 812, 835-7, 836 PET 224
flex modulus 86 absorption,anhydride-cured recycled SMC 890, 892, 893
flex strength 86 system 72 S2-g1ass 383
insulation resistance 95 aramid fibers 204 Spectra 225
processing conditions 84 impregnation 117, 118 SYM 208
surface resistivity 95 polar 175 Technora 209
tensile strength 86 removal 159, 175 Teijinconex 209
volume resistivity 95 resistance Tekmilon 225
weight loss 86 PBI237 Twaron 208
Skylab oxygen tank 471, 471 thermoplastics 115, 127 Vectran 235
Slip forming 547, 547 toxicity 361, 362 Specific heat
Slot camera, X-ray backscattering trichloroethylene 64 aromatic system 71
846,846 Sonar domes, marine applications definition 532-3
Sloughing 508, 510 922 fillers 533
Slurry-spinning 159 Sonotrode Kevlar 206
Small boats 921 material 607, 608 Nomex 206
SMC 380-1,890-3,892,893,895-6, non-rotating steel 608 reinforcements 533
897 Sources of information Teijinconex 206
automobile industry 381 Armos 221 Twaron 206
current use 381 Kevlar 221 Specific modulus 715, 716
early applications 380 PBO fibers 236 aluminum 1049
fatigue properties 386, 387 PDI fibers 236 carbon 1049
flexural modulus 386, 386 SYM 221 definition 715
flexural strength 386, 386 Technora 221 E-glass 1049
formulation 384 Twaron 221 Kevlar 1049
 Index 1109

Kulon 305 Spectra Shield 234 laminate 616, 630

PET 224 Spectroscopy notation 619
S2-glass 1049 fluorescence 69 Stackup, laminate pre-bleed 569
Spectra 1049 infrared 69 Stainless steel, thermal data 818
steel 1049 Raman 69 Stamping 545
titanium 1049 Speed sensitivity 463 Standard tests, composites 880
VMN-4305 Speed tape repairs, aircraft 859 Standards, organizations 1060-1
Specific stiffness, aramid fibers 207 Spherical shapes, mament winding Staple 159--60,216
Specific strength 466-7 alumina 159
aluminum 1049 Spherulites 120 alumina-silica 160
aramid fibers 207 Spin process, Technora 213 density 162
carbon 1049 Spinning 175-7 description 162
continuous fiber 162 conditions, pitch 175-7, 176 diameter 162
E-glass 1049 PAN 175 elastic modulus 162
Kevlar 1049 temperatures 175 green 159
metals, vs. composites 1047,1048 Spirit of Australia 354, 376-7, 376, manufacturers 162
PET 224 377 specific strength 162
S2-glass 1049 carboni epoxy mast 377 thermal expansion coefficient 162
Spectra 1049 Split rings, testing 784 trade names 162
staple 162 Sporting goods 1044-52 unfired 159
steel 1049 braiding 1045--6 Static mixers 441
titanium 1049 mament winding 1046, 1047 Steel
whiskers 162 manufacturing techniques CTE 558
Specifications 1045-9 density 170, 558
materials 722 pultrusion 1045--6 dies 426-7,429,430
and standards 1059--67 roll wrap process 1045 thermal expansion 507
Specimen geometry, honeycomb 263 RTM 1046 elastic modulus 170
Specimen width 780 Sports cars 910 fracture elongation 170
Spectra 23,26,223,228,229,761 Spray lay-up, short fibers 337,338 linear thermal expansion 705
attributes 358 Spray metal, tooling 442 rehabilitation 988
availability 231 Spray-up 436,437 selection, mold 393
axial compressive properties 226, Spraying specific modulus 1049
226 ionic 300 specific strength 1049
axial tensile properties 226, 226 plasma 300 tendons, prestressed concrete 985
carbon 359 Springback 557 tensile strength 170
constituent properties 767 GFRP 1013 tension-tension ratio 1049
creep resistance 227,228-9,230, Springer's model 578,581, 582, 585 thermal conductivity 558
233 compacted plies 588 Steering 482, 483
crystallite melting point 26 numerical schemes 586 definition 482
density 503 Sprue and runner 541,542 radii 482
dielectric constant 231 Square packing 403, 403 Stefan-Boltzmann law 850
elastic properties 764-7 Squeeze flow 548 STEL 826
elongation at break 225, 503 SRlM433 Step lap
fiber type 225 machine 434 joints 628,629,645--6,645,649,
flexural properties 226, 226 Stability 650,651,652,659
graphite, hybrids 354 color 43 design 652
marine applications 917 PBO fiber 238 Stiffness
for oil containment 233-4 and temperature 49 adherends 636
pricing 232 thermal-oxidative 75,92 carbon-carbon composites 333
references 774-7 Stabilization changes
short beam shear properties 226, anisotropic pitches 178 microcracking 796
226 commercial 178 with time 800
specific oxygen addition 173 coefficients
gravity 225 PAN 172,177-80,178,179 constitutive equations 743-4
modulus 1049 rate of 180 determining 737
strength 1049 Stabilizers ~6/epoxy 702,703
strength retention 230 covers, boron-epoxy 1030-1 and coupling 700
tensile modulus 225,503 UV 43 cylindrical shells 756
tensile strength 225, 503 Stacking sequence matrix, laminates 693
1110 Index

plane stress 689 T50 graphite 724, 724 polyamide-imides 121

to weight ratio, PET 223 weight ratio 21 polyarylene ether 121
tooling materials 577 aramid fibers 223 polyester 121
unbalance, adherends 629 PET 223 polyimides 121
weight ratio 21 Stress polysulfones 121
Stitched fabric 436, 436 analysis prepreg 109
Stitching 801 joints 611 thermoplastics 119
Storage 357--8 one-dimensional 627 Supported films, adhesives 276
freezer 425 concentration 610,613,614,788 Surface
Strains distribution, double lap joint 654 area 607--8
at break, PP /PCI-glass 898 interlaminar 738 compressive stress 321
displacment, equations 739 patterns, laminates 322 crusting 339
energy, adhesives 653 residual 466 of curvature 479
graphite/epoxies 4 resultants, laminates 691-3 damage 322
measuring 779 rupture defect healing 295
to failure 319 failure 24 disbonded 677
Strand mat Kevlar 212-13,213 energies 316,316
continuous 503 para-aramids 212 finish
fiber structure 503 5-glass fibers 212-13, 213 pultrusion 515
Strand properties strain curve 319 tools 597
Kulon 306 thermally induced 640 flaws 840
VMN-4306 thermoelastic 466 geometry
Strands 164 transfer 149 concave surface 484
Strap joints, tapering 645 Stretching elongations 633 fiber placement 484--6
Strength Strip heaters 499 radii of curvature 484, 486
alumina silicate 309 Structural types of 484
aramid 718 applications 709 ply, veil 503
bending 791 aircraft 1027--8,1027,1028 polish, compression molds 395
and density 23 formula, POO 237 preparation 630-1,667--84
efficiency, carbon-carbon geometry abrading 669
composites 344 braiding 415-16 adhesive joints 611
envelopes 802, 802 nonwoven textiles 419 aerospace industry 669
evaluation 737 hierarchy 778 aluminum 871, 876
M60J tape 718 reaction injection molding see automotive industry 669
P100 tape 718 SRIM cutting plies 870
properties testing 792 grit blasting 870-1
E-glass 764-7 Structure history of 667-71
graphite/epoxies 4,768-71 aramid fibers 204-5 non-autoclave 871-2
laminate 764-7 PET 224,224 peelplies 683,870
predicting 318, 318, 319 polyamide-imides 121 removal 631
S2 glass 764-7 polyarylene ether 121 repairs 870-1
T50/F584 epoxy 726,726 polyester 121 testing 669
ratio polyimides 121 properties, particulate fillers
lM6/epoxy 701 polysulfones 121 250-1
laminates 696 Styrene quality, reinforced composites
retention 213 addition 34 387
aramids 230 polyester 382 resistivity
definition 241 Submarines 473 fiberglass 135
Kevlar 205, 215 Substrates 157 Skybond 95
medical applications 957 heating 293 roughness, tooling materials 577
Spectra 230 Subsurface damage, inspection 839 slick 682
Technora 205 Succinic anhydride (SA), melting speed~grinding 600
structural 795 point 58 treatment 190-1
T300 fabric 718 Superconducting waviness 603
tests, joints 626 magnetic energy storage systems, weave, yarns 407,407
three-dimensional 489 see SMEs Sutures, PET 234
tooling materials 577 SupeM:ollidermagnets 978,978 SVM
values Supennite 247, 247 availability 220
carpet plots 719 Suppliers elongation at break 208
 Index 1111

initial tensile modulus 208 adherends 653 long-tenn use temperature 206
PBIA-based fibers 210 adhesive joints 628 melting temperature 206
sources of infonnation 221 Tapering moisture content 206
specific gravity 208 adherends 627, 629, 644-51, 645, specific gravity 209
tensile strength 208 647,648,649,650,656,658 specific heat 206
Sweden, recycling 887 advantages 646, 647 tensile strength 209
Sweep method, plaster masters peel stresses 646 thenna! conductivity 206
565-6 strap joints 645, 647, 648 Tekmilon 223,228
Swelling stresses, water absorption double 646, 647, 648 availability 231
811 Tappers, mechanical 840 elongation at break 225
Symmetric, laminates 692, 749 Tateho fiber type 225
Symmetry 9,1(}-11 fracture strength 325 specific gravity 225
Synergism 43 fracture toughness 325 tensile modulus 225
Syntactic foams, adhesives 275--6 work of fracture 325 tensile strength 225
Synthetic marble 253 Young's modulus 325 Telescopes 1014-15
Te glass 134 Temperature 817-19
T50 graphite Technical training, repairs 878 consolidation 577
material properties 724, 724 Technology control, dies 511-13,512,513
strength values 724, 724 reuse 885-9 cryogenic 817
TSO/F584 epoxy winding 458-66 modulus 212
material properties 726,726 Technora 204 decomposition 76
strength properties 726,726 availability 220 elevated 817
T300 fabric breaking strength 214 extremes, GFRP 1004-21
mechanical properties 718 chemical resistance 214 glass transition 75
strengths 718 chemical stability 214 gradients 470-1
Table rolling 425-32 decomposition temperature 206 profile, dies 511, 512
equipment, suppliers 431 density 206 resistance, aramid fibers 205
mandrels 429 elongation at break 209 spinning 175
materials 429-30 ether linkages 210 stability 49
Tabs 787 fatigue resistance 213 dies 514
Tack 356-7,479 filament diameter 206 and viscosity 173-4, 174
controls 356-7 filament shape 206 Template method, plaster masters
desired 479 fonns of 216 564-5
excessive 357 initial tensile modulus 209 Templates
heavy on prepregs 357 linear thennal expansion ply locating 574-5
requirements 357 coefficient 206 trim and router 574
Tactical aircraft, material weights long-tenn use temperature 206 Tenacity, definition 241
1027 melting temperature 206 Tennis rackets 463
Talc 246 modulus and temperature 210, Tensile
composites 246 211 elongation 41, 42
impact strength 246 moisture content 206 recycled PET 894
Italian 248-9,248 properties 205 RN 440
low cost 246 sources of infonnation 221 failure 383
110ntana 248-9,248 specific gravity 209 forces, transmission 787
New York 248-9, 248 spin process 213 load, fibreglass epoxies 153
polypropylene filled 246 strength retention 205 modulus
Tank solution processing 871 tensile strength 209 aramid fiber 715, 716
Tap test 840 and temperature 210, 211 AS-4 carbon fiber 123
laminar type flaws 840 thennal conductivity 206 B11C 386, 386
Tape 164 Teflon 250,360,834 carbon fibers 335
braiding 418 Teijinconex 205 carbon (Type BOO) 503
collimated 359 decomposition temperature 206 carpet plots 719
fiberglass 146 density 206 Celion 87
non-woven 359 elongation at break 209 continuous ceramic fibers 309
placement 456 filament diameter 206 Dyneema 225
reinforcement 15 filament shape 206 E-glass 360,383,503,715, 716
slit 478 initial tensile modulus 209 epoxy 511
unidirectional 298,359 linear thennal expansion graphite 360,383,715,716
Tapered coefficient 206 Hoechst Celanese 225
1112 Index

Kevlar 383, 503 Kevlar 208, 383,503 carbon-carbon composites 349

Kulon 306 Kulon 305, 306 compression 346
lot-to-lot variation 725 lot-to-lot variation 725 flexure 346
LPMC 386, 386 LPMC 386, 386 comer 346
marine laminates 920 marine laminates 920 impact 346
PBO 235,238 and microtexture 188 impregnation 341, 343-4, 344
PMR-1582 and modulus 21, 21 interlaminar shear 346
laminates 88 and nitrogen release 182 interlaminar tension 346
polyester 511 Nomex 209 process 341
PP and granulated SMC 897 NR-150 82 shear 346
PP IPCI-glass 898 nylon 6/6 PCI-glass 899 tension 346
recycled NBC 897 PAN-based fibers 169, 170 thermal expansion 346
recycled PP 895, 896 PBO 235,238 Terephthalic acid 35
recycled SMC 890,892, 893 phenolics 258 Terephthalics 36, 39
regrind RIM 895 pitch-based fibers 170 processing time 36
RTM 440 plied-yarn 151 reactivity 36
52-glass 360,383,503 PMR-15 82,88 solubility 36
SMC 386,386 polyester 511 Tertiary amines 104
Spectra 225, 503 polyurethane 258 Test
and strength 21,21 polyvinyl chloride 258 direction, yam distribution 154
Tekmilon 225 pp IPCI-glass 898 methods 721
thermoplastic resins 122 recycled PET 894 adhesives 630
Vectran 235 recycled SMC 890, 892, 893 joints 626
vinylester 511 regrind RIM 895 specimen
VMN-4306 52-glass 383 dimensions 786-7
ZMC 386,386 S-glass 503 loading sections 786
properties epoxies 152 transition sections 786
aliphatic system 70 single-yam 151 Testing
anhydride-cured system 72 skinned molded foam 258 barcol hardness 38
E-glass 360 5kybond 86 bars 784
graphite 360 5MC 386,386 compression 38, 781
testing 38 5pectra 225, 503 distortion temperature 38
stiffness, aramid fibers 209 steel 170 eddy current 849-50
strength SVM 208 flat specimens 786-7
ABS 258 Technora 209 flexural properties 38
AFR700B/S21aminates 93 Teijinconex 209 half-discs 782
aluminum 170 Tekmilon 225 history 779
aramid fibers 209 thermoplastic resins 122 hydraulic system 782
Armos 208 titanium 170 impact strength 38
AS-4 carbon fiber 123 Twaron 208 in-plane shear methods 783
BMC 386,386 Vectran 235 incoming materials 20
carbon fibers 169 vinylester 511 microwave 853
carbon (Type T300) 503 VMN-4 305, 306 non-symmetric bending 783
carbon-carbon composites yam distribution 154 off-axis tension 781
333 ZMC 386,386 pressure compliant ring 782
carboni aluminum MMC 300 stress rail shear 783
carpet plots 720,720 PP and granulated SMC 897 requirements 779-80
CE 108 recycled NBC 897 ring specimens 782, 784
Celion 87 recycled PP 895, 896 rods 784
cellulois acetate 258 Tension selecting techniques 779
continuous ceramic fibers 309 and compression 786-9 special tests 792
Dyneema 225 failure in 787 specimens 780,786
E-glass 24, 383,503 head, fasteners 625 split rings 784
epoxies 258, 511 tension ratio 1049 structural 792
fiberglass 134 TEOS 346 tensile properties 38
epoxies 153 testing 781, 782, 788 tension 781
graphite 383 rings 782, 788 torsion of square plate 783
and higher temperatures 182 Tensioning 469 tubes 784
Hoechst Celanese 225 TEOS 340 ultrasonic through-transmission
 Index 1113

840-2,841 polypropylene 259 stress 659,660,661-2

Tetraethylenepentamine (TEPA), polyurethane 259 laminates 322
viscosity 54 polyvinyl chloride 259 Thermally
Tetraethylorthosilicate, see TEOS skinned molded foams 259 induced stresses 640
Tetramethylbisphenol F steel 558 resistant gloves 829
homopolymer property 102 Technora 206 Thermoelastic behavior, anisotropic
structure precursor 102 Teijinconex 206 bodies 470-1
supplier 102 thermoplastics 531-2 Thermoforming 436, 437, 526,
trade name 102 Twaron 206 529-30,529,532,545-9
Tex,defirrition 136,241 urethane board stock 558 polymer properties 545
Textile control tape 819 Thermoplastics 356
fibers 307 cycling 89,89,816,817 aerospace 115
preforming aircraft 818 aging 127
defirrition 397 spacecraft 818 amorphous, HDT 126
linear 398 data 818 automotive market 115
planar 398 decomposition, carbonyls 293-4 chemical resistance 126
three-dimensional 398 degradation 31 composite density 534
preforms 401-2 diffusivity, thermoplastics 534 composites
role of 398-401 expansion 3,5, 19 extrusion 526
selecting 401 adherends 630, 661 injection molding 526
T 120 carbon-carbon composites processing 525-52
g NR-150 82 333 compression molding 116, 544-9
PMR-1582 ceramic composites 312 compressive properties 124, 125
polyarylene ether 121 coefficient diaphragm forming 116
polyester 121 continuous fiber 162 dielectric constant 107
polyimides 121 staple 162 dissipation factor 107
polysulfones 121 whiskers 162 extrusion 534-8
resins 440 composite tools 592 fatigue resistance 115
Thermal correction, tooling design 590 fiber content 117
accelerators 509 E-glass 383 folding 116
coefficient graphite 383 fracture toughness 124
BMC 386,386 molding 593 heat capacity 532-4
Kevlar 383 NR-150 82 high price of 1024
LPMC 386, 386 PMR-1582 impact energy 125
ZMC 386,386 52-glass 383 injection molding 116
coefficients, orthotropic layers SMC 386,386 joining 127
742 steel dies 507 low toxicity 115
conductivity TEOS 346 manufacturing 116-17
ABS 259 thermoplastics 127 matrices, aramid fibers 221
aliphatic system 70 tooling materials 577 melt, high viscosity 116
aluminum 558 extremes 816-17 molecular weight 116
aromatic system 71 gradients, carbon-carbon polymers
carbon fibers 169 composites 333 adducts 526
carbon-carbon composites inertia technology 444 fillers for 527
333,349 mismatch 638, 661, 662 reinforcements 526
carboni epoxy prepreg 558 oxidation 79 prepregs 357
cellulois acetate 259 carbon-carbon composites processing 115
ceramics 307 345,347-8 properties 115
electroformed nickel 558 properties pultrusion 116
epoxies 259 aramid fibers 205-7,206 reinforcing 528
glass/epoxy 558 BMI 106,109 remelting 115
high carbon cast steel 558 carbon fibers 184-5, 185, 186 resins 31-2
Invar 558 CE 106,109 fracture toughness 122
Kevlar 206 composites 661 health effects 834
mahogany 558 fiberglass 135 injection 116
monolithic graphite 558 rotational molding 551 tensile modulus 122
Nomex 206 radiation 816 tensile strength 122
particulate fillers 242 resistance, aramid fibers 205 solvent resistance 115, 127
phenolics 259 stability, composite tools 592 structure 118
1114 Index

suppliers 119 Three zone model, pultrusion 505, use of caul plates 590
thermal 505 development, space 1015-16
conductivity 531-2 Three-dimensional elastomer 14
diffusivity 534 fiberglass 146 elastomeric 593-4, 593
expansion 127 lamina properties 761 electroformed nickel 591-2,592
thennoforming 544-9 reinforcement 456, 801 employee induced damage 563
tougheners 106 Threshold limit values, see TLV epoxy laminates 562
trade names 124, 125 Through-the-facesheet evacuation, fiber placement 479-80
transverse modulus 123 aircraft 865 fiber separation 562
transverse strength 123 TIme weighted average, see TWA fixed-volume method 593, 593
unidirectional Tune window, consolidation 588 graphite-epoxy 592-3,593
compression strength 125 Tltanates, coupling agents 251 hole fittings 562
trade names 125 TItanium leakage 562
winding 458 alloy- linear thermal expansion master models 563-ii
Thermosets 705 materials 589
amorphous 118 carbide 319 bleeder 577
compressive strength 124 density 170 consolidation 577
joining 127 diboride 319 hardness 577
polymers 34 elastic modulus 170 metals 577
Thermosetting 37,356 fracture elongation 170 properties of 558
binders 458 specific modulus 1049 stiffness 577
dielectric constant 107 specific strength 1049 strength 577
dissipation factor 107 tensile strength 170 surface roughness 577
reaction, definition 49 tension-tension ratio 1049 thermal behavior 589
resins 518 TLV 825,826 thermal expansion 577
Thermostamping 22 Tm moisture absorption 561
Thick composites, consolidation 585 polyamide-imides 121 and part shape 589-94
Thick molding compound, see 1MC polyarylene ether 121 prepreg 562
Thick-walled structures 456, 46&-71 polyester 121 resin system 562
loaded in torsion 473-4 polysulfones 121 rubber 366
monolithicity 456 1MC380 sealers 561
residual stresses 469 Tolerance selecting materials 557
torsion strength 473 stackup 561 sheet wax 574
Thickeners 383 to flaws, ceramics 307 solid rubber 670
Thickness Tonox 60-40 (Uniroyal Inc) 56 spray metal 442
adherends 627-9,628,628,658 Tool urethane based 567
bond 631 geometry 439 variable volume method 593, 594
layer 628, 656 life 599 wet lay-up repairs 867
to adherend 641 factors 596-7 Tools
drawings 731 pins 574 composite 361
flberglass 141-4,153 separating from master 570 cure cycle times 559
Hercules 1M7 fiber 478 steel, erE 589 depth of cut 597
honeycomb 263 surface design 556-63
preform 438-9, 438 area 480 durability 558
sandwich 284 strength 480 fabrication, cost 558, 559-60
section 512 usage 559 life expectancy 558,559,561
variation 372-3 wear 599 machinability 558
Thin composites, consolidation 585 Tooling 14, 361-ii, 392-5,429-30, materials available 560
Thin-walled 447, 479-80, 480, 556-75 materials selection S59
beams 747-9,748 alternative 449-50,449,450 methods of fabrication 560
shear stress resultants 748, balls 565 mounted 480
748 bushings 562 multi-cavity 500
uses 747 cast aluminum 442 resistance to failure 561-3
shells 456 correction method 590-1 selection 598
structures 466-8 eTE 556-7, 557-8 surface finish 597
Thinning, organic solvents 65 design 590 usage rates 558
Thixotropic liquids, adhesives 275-6 coordinating partial plies 590 wear ratio 597
Thornel 1054 thermal expansion correction Torayca 1054
properties of 170 590 fiber types 1054
 Index 1115

properties of 170 whiskers 158,162 density 206

Torque levels 625 Transfer molding 22 elongation at break 208
bolt tensions 626 Transformation relations 707 filament diameter 206
Torsion 784, 790 Transition, areas of 731-2,732 filament shape 206
square plate, testing 783 Transportation 905-15 forms of 216
strength, thick-walled structures future directions 915 initial tensile modulus 208
473 land 905-15 linear thermal expansion
Total hip arthroplasty 960-1 pultrusion 520-1 coefficient 206
femoral components 960-1 Transverse long-term use temperature 206
FRP 959 compression, Kevlar 212 melting temperature 206
metallic devices 960 erE 699 moisture content 206
Total overlap length 658 extension, IM6/epoxy 701 properties 205
Tougheners modulus 12 sources of information 221
reactive rubber 106 IM6 / epoxy 699 specific gravity 208
thermoplastic 106 M401/F854 699 specific heat 206
Toughening 104-5 resins 1056 tensile strength 208
interphase 802, 802 thermoplastics 123 thermal conductivity 206
Toughness 801 shear Twill, weave 145,145,405,406
a~ives 274-5,274 adherends 630 Twin-screw extruders 535-6
aramid fibers 221 failure, sandwich 282 advantages 539
interlaminar 124, 124 strain, IM6/epoxy 701 classification 537
measuring 274,274 strength disadvantages 539
resins 440 moisture 811 Twist
Tows 164 thermoplastics 123 drill 600
bridging of 487 tensile strength, resins 1056 Kevlar 215-16
buckled 480 tension Two-dimensional
carbon fiber 292, 293 Kevlar 212 flow, compression molding 584
cutting 483 test 688 lamina properties 760-1
definition 476 Treatments, Vectran 236 reinforcement 456
dropping and adding 483-4 Triaxial, plain-weave 399 'JYpical
end locations 481 Triethylenetriamine (TETA), compression strength, carbon
folded 487 viscosity 54 fiber composites 1055
selecting 478 Triglycidyl p-aminophenol 53 properties 1053-8
thickness 478 Trimellitic anhydride (TMA), E-glass 1055
twisted 487 melting point 58 selecting 1057-8
wandering 487 Trimming 446 'JYranno 316
width Kevlar 600
histogram 479,479 2,4,6-Tris(dimethylaminomethyl)- UHMWPE 959, 962
of material 479 phenyl 60 compression molding 545
Toxic 822-4 Trucks 912 reinforcement 963
definition 822 Truss structure, spacecraft 968-70 Ultimates, Kevlar 210
Toxicity TRW-R-8XX 94 Ultrahigh molecular weight
acute and chronic 823-4 chemistry of 81 polyethylene, see UHMWPE
fluorine 362 Tubes, testing 784 Ultrasonic
particulate fillers 252 Tubular structures 425-32 correlation 843-4, 844
reproductive 824 pultrusion 496-7 cutting 601
solvents 361, 362 Tunnels 950 machining (USM) 605-8
Trade names Turning 597,598,601 advantages 608
continuous fiber 162 advantages 597 applications 607-8
fibers 158 withAWJ 603-4 brittle materials 605
PAN-based tow 198 disadvantages 597 ceramic matrix composites
pitch-based tow 199 Kevlar 600 605
polyamide-imides 121 machining 602 disadvantages 608
polyarylene ether 121 repeatability 603 rates 608
polyester 121 surface finish 603-4 polar backscatter 843
polyimides 121 TWA 826 pulse-echo testing 842-3,842
polysulfones 121 Twaron resonance 843
staple 162 availability 219-20 signal 840
thermoplastics 124,125 decomposition temperature 206 to noise ratio 834
1116 Index

through-transmission testing pressure 918 pultrusion 495-6

840-2,841 processing 919-20 advantages 495
automation 840 silicone rubber 577 tape wrapper 427,428
C-scan 841,841 bagging 569,593,872-3 Vmyl cyclohexene dioxide, viscosity
vibrations 607 applications, aerospace 375 63
Ultraviolet Coronagraph with autoclave 593 Vmyl esters
Spectrometer, see UVCS bridging 369-71,369 applications 41-2
Ultraviolet resistance, PET 224 eliminating 369, 370, 371 corrosion resistance 41, 42, 42
UMC 381 edge bleeder 368 density 505
Uniaxial loading, allowable 802 expendable 367-71 elongation at break 505
Unidirectional bags 371 flexural modulus 505
composites 107-8 precautions 363 flexural strength 505
compression testing 787 reusable 371-2 general purpose 40
fabrics, weave 145,145 thermoset composites 371 heat distortion 505
laminates 4, 8 vertical bleeder 368 heat resistant 40-1
molding compound, see UMC drying 818 marine applications 916
ply 741 infiltration 299-300 properties 41-2
tape, flexural strength 401 pressure resins 28,382,504
Uniformity 301, 301, 302 cures 870 structure 40
Unit costs, polymer processes 532 repairs 869 tensile modulus 505
Unsaturated polyesters, resins 382 Valves 954 tensile strength 505
Unsaturation 39 Vapor VIscoelasticity 122
Unsupported barriers 817 VISCosity 37
films, adhesives 276 processing 316-17,316 aliphatic system 70
single lap, adhesive joints 628 VARI 910 anhydride-cured system 72
Untapered, joints 624 VARTM 919 aromatic system 71
Uranium enrichment centrifuges advantages 919 coupling agents 250
953-4 Vectran 234-7, 234, 235 definition 527
Urea-formaldehyde, adhesives 255 aerospace applications 236 elongational 529-30
Urethane board stock applications 236-7 low density polyethylene 528
ceofficient of thermal expansion availability 236 polypropylene 530, 531
558 cost of 234-5 predicting 528-9
density 558 dielectric constant 236 resins 440, 578
thermal conductivity 558 elongation at break 235 RTM 440
US Air Force Materials Laboratory forms 236 shear rate-dependent 528
22-3 melt spun 235 and temperature 173-4, 174
US Federal Aviation Regulations melting temperature 235 vs. time, epoxy resins 509, 510
(FAR) 1022 properties 235-6 VIsual inspection 839-40
Use temperature, PET 225 protective garments 236 VIS 160
UV resin 235 time needed 160
absorption, aramid fibers 221 safety materials 236 VMN-4
radiation, carbon black 245 sources of information 236 bending strength 305
resistance, PET 230,233 specific gravity 235 carbon content 306
stability tensile modulus 235 coating thickness 306
PBO fiber 238 tensile strength 235 compressive strength 305
PET 233 treatments 236 density 305, 306
UVCS 1017-18,1018 Veils 436, 438 elastic modulus 305
surface ply 503 elongation at break 306
V-cap 93 surfacing, application 436 filament diameter 306
chemistry of 81 Velocity, meteoroids 815 specific modulus 305
processing conditions 84-5 Velsicol 58 strand properties 306
Vacuum Venetian blinding 482 tensile modulus 306
assisted resin injection, see VARI Vent design 442-3 tensile strength 305,306
assisted resin transfer molding, Ventilation 830 Voids 430,548,588,630--1
seeVARTM volatiles 683 capture 450--1
bag 14,14 Venting 450-1 content
autoclave tooling 589 ports 443 checks 430
molding 361, 577 Verification, design 733-4, 733 marine laminates 920
nylon 577 Vertical exposed surface 432
 Index 1117

stability map 588, 588 Weather resistant, resins 813 core plug 866
Volatiles Weathering 813 damage assessment 866
content 356 climate 813 elevated temperatures 868
elimination 78, 79 graphite/epoxy 813 equipment 867
ventilation 683 graphite/polysulfone 813 face sheet 866-7
Volkersen Kevlar/epoxy 813 heat lamps 865, 866
model 662-3 Weave heating blankets 865
solution 635 aramid fibers 222 hot air blowers 865
Volume basket 140,145 materials 867
expansion, water absorption 811 crowfoot satin 145, 145 precured patching 867-8
fraction eight-harness satin 145, 145 procedures 866-7
fiber reinforcement 435 fiberglass 141-4 repair environment 867
IM6/epoxy 699 non-woven 145,146 restoration of coatings 867
M401/F854 699 patterns 140, 145, 150 tooling 867
resistivity and mechanical properties two resins 571
fiberglass 135 152 Wet mat molding 43
Skybond 95 plain 140,145 Wetout 443
Volumetric shrinkage biaxial 407 Wettability, fiber reinforcement 435
aliphatic system 70 plied-yam 151 Wetting agents, additives 516
aromatic system 71 shearing 439 Wetting speed 495
Von Mises criterion 708 single-yam 151 Whipple-type shield 814-15,814
Vortex, tennis rackets 1046 twill 145, 145 Whiskers 160-1,308,308,308,314
Voyager aircraft 32, 32 unidirectional fabrics 145, 145 alumina 161
Weavers 407, 407 coating 166
Wall effects 153 Weaves commercially available 308
Warp 541 hybridized 118 composition 158
knitting 408,409 reinforcing 118 continuous 156, 160
plied-yam 151 Weaving 18,402,402,404-8,405, CTE 308
single-yam 151 406,407,801 density 162, 308
yam 140 biaxial 406 description 162
fiberglass 141-4 plain weave 405 diameter 162,308
Warpage 374,467,557 satin weave 405 discontinuous 156,160-1
Warping twill weave 405 elastic modulus 162
GFRP 1013, 1013 definition 404 length 308
polyoxymethylene 541 design methodology 407-8 manufacturers 158, 162
Wash resistance, fiber reinforcement multiwarp 406 manufacturing processes 158
435 loom 405 for reinforcing 166
Wastage three-dimensional 405, 405 SiC 308
cut-off 502 triaxial 405,405,406 elastic modulus 327
pultrusion 489 two-harness loom 405 fracture strength 325, 326,
Water Web, yarns 407, 407 326,327
absorption 110,811 Weft, knitting 408,409 fracture toughness 325, 326,
aliphatic system 70 Weibull distribution 189 326,327
aramid fibers 215 Weight silicon nitride 161
aromatic system 71 aramid fibers 223 specific strength 162
rate of relaxation 811 fiberglass 141-4 thermal expansion coefficient 162
recycled NBC 897 fraction 1027-8 trade names 158, 162
recycled PET 894 loss vapor-liquid-solid process 160-1,
recycled SMC 890 AFR700B/S21aminates 94 326
swelling stresses 811 PBO fiber 238 Young's modulus 308
volume expansion 811 Skybond 86 Whole rings, bending 785
evacuation, aircraft 863 savings 1028, 1028 Width of material 479
immersion 110 Weld Wind
ingestion, aircraft 859 lines 542, 544 eye 461
intrusion, NDE 854 strength 544 tunnel blades 975-6,976
jet 601 Wet lay-up 355,571-3 turbine blades 954
resistance 3 composite tools 566 Winders
soluble sand mandrels 465--6 repairs lathe type 461
Weak bonds 671-5,671 aircraft 865-7 race-track 462
1118 Index

stationary whirling arm type 461 Silar 325 slippage 399

Winding Tateho 325 surface weave 407, 407
angle 496-7 Work-around technique 1006,1011 textured 138
with braiding 473 Workmanship 715 twisted, geometry 404
chord 462 Woven warp 14{)
direction 496-7 fabric 359 web 4{)7, 407
dry 458 laminates 818 Z twist 138
extra pressure 470, 470 flexural strength 4{)1 Yield
geometry 459 reinforcements 91~17 fiberglass 139-40
helical angle 460 Wrapping tapes 429 stress at break, PP /PCI-glass 898
history of 458 Wrinkles 487 Young's modulus
inaccuracies 463 alumina composites 325
isotropic metallic tape 468 alumina silicate 309
X-ray
layer-by-layer curing 470 carbon fIbers 295
backscattering 845-6,846,847
materials for 458-9 CE 108
imaging 844-9
non-geodesic 457 ceramic composites 312
cost advantages 845
pattern 459 ceramics 307
foreign materials 844
closure 459
honeycomb defects 845
E-glass 24
control 463-4 and microtexture 189
matrix cracks 844
pitch of 496-7 resins 440
moisture contamination 858
planar-polar 462 Silar 325
non-film 845
preform 458 Tateho 325
porosity 844
process 456, 457 whiskers 308
XMC 381
programmed 469-70
pull 489, 490 Z molding compound, see ZMC
and pultrusion 496-7 Yarns Z twist, yarn 138
semifabricated composite 468 classification 4{)2 Zero-shear viscosity, rotational
speeds 463 content, distribution 154 molding 551
stages 469 crimp 416 Zirconates, coupling agents 251
technology 458-66 cross sections 411 ZMC380
theory of 468-9 designation, fiberglass 139-40 flexural modulus 386, 386
thermoplastics 458 distribution 154 flexural strength 386, 386
of a torus 462 filament 164 formulation 384
wet 458 fill 4{)7, 407 injection machine 380
Wmdmills 521 filling 140 lZOD 386, 386
Wood 256, 838, 958 geometry 4{)3-4, 403 low viscosity 380
durability 256 identification 138-9 specific gravity expansion 386,
end-grain balsa 256 inclination angle 4{)7 386
low cost of 256 jamming 412 tensile modulus 386, 386
rehabilitation 988 longitudinal 415 tensile strength 386, 386
Work of fracture orientation angle 412 thermal coefficient 386, 386
alumina composites 325 S twist 138 Zonyl 251