POLYURETHANES EXPO 2001

September 3 0 - October 3, 2001

G reater Columbus Convention Center
Columbus, O hio

CONFERENCE PROCEEDINGS

Sponsored b y the Alliance for the Polyurethanes Industry

a business unit of the American Plastics Council

A lliance for the

Am erican
P olyurethanes
Plastics
I ndustry
Council®

Boca Raton London New York

CRC Press is an imprint of the

Taylor & Francis Group, an informa business
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2001 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an lnforma business

No claim to original U.S. Government works

ISBN 13: 978-1-58716-134-6 (pbk)

ISBN 13: 978-1-138-45961-8 (hbk)

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have
been made to publish reliable data and information, but the author and publisher cannot assume responsibility
for the validity of all materials or the consequences of their use. The authors and publishers have attempted to
trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if
permission to publish in this form has not been obtained. If any copyright material has not been acknowledged
please write and let us know so we may rectify in any future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted,
or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including
photocopying, microfilming, and recording, or in any information storage or retrieval system, without written
permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www. copyright.com
(http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers,
MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety
of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment
has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only
for identification and explanation without intent to infringe.

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com

and the CRC Press Web site at

http://www.crcpress.com

Disclaimer

The papers contained in these conference proceedings are reprinted with the permission of the individual
authors who are solely responsible for their content. The papers have not been subject to peer review.
API and APC assume no responsibility for the accuracy or use of the information presented in these
papers, and hereby disclaim all liability in connection therewith. API and APC make no warranties or
guarantees, expressed or implied, related to the material presented in these conference proceedings.
Table of Contents

TECHNICAL SESSION A— Environmental

Coatings Prepared from Polyurethane Soft Foam Recycling Polyols.................................................. 3
R. Langenstrassen, H. Huth, M. Pohl, K.-H. Schmidt, G. Behrendt, Institut fur
Kreislaufwirtschaft der Technischen Fachhochschule Wildau
H. Goering, Bundesanstalt fur Materialforschung und -prufung
S. I. Ivanyi, University of Chemical Technology and Metallurgy
Improved Comfort Factor with Recycled Content by Reformulation of Flexible
Slabstock Foams.......................................................................................................................................11
R. Villwock, Mobius Technologies, Inc.
Recent Developments in the Recycling of Flexible PU Foams Back into New PU
Foam P ro d u cts...................................................................................................
R Berthevas, F. Aguirre, Dow Europe S.A.
J. Tu, The Dow Chemical Company

TECHNICAL SESSION B— Automotive Low V.O.C. Emissions

Non-VOC Catalysts for HR Molded Foams Used in Automotive Seating A pplications............... 33
M. Sakai, A. Ishikawa, M. Morii, K. Kiuchi, Kao Corporation
New Silicone Surfactants for TSI Based HR Molded Foams—Optimization of Bulk
Stabilization and Processing Latitude.................................................................................................. 41
T. Boinowitz, G. Burkhart, J. Klietsch, Goldschmidt AG
W. Bunting, Goldschmidt Chemical Corp.
Elimination of Amine Emissions from Polyurethane Foams: Challenges
and O pportunities................................................................................................................................ 47
F. M. Casati, J. M. Sonney, H. Mispreuve, A. Fanget, Dow Europe S.A.
R. Herrington, J. Tu, The Dow Chemical Company

TECHNICAL SESSION C—Testing

Monitoring of Polyurethane Foam C u r e ............................................................................................ 61
S. A. Jones, K. W. Scott, B. G. Willoughby, Rapra Technology Ltd.
E. A. Sheard, Lescon, Inc.
Three-Way Thermal Conductivity Instrument C om parison........................................................... 71
B. Canney, C. Dixon, N. Mathis, Mathis Instruments Ltd.
Testing for MDI Emissions from Commercial Boards and Polyurethane G lu e ........................... 75
P. P. Parekh, B. Karoly, Huntsman Polyurethanes
Report on the First ASTM Round Robin of Vacuum P a n e ls......................................................... 81
A. Brzezinski, LaserComp
T. Stovall, Oak Ridge National Laboratory

Table of Contents / iii

TECHNICAL SESSION D— Construction
The Development of All Water-Blown Polyisocyanurate Foam System for
Metal-Faced Continuous Sandwich P an els......................................................................................... 89
A. Naruse, H. Nanno, M. Kurita, H. Inohara, T. Fukami, Nippon Polyurethane
Industry Co., Ltd.
Alternative Blowing Agent Solutions for Polyisocyanurate Continuous Laminate
Metal Panels ......................................................................................................................................... 97
M. J. Skowronski, M. R. Huspek, The Dow Chemical Company
C. Righi, Dow Italia S.p.A.
Processing Gaseous Blown Spray Applied Polyurethane F o a m ......................................................107
I. Wheeler, R. Crooker, D. Wickwire, ATOFINA Chemicals, Inc.
Meeting the Needs of the Future with HFC-245fa and HFC-245fa Blend Technology
in the Construction Industry...............................................................................................................115
M. Bogdan, D. Williams, Honeywell
Pour-in-Place Discontinuous Panels Using P entanes........................................................................125
A. DeLeon, D. Shieh, Oxid L. P.
E. F. Feske, Albemarle Corporation

TECHNICAL SESSION E— Innovative Polyurethane Solutions for

Automotive Applications
Innovative Polyurethane Seat Trim Cover S o lu tio n ..................................................................... 133
E. Kurtycz, S. English, A. Larre, The Dow Chemical Company
J. Bladon, Dow Chemical Canada
A. Hogg, J. Godoy, Dow Europe S.A.
Automotive, Moulded Visco-Elastic F o a m s.......................................................................................143
R Farkas, R. Stanciu, L. Mendoza, Woodbridge Foam Corporation
HFC-245fa: An Ideal Blowing Agent for Integral Skin F o a m ...................................................... 151
B. Zhong, J. M. Bowman, D. Williams, Honeywell
Advances in Polyurethane RRIM for Automotive Exteriors......................................................... 161
M. F. Hurley, Bayer Corporation
Next Generation E-Coat Capable Exterior Body P an els............................................................... 171
C. Bam brick, Dow Chemical Canada
K. Arnold, The Dow Chemical Company
High Performance Structural RIM for Large Composites: Beams, Boxes,
and Beyond ....................................................................................................................................... 177
G. Symosko, R. Cageao, D. D ’Errico, B. Lee, Bayer Corporation
W. Guarnieri, Bayer AG Leverkusen
R DePalma, Meridian Automotive Systems
New PU Cavity Filling Foam for Sound Abatement in Car Body S h e lls .................................... 183
S. Meyer-Ahrens, C. M. Milliren, Bayer Corporation
G. S. Karas, M. Lupini, Orbseal LLC, Inc.

TECHNICAL SESSION F— Fundamentals

Structure-Property Relationships of Poly(urethane-urea)s with Soft Segments
Consisting of Ultra-Low Monol Content Poly(propylene glycol) with and
without Tri(propylene glycol) ......................................................................................................... 191
M. J. O ’Sickey, G. L. Wilkes, Virginia Polytechnic Institute and State University
B. D. Lawrey, Bayer Corporation

iv / Table o f Contents
Exploring Urea Phase Connectivity in Flexible Polyurethane Foams Using Lithium
Chloride as a P ro b e..............................................................................................................................195
A. Aneja, G. L. Wilkes, Virginia Polytechnic Institute and State University
Fundamental Investigation of the Factors That Influence the Properties of Slow
Recovery Polyurethane F oam s........................................................................................................... 201
S. Narayan, A. Berube, Rogers Corporation
Polyurethane Elastomers with Multiple Chain Extenders...............................................................211
J. McClusky, S. Ghosh, X. Wei, University of Texas at San Antonio
M. A. Pocol-Savage, Kaneka Texas Corporation
The Effects of Temperature/Pressure Gradients in the Rigid Polyurethane
Foaming Process................................................................................................................................... 219
B. E. Obi, I. Latham, L. J. Brown, The Dow Chemical Company
New DMC Catalysts for Manufacturing P o ly o ls..............................................................................227
J. M. O ’Connor, SynUthane International Inc.
M. H. McAdon, The Dow Chemical Company
D. E. Laycock, Dow Chemical Canada Inc.
Nano- and Micro-Fillers for Polyurethane Foams: Effect on Density and
Mechanical P ro p erties..................................................................................................................... 239
B. Krishnamurthi, S. Bharadwaj- Somaskandan, F. Shutov, Tennessee
Technological University

TECHNICAL SESSION G—Automotive Seating Comfort and Durability

Benchmarking of Polyurethane Technologies for Automotive Seat Cushions ........................... 247
M. A. Koshute, M. Blaszkiewicz, B. L. Neal, Bayer Corporation
New Comfort MDI Molded Foam for Automotive Seating—A Study of Foam
Properties for Riding Comfort P erform ance................................................................................. 255
K. Saiki, S. Murakami, M. Hayashi, S. Egawa, T. Fukami, Nippon Polyurethane
Industry Co., Ltd.
A Study of the Influence of Polymer and Cell Structure on Polyurethane Foam
Properties for Better Riding C om fort............................................................................................... 267
M. Isobe, K. Usaka, H. Utsumi, K. Ohkubo, Mitsui Takeda Chemicals, Inc.
Alternative Methods for Durability Specification..............................................................................275
A. LeFever, J. McEvoy, Johnson Controls, Inc.
Accelerated Aging and Durability Testing of Polyurethane F o a m s............................................. 281
J. T. McEvoy, R. Yamasaki, Johnson Controls, Inc.
Measurement of High Resiliency Moulded Foam Properties as a Function of
Climatic Conditions..............................................................................................................................285
G. R. Blair, A. Russ, D. E. Bradt, R. J. Bailey, Woodbridge Foam Corporation

TECHNICAL SESSION H— Blowing Agents

Aging of Polyurethane Foam Insulation in Simulated Refrigerator Panels—Three-Year
Results with Third-Generation Blowing A g e n ts ..............................................................................313
K. E. Wilkes, D. W. Yarbrough, W. A. Gabbard, G. E. Nelson, Oak Ridge National Laboratory
J. R. Booth, Tennessee Technological University
Analysis of the Evolution of PIR Foams in the Context of the Phaseout of H C F C s.................. 325
R Dournel, SOLVAYResearch and Technology
L. Zipfel, SOLVAYFLUOR und DERIVATE
HFC 365mfc-Blown Spray Foams: Market Positioning and Environmental B en efits............... 333
H. Krahling, SOLVAYManagement Support GmbH
L. Zipfel, SOLVAYFLUOR und DERIVATE GmbH

Table o f Contents / V
HFC-134a as a Coblowing Agent with P en ta n es............................................................................339
J. Wu, D. Dillon, ATOFINA Chemicals, Inc.

TECHNICAL SESSION J—Innovation and New Application in CASE

New UV-Curable Coatings for Caul Paper M anufacturing......................................................... 347
D. Bontinck, G. Vanmeulder, UCB Chemicals
M. Idacavage, UCB Chemicals, Inc.
A Novel Plasticizer-Free PU Gel Technology for Improved C om fort.......................................... 353
L. Pellacani, V. M. T. Thiede, Dow Deutschland GmbH and Co., OHG
New Non-Isocyanate Curatives for 2K Solvent-Borne Urethane C oatin gs................................... 363
K. B. Chandalia, S. L. Goldstein, M. J. Morgan, R. T. Wojcik, Lyondell Chemical Company
The Evaluation of Metal and Tertiary Amine Catalyst in CASE A p p lication ........................... 373
H. Kometani, Y. Tamano, TOSOH Corporation
R. Van Maris, TOSOH EUROPE BM
K. M. Gay, TOSOH USA, Inc.
RimSpray™: A Unique Method for Applying Polyurethane Foam to Molds
and Substrates...................................................................................................................................... 381
M. F. Bartenstein, Gusmer-Admiral, Inc.
Investigation of Polyurethane Adhesives for Use in the Apache Aviator NBC
Protective M a sk ................................................................................................................................... 389
G. RYoung, S. Ciborowski, D. May, L. Oswald, S. Gross, U.S. Army Soldier
Chemical-Biological Command

TECHNICAL SESSION K— Furnishings

Less Temperature-Sensitive LR Slabstock F o a m ............................................................................. 403
M. Kageoka, H. Ootsuka, Y. Sato, T. Kumaki, Mitsui Takeda Chemicals, Inc.
Flame-Retardant and Physical Characteristics of Flexible Hybrid Polyurethane Foam
Materials Prepared Utilizing Foam-One Technology....................................................................... 409
C. Jayakody, D. Myers, M. Crocker, K. Bures, J. Bridge, M. Brown, Chestnut Ridge Foam, Inc.
A New Generation of Silicone Surfactants Optimized for Liquid Carbon Dioxide Blown
Slabstock Foam s................................................................................................................................... 417
O. Eyrisch, G. Burkhart, Goldschmidt AG
R. Borgogelli, Goldschmidt Chemical Corporation
DBTDL Replacement in High Resilience Slabstock F o a m s........................................................... 423
E. -J. Gerard, H. Verstraete, W. Maas, B. Schlenter, Shell Chemicals
Catalysts and Silicone Surfactants for Reduced VOC Emissions of Polyester
Slabstock F o a m ...................................................................................................................................431
M. S. Huhtasaari, R. Plaumann, J. Grimminger, Air Products GmbH
J. G. Kniss, F. D. Womack, Air Products and Chemicals, Inc.
Novel Polymer Polyol with High Polymer Content and Low V isco sity ....................................... 443
T. Tomosada, H. Ohta, S. Kono, T. Furuta, H. Horiguchi, Sanyo Chemical Industries, Ltd.
Flexible Foam Formulating Technology Development via a New Modeling A p p ro a ch ............ 451
R. J. Lockwood, Huntsman Polyurethanes
TECHNICAL SESSION L—Appliance
An Assessment of the Performance of Rigid Polyurethane Foam Insulation for
Use in Residential Water Heaters Produced with SNAP-Approved Non-ODP
Blowing Agents....................................................................................................................................461
K. A. Ingold, R. A. Yourd, V. M. Mautino, Bayer Corporation
HFC-134a: The Other H F C ............................................................................................................473
J. King, I. Latham, C. Martin, The Dow Chemical Company

vi / Table o f Contents
HFC-245fa: Foam Equipment C onsiderations...............................................................................483
J. M. Bowman, D. J. Williams, Honeywell
Development of Amine Catalysts for the Next Generation Blowing A g e n ts .............................. 491
H. Kiso, K. Tokumoto, Y. Tamano, H. Yoshimura, TOSOH Corporation
K. M. Gay, TOSOH USA, Inc.
R. Van Maris, TOSOH EUROPE B.V.
Do the Existing Silicone Surfactants Match the Requirements of HFC-245fa-Blown
Appliance S y te m s? ........................................................................................................................... 501
G. Burkhart, M. Klincke, Goldschmidt AG
K. Willoughby, Goldschmidt Chemical Corp.
“LinFlex™”, a Proven Success, New-Innovative Refrigerator Cabinet
Polyurethane M achinery.................................................................................................................... 513
D. R. Hanne, J. Willison, Hennecke Machinery
K. Pieper, B. Geiger, Hennecke GmbH

POSTER SESSION—CASE
A New Anti-Hydrolysis Ether-Based PU System for Uni-sole A pplication................................. 519
C. C. T. Huang, S. C. R Liu, Pou Chien Chemical Co., Ltd.
Investigation of Polyurethane Adhesives for Use in the Apache Aviator NBC
Protective M a sk ................................................................................................................................... 521
G. R Young, S. Ciborowski, D. May, L. Oswald, S. Gross, U.S. Army Soldier
Chemical-Biological Command
Low Gloss, Low Build Up Water Based Release Agents for Elastomers and
Microcellular Systems ....................................................................................................................... 533
M. Johnson, Chem-Trend, Inc.
Lower Durometer TDI Elastomers “Plasticizer Free” for Solvent R esistan ce............................. 535
R. B. Durairaj, J. DeAlmeida, INDSPEC Chemical Corporation
New Developments in Light and Thermal Stabilisation Packages for Flexible
Polyurethane Integral Skin A pplications......................................................................................... 537
R Jacobs, B. Johnson, M. de la Cruz, H. Goyvaerts, Great Lakes Technology Belgium
R Orizondo, Great Lakes Chemical Corp.
RimSpray™: A Unique Method for Applying Polyurethane Foam to Molds
and Substrates...................................................................................................................................... 539
M. F. Bartenstein, Gusmer-Admiral, Inc.

POSTER SESSION— FLEXIBLE FOAM/AUTOMOTIVE

Amine Emission Free Catalyst for Polyester Slabstock Flexible F o a m ....................................... 547
R. Van Maris, TOSOH Europe B. V.
H. Nakamura, Y Tamano, TOSOH Corporation
K. M. Gay, TOSOH USA Inc.
DMAMP-80™ A Superior Reactive Amine Catalyst for Polyurethane F oam s............................. 551
K. Little, R. J. Swedo, ANGUS Chemical Company
F. M. Casati, A. Fanget, H. Mispreuve, Dow Europe S.A.
Effect of Nano and Micro-Silica Fillers on Polyurethane Foam Properties................................. 557
I. Javni, W. Zhang, V. Karajkov, Z. S. Petrovic, Pittsburg State University
V. Divjakovic, University of Novi Sad
Effect of Particle Size Distribution on Performance of Copolymer P olyols................................. 563
R. Neff, W. Apichatachutapan, BASF Corporation
Flexible Polyurethanes Using a Castor Oil Based Polyester P o ly o l............................................. 565
R Penczek, R. Ostrysz, D. Wieczorek, Z. Boncza -Tomaszewski, Industrial Chemistry
Research Institute
S. V. Udeshi, Jayant Oil Mills
Table o f Contents / vii
Innovative Polymers for Optimized Release Performance for HR Molded and
Flexible F o a m .................................................................................................................................... 569
R. Rossio, Chem-Trend, Inc.
New Styrene-AcryloNitrile Polymer Polyols for Cold Cure Moulding
Automotive A pplications.................................................................................................................. ....
J. Cenens, E. Van Eetvelde, T. Schroder, B. Flammang, J. C. Vandichel, Shell Chemicals
Spray Polyurethane Elastomer for Use in Automotive Dash Insulator Applications.................. 573
A. J. Messina, G. Gebreselassie, Collins andAikman Corporation
Synthesis of PU High Resilience Flexible Foams with Dispersing Liquid Prepared
by Polyether Grafting Starch Copolym er....................................................................................... 575
C. Su, X. Xue-Wei, Nanjing University of Chemical Technology
Whiplash Countermeasures: SafeG uard™ .................................................................................... ....
E. Kuczynski, R Farkas, R. Stanciu, Woodbridge Foam Corporation

POSTER SESSION— FUNDAMENTALS/EQUIPMENT

Compatibility Study of Poly bd® Resin with Other Polyols and Property Evaluation
of Polyurethanes Perpared Therefrom............................................................................................. ....
H. S. Chao, Sartomer Co., Inc.
Dielectric Polarization Measurement with CMD Provides New Insight into the
Foaming Process........................................................................................................................ 589
B. H. W. Hofmann, Format Messtechnik GmbH
D. J. Gross, Electrotech Industrial, Inc.
IC Fluid Power for Fluid Measurement and Control..................................................................... 593
T. Marsaglia, IC Fluid Power
Mixing Equipment for the In-Line Addition of Scrap Regrind and Other
Fillers into P U R ................................................................................................................................. ....
D. Ramazzotti, Edge Sweets Company
Structures and Properties of Polyurethanes Based on Model T riglycerid es.............................. 597
Z. S. Petrovic, A. Zlatanic, Pittsburg State University
Three-Way Thermal Conductivity Instrument Comparison......................................................... 601
B. Canney, C. Dixon, N. Mathis, Mathis Instruments Ltd.

POSTER SESSION— RIGID FOAM

Amidic Polyols, New Polyhydroxylic Raw Materials for Rigid Polyurethane F o a m s ............... 607
M. lonescu, S. Mihai, E. Stepan, Institute of Chemical Research
C. Roibu, V. Preoteasa, S. C. Oltchim, S.A.
Diffusion of Gases in Rigid Polyurethane Foam at Different Tem peratures.............................. 609
M. Froling, S. Mangs, O. Ramnas, M. E. Oisson, U. Jarfelt, Chalmers University
of Technology
Dimethoxymethane as Co-Blowing Agent with HFC134a for Polyurethane
Rigid F oam s.................................................................................... ...................................................61 1
M. Modesti, A. Lorenzetti, Padova University
G. Basile, Fluorocarbons and Development Service
Glycolysis Reactivity of Rigid Polyurethane Foam for R efrigerator.......................................... 615
M. Murai, M. Sanou, T. Fujimoto, F. Baba, Mitsubishi Electric Corporation
HFC-134a as a Coblowing Agent with P en ta n es........................................................................... 617
J. Wu, D. Dillon, ATOFINA Chemicals, Inc.

viii / Table o f Contents

HFEs as New Generation Blowing Agents....................................................................................... 619
R. Tamai, S. Urata, A. Sekiya, National Institute of Advanced Industrial Science and
Technology (AIST)
H. Takeyasu, H. Sato, Asahi Glass Co., Ltd.
Hydrocarbon Blown Rigid Foam Formulations and Physical Properties.................................... 621
J. Canady, C. McAdams, K. Brown, Kosa
R Merchant, Exxon Mobil
Influence of Expandable Graphite on the Physical-Mechanical Properties and Fire
Behaviour of Flame Retarded PIR-PUR F o a m s ........................................................................... 623
M. Modesti, A. Lorenzetti, F. Simioni, Padova University
M. Gilbert, Graphtech, Inc.
Optimization of Flame-Retardants for Rigid PIR Foams: A New Screening Apparatus,
and Correlation to Large-Scale Flammability T e s ts ..................................................................... 627
B. Feske, Albemarle Corporation
J. Canaday, KoSa
Performance of Non-ODP Rigid Foams in Picnic C oolers............................................................ 629
J. Cionni, V. M. Mautino, Bayer Corporation
Sandwich Panel Lines: Comparison Between Discontinuous and Continuous Process.
Advantages Given by the Continuous Production Method Compared to the Manufacture
of Panels by Discontinuous M eans..................................................................................................... 631
E. Greco, V. Mariani, Impianti OMS Spa
Solkane ® 365/227 Blown Rigid Foams: Optimized Insulation and Cost Performance . . . . 637
L. Zipfel, K. Boemer, SOLVAY FLUOR undDERIVATE GmbH
P. Doumel, SOLVAY Research and Technology
The Effective Diffusivity of C 02and HCFC-141b in PU Foam Prepared with HCFC-141b
as Insulating Blowing Agent at 20,25 and 40 C .............................................................................. 643
J. R. Booth, Tennessee Technological University
K. Wilkes, Oak Ridge National Laboratory
Unique, Enviromentally Friendly Release Agents for MDI Foam A rticles................................... 645
B. Smith, S. Lytle, L. Reams, Franklynn Industries, Inc.

Author Index..........................................................................................................................................647
Company In d e x ....................................................................................................................................649

Table o f Contents / ix
Expo2001 Planning Committee
Joseph M. Sutej Sean Kelly
Bayer Corporation Bayer Corporation
(Chair) Alfredo B. Larre
Amy Ahlich The Dow Chemical Company
The Dow Chemical Company Elvira Monteleone
Brenda Cage Huntsman Polyurethanes
Bayer Corporation Daniel Nott
Steve Corbett Honeywell
Oxid, L.P. F. Michael Pfister
A1 DeLeon Arch Chemicals, Inc.
Oxid L.P. Wendy C. Rossman
Gary Edwards Air Products & Chemicals, Inc.
BASF Corporation Paulette Smith
Christopher Godinich Gusmer-Admiral, Inc.
Huntsman Corporation Ray Thomas
George Gula Goldschmidt Chemical Corporation
Cannon USA David Williams
Erin Marie Kelly Honeywell International
Chem -Trend, Inc.

Expo2001 Technical Program Committee

Michael J. Cartmell Peter H. Markusch
Huntsman Polyurethanes Bayer Corporation
Mark Clark Ken G. McDaniel
Cannon Bayer Corporation
Sue Clift Don McElheney
Bayer Corporation Arch Chemicals Inc
Barry Collins Mark S Prichard
Cannon Gusmer-Admiral, Inc.
A1 DeLeon Joseph M. Sutej
Oxid L.P. Bayer Corporation
Dana Gier David Williams
The Dow Chemical Company Honeywell International
David W House Steve Willis
UOP Krauss Maffei
Erin Marie Kelly Kenneth Willoughby
Chem -Trend, Inc. Goldschmidt Chemical Corporation
Alfredo B. Larre
The Dow Chemical Company

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001

 Expo2001 Issues Session Organizers
Amy Ahlich Alfredo Larre
The Dow Chemical Company The Dow Chemical Company
Charlie Bartish Robert Lockwood
Air Products & Chemicals, Inc. Huntsman Polyurethanes
Bob Blake Ken G. McDaniel
Bayer Corporation Bayer Corporation
Sue Clift Patrick H. McNamara
Bayer Corporation Huntsman Corporation
A1 DeLeon Stan Moore
OxidL.P. The Dow Chemical Company
Kurt C. Frisch, Jr. William Nicola
Bayer Corporation Bayer Corporation
Christopher Godinich F. Michael Pfister
Huntsman Corporation Arch Chemicals, Inc.
George M. Gula William P. Robert
Cannon BASF Corporation
Allan James Rich Rossio
Dow Chemical Canada Inc. Chem -Trend, Inc.
Linda Jeng Alan Schrock
Dow Chemical Canada Inc. The Dow Chemical Company
Jean Kasakevich Mark Spence
The Dow Chemical Company The Dow Chemical Company
Sean Kelly Joseph M. Sutej
Bayer Corporation Bayer Corporation
Jeff A. King Ray Thomas
The Dow Chemical Company Goldschmidt Chemical Corporation
James L. Lambach Jim Tobias
Bayer Corporation Air Products & Chemicals, Inc.
David Williams
Honeywell International

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001

 Expo2001 Issues Session Moderators/Vice-Moderators
Amy Ahlich Jayne Mallwitz
The Dow Chemical Company Arch Chemicals
Greg Banks Kirk McCabe
Huntsman Polyurethanes Bayer Corporation
Charlie Bartish Elvira Monteleone
Air Products & Chemicals, Inc. Huntsman Polyurethanes
Sue Clift Stan Moore
Bayer Corporation The Dow Chemical Company
Ken Davis Paul Neill
General Electric Appliances Stepan Company
Inna Dolgopolsky Robert Norris
Woodbridge Foam Corporation Price WateHouseCooper
Dimitri Dounis Ralph Priester
Hickory Springs Manufacturing The Dow Chemical Company
Company Steve Robbins
Michel Drouin Air Products & Chemicals, Inc.
Johns Manville William P. Robert
Tony Duley BASF Corporation
Woodbridge Foam Corporation Don Schomer
Joseph F. Jadlocki Bayer Corporation
Huntsman Polyurethanes Alan Schrock
Allan James The Dow Chemical Company
Dow Chemical Canada Inc. Richard Skorpenske
Linda Jeng Bayer Corporation
Dow Chemical Canada Inc. Steve Seneker
Tom Jensen Anderson Development Co.
General Motors Craig Sungail
Jean Kasakevich Huntsman Polyurethanes
The Dow Chemical Company Jim Tobias
Hamdy Khalil Air Products & Chemicals, Inc.
Woodbridge Foam Corporation Larry Wethje
Jim Lambach Association o f Home Appliance
Bayer Corporation Manufacturers

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001

Expo2001
Public Relations Counsel
Josiah McClellan
Gibbs & Soell

Alliancefor the Polyurethanes Industry

Virtual Trade Show Management
Mary Novack
Kenworthy Management

Alliancefor the Polyurethanes Industry

Industry Officers
API Chairman
Joseph F. Jadlocki
Huntsman Polyurethanes

Steering Committee Chairman

William A. Bernstein
BASF Corporation

Alliancefor the Polyurethanes Industry

Conference Staff
Fran W. Lichtenberg
Executive Director

Lisa B. Smith
Assistant to the Executive Director

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001

 SESSION A
ENVIRONMENTAL

TECHNICAL SESSION ORGANIZERS

William P. Robert, Chair

BASF CORPORATION

Jean Kasakevich
THE DOW CHEMICAL COMPANY

Patrick H. McNamara
HUNTSMAN CORPORATION

TECHNICAL SESSION MODERATOR

William P Robert
BASF CORPORATION

TECHNICAL SESSION VICE - MODERATOR

Jean Kasakevich
THE DOW CHEMICAL COMPANY
Coatings Prepared from Polyurethane Soft Foam
Recycling Polyols

RAINER LANGENSTRASSEN, HANNELORE HARALD GOERING

HUTH, MARTIN POHL, KARL-HEINZ Bundesanstaltfiir Materialforschung und -prufung
SCHMIDT and GERHARD BEHRENDT Unter des Eichen 87
Institutfiir Kreislaufwirtschaft der Technischen D -12205 Berlin
Fachhochschule Wildau Germany
Friedrich-Engels-Strafie 63
D -15745 Wildau
Germany

SVETLANA I. IVANYI
University o f Chemical Technology and Metallurgy
Sofia
Bulgaria

ABSTRACT produced have better tensile strength than current trade

products, an optimal elongation behavior and can be used
Recycling polyols are derived from polyurethane soft in the temperature range o f - 20 to +60 °C.
foams by the combination of glycolysis and aminolysis. In
this process the cleavage of the urethane bonds results in INTRODUCTION
urea groups and the hydroxyl compounds. The polyureas
from the water reaction of the polyisocyanate are in general At present the most of polyurethane soft foam wastes
not cleaved due to short reaction times and moderate tem ­ from production or car disassembly is burned or brought to
peratures. By this process homogenous polyols of low landfills. Possibilities o f recycling of this type of condensa­
glycol content are obtained the hydroxyl number o f which tion polymers consist only in direct recycling of the mate­
is adjusted to the properties of the coatings aimed at and is rial, e.g. direct use o f the foam flakes, combustion or
in the range of 180 to 300 mg KOH/g depending on the chemical degradation of the macromolecules, e.g. by pyro­
Shore hardness to be produced. The primary amine con­ lysis, hydrolysis, solvolysis or hydrogenation. The chemi­
tent, a limiting parameter for distribution in Europe, is cal degradation procedures have the advantage o f convert­
lower than one percent leading to no limitations in this ing the macromolecules to oligomers or monomers which
respect. The viscosity of the recycling polyols is in the can be brought to another application.
range of 3000 to 6000 mPas (25 °C) with some influence Object o f our work is the synthesis of new raw materials
o f the amount of dissolved polyureas originating from the for the production of specialty polyurethanes plastics. Our
reaction mentioned above. ecological aim is to recycle hitherto not used wastes with­
Coatings are produced by simple mixing o f the recycling out any depreciation by using a new synthesis route and
polyols plus additives in low concentrations with a di- or new technologies. The environmental impact is reduced by
polyisocyanate. The coating is either sprayed by gun onto decreasing the necessity of landfill or combustion. The
the surface like concrete, paper, sheet metal, or plastic amount of primary raw material can be reduced by using
films or by applying a knife for thin film techniques using recycling polyols so that material and energy resources are
the Mathis LabDiyer™ for knife coating and hardening. saved. Considered economically, new application fields are
The films of 0.2 to 4 mm thickness thus produced are hard­ opened by the inexpensive production o f specialty products
ened either at room temperature being tack-free after 30 to from the recycled materials.
120 minutes depending on catalysis or by applying warm Our work was directed to a modified solvolysis proce­
air of 70 °C in the LabDiyer™ for five to 30 minutes. dure which enables us to produce new polyols from poly­
Hardness and elasticity o f the coatings can be adjusted urethane soft foams which, according to the aim mentioned
by simply varying die isocyanate index. The coatings thus above, can be processed to high-grade polyurethanes, espe-

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 3

daily coatings, adhesives, and sealants as well as compos­
ites, by formulation with polyisocyanates. In this way,
polyurethane soft foams are recycled not to give the origi­
nal products but to new materials having an essentially
higher value.
This aim shall be approached by a modified solvolysis Figure 1. Cleavage o f urethane by hydroxyl compounds.
consisting in the conversion of the polyurethane soft
foams with a reaction mixture of higher glycols (molecular
weight > 100 g/mole) and aliphatic primary and/or secon­ In this reaction urethanes containing the glycol used as
dary amines within the temperature range of 120 to 180 °C. the solvolysis agent as esterification component and poly­
The recycling polyols thus produced have a set of fitting ester or polyether alcohols, resp., which were used as the
properties (hydroxyl number, viscosity, reactivity) for the polyol component in the polyurethane, are formed. As a
proposed application and are converted to the desired side reaction the urea groups can be cleaved by the glycols
polyurethanes by especially developed formulations, proc­ (figure 2).
essed to coatings as the basic example in the laboratory
with a Mathis LabDryer™ and characterized by physical
measurements and mechanical tests.

NEW PATHWAYS OF POLYURETHANE FOAM

RECYCLING

The authors studied three new pathways of the material

recycling of polyurethane foams in co-operation with part­
ners in small and medium-sized enterprises:
Figure 2. Cleavage o f urea by hydroxyl compounds.
1. Low molecular weight wastes from the polyester syn­
thesis (so-called oligoester condensates) are used in
the glycolysis instead of the glycols in order to lower In this reaction primary aromatic amines are produced
the costs of the procedure as well as to recycle an­ compulsory, being undesired because of their toxicologi­
other, not used waste stream [1]. cal, especially cancerogenic properties. In the case of using
2. A combination of glycolysis and aminolysis is applied aliphatic amines in combination with glycols, the urethane
to the polyurethane foams. With this procedure, di- groups are preferably reacted with the amine providing
and/or trisubstituted ureas are formed which are solu­ araliphatic ureas (figure 3), which are, depending on the
ble in the polyol mixture, and polyether alcohols structure of the selected amine, di- or trisubstituted.
originating of the waste of soft foams.
3. The soft foams react with dicarboxylic acids (acidoly-
sis) forming acylureas and the polyether alcohols in
their original form. As for expensive diisocyanates, the
separation o f the acylureas and the cleavage of them to
diamine is recommended. Acylureas based on com­
mon diisocyanates should be kept in the polyol mix­ Figure 3. Cleavage o f urethanes by amines.
ture. Excess carboxylic acid groups are esterified after
reaction. This reaction is for instance carried out with
phthalic anhydride and provides rigid foam polyols. In this reaction no primary aromatic amines are formed.

CHEMISTRY OF POLYURETHANE RECYCLING METHODS AND PROCEDURES

The degradation of polyurethanes is a series of reaction The recycling of polyurethane soft foams using a combi­
steps based on the special properties of the reaction prod­ nation of glycolysis and aminolysis providing a polyol
ucts contained in them. Depending on the nature of the mixture is carried out as a one-step reaction in a four-
reagent, different reaction products are formed by the con­ necked flask (0.75, 2.5 or 6 l) or in a heated stirring vessel
version of the products with the solvolysis agents. In the (high-alloyed steel, 90 or 160 1). The apparatuses are fitted
simple glycolysis with excess glycol(s) the urethane groups with a demister, cooler (which can be switched between
are simply transesterified as is shown in figure 1. refluxing and descending operation), inside thermometer
and stirrer. The reactors are flushed with nitrogen.
The raw materials are polyurethane soft foams which are
provided by car recycling enterprises or from test produc­
tions.

4 / Langenstrassen, Huth, Pohl, Schmidt, Behrendt, Goering, Ivanyi

 The general procedure was the following [2]: The including those originally present in the foams from the
weighed amount of the glycol component is filled into the reaction with water. The high-molecular weight polyureas
vessel and heated with stirring. During the heating a part of thus formed are slightly soluble in the polyols originally
the calculated amount of the amine is added. At a tempera­ forming the soft segments as well as in the polyols used for
ture o f 160 °C the addition o f the polyurethane soft foam glycolysis with different ratios. Their solubility depends on
flakes begins. As soon as the temperature reaches 180 °C, their degree of branching and their molecular weight and
the residual amount of amine is added and heating of the further on the type and chain length of the glycol. Glycols
mixture is continued. The foam flakes are added continu­ based on propylene oxide dissolve the polyureas better
ously at the defined temperature or temperature program than those based on ethylene oxide. Hence, the homogene­
until the calculated amount is added. The reaction mixture ity o f the recycling polyols is substantially determined by
is stirred for another 0.2 to 5 hours at the defined tempera­ the choice of the glycols used.
ture. Eventually, heating is stopped and the mixture stirred The reaction o f glycols with the urethane groups can be
until a temperature of 120 °C is reached. Finally the liquid considered to be a transesterification in which the short
product is filled into the prepared containers. The charac­ chain diols (used in excess) replace the long chain poly­
teristic data (viscosity, hydroxyl number, amine number, ether alcohols. This reaction, like every transesterification,
amine content by GC, water content) of the polyols are is a balance reaction. As a side reaction, urea groups may
determined. be cleaved by the hydroxyl groups. Here the hydroxyl
The polyurethanes are synthesized from the recycling groups act as nucleophilic reagents and form by this slow-
polyols according to [3]. For the manufacture of films a lier cleavage reaction primary aromatic amines. Hence, a
component A is prepared and reacted with component B, short reaction time with glycols of only low nucleophily is
which have the following composition: desired for the production of high-quality recycling poly­
ols. Our experimental series resulted in dipropylene glycol
Component A: recycling polyol 97.66 parts or a mixture from dipropylene and hexapropylene glycol to
acetylacetone 2.20 parts be optimal. Table 1 gives the formulations and some prop­
TM
Vinycene 400 0.14 parts erties of recycling polyols.
Component B: Lupranat™ M 20 A 90 to 17 parts As is shown, recycling polyols could be produced having
a wide range of hydroxyl numbers and viscosities useful
The amount o f Lupranat™ M 20 A used depends on the for polyurethane production. The family of reactive polyols
hydroxyl number of the recycling polyol as well as on the so far produced includes the range o f hydroxyl numbers
defined isocyanate index. The essential mechanical proper­ from 80 to 400. Table 2 shows a small group of selected
ties, e. g. hardness of the coating, are adjusted mainly by polyols.
the isocyanate index. The combination of different glycols allows to produce
Analysis of hydroxyl and amine numbers, resp., was car­ recycling polyols with properties advantageous for the
ried out according to standard methods. Viscosities were requirements of specialty products. The amine number of
measured by using a Hoppler viscosimeter. Amine deter­ the polyols is mainly a result of residues o f the reactants
minations were performed using a Hewlett Packard HP used and can be reduced by subsequent degassing by a
5890 Series II with MSD 5970 instrument. Mechanical rotating evaporator (120 °C/25 mPa/2 h). The reactivity of
tests were made using a tensile tester FPG 7/20 of Kogel the recycling polyols against isocyanates is veiy high so
GmbH, Leipzig, with automatic recording of force and that the starting times without any catalyst are 30 to 90
elongation. Relaxation spectra were obtained by an auto­ seconds. The reaction rate can be controlled by the addition
matic torsion pendulum of Myrenne GmbH, Aachen of 1 to 3 wt.-% o f acetylacetone.
(stepwise temperature program with 5 K per step and fre­ According to the present state of research, the recycling
quency 1 Hz). polyols may be used in several fields of polyurethane pro­
The polyisocyanate Lupranat™ M 20 A was kindly pro­ duction such as rigid foams, semi-rigid foams or elas­
vided by BASF AG and is based on aniline formic alde­ tomeric products. Their main application field is now con­
hyde condensates with an average isocyanate functionality sidered to be in elastic coatings using the polyols with
o f about 3.9.
Table 1. Formulation and Properties o f Recycling Polyols.
RESULTS ComponenfiNumber P 311 P 17 P 117 P 338
Soft foam (g) 1720 430 26500 430
Cleavage o f polyurethane soft foams is a series of reac­ Di'propylene glycol (g) 1120 140 9100 80
tive steps which proceed in the case of reacting them with Hexapropylene glycol 285
the solvolysis mixture glycol and amine as competing (g )
consecutive reactions. The faster reaction is the cleavage of Di-n-butylamine (g) 64 16 1650 10
the urethane group by the aliphatic amine resulting in the Hydroxyl number (mg 356 224 224 173
formation o f substituted ureas in which the degree of sub­ KOH/g)
stitution depends on the amine used. Because of the func­ Viscosity (mPas, 25 °C) 3370 8120 8100 9560
tionality o f the polyisocyanate branched ureas are formed Amine number (mg 48 51 39 36
KOH/q)

Langenstrassen, Huth, Pohl, Schmidt, Behrendt, Goering, Ivanyil 5

 Table 2. Properties o f Recycling Polyols.
P 318 P 338 P 326 P 223 P 225 P 222
Hydroxyl number (mg KOH/g) 81 173 210 283 328 382
Viscosity (mPas, 25 °C) 23,500 9,560 2,560 7,860 7,700 1,950
Amine number (mg KOH/g) 44 36 17 43 43 31
Water content (%) 0.06 0.07 0.07 0.03 0.05 0.06

hydroxyl numbers between 180 and 320 mg KOH/g. The the recycling polyols show some peculiarities compared to
polyols having hydroxyl numbers beyond 300 were used a typical polyurethane elastomer which are due to the com­
for the manufacture of rigid foams using the standard for­ position of the polyols and the special procedure of their
mulation. In this way, rigid foams could be produced from production. The new synthesis route of the polyols results
the polyol P 222 for instance. The following properties in a composition of excess glycols from the glycolysis
were achieved at an aimed raw density of 40 g/dm3: reaction, the polyether polyols originally used in the soft
foams (molecular weight in the range of 5,000 to 6,000
raw density (g/dm3) 36.4 g/mole and hydroxyl functionality 2.7 to 3.0), small
compression strength (kN/mm2) 2.87 amounts of hydroxyl reactive additives, e.g. chain pro­
dimension stability (24 h/140 °C) +1.6 % longed (short-chained glycols), cell openers (in general
polyethylene glycol 600 with 2 to 5 wt.-% of the polyol
From recycling polyols with lower hydroxyl numbers component) and curing agents (frequently amine polyether
and isocyanate indices below 80 semi-rigid foams are ob­ alcohols having an OH-functionality of 4 to 5). Besides the
tained. These foams were first produced in the density hydroxyl compounds, in the recycling polyols catalysts
range o f about 60 g/dm3 in order to demonstrate the poly­ from the soft foam production (e.g. diethylenetriamine and
ols to be usual for this application. other tertiary amines as well as tin-organic compounds as
Primarily, semi-rigid films are produced from the recy­ dibutyltindilaurate), which accelerate the polyurethane
cling polyols at the Mathis LabDryer™ in order to simulate formation reaction, and finally dissolved or dispersed
reactive coatings. The films are drawn with a knife coater polyureas are present.
on silicone release paper according to the standard formu­ Two examples show the characterization of the network
lation with isocyanate indices between 90 and 75 and hard­ structure by means of the dynamic-mechanical analysis:
ened at 90 °C for 25 minutes. Two example formulations polyurethanes based on polyol P 15 or P 134, resp., the for­
based on recycling polyols P 223 and P 224 (hydroxyl mulations and analytic data of which are given in table 4.
number 272 mg KOH/g) and their film properties are The relaxation spectrum o f the film based on polyol P 15
shown in table 3. shows four distinct maxima of the storage module G’
The stress-strain curves o f the two film materials de­ against temperature at -153.2, -88.2, -21.0 and 54.9 °C
scribed here (standard rods with 40 mm length and cross (figure 5).
section of 4x4 mm) show Hook behavior by an almost The transition areas at -153 °C and -88 °C may be attrib­
linear ascent of the strain against the elongation in range up uted to the onset of group mobility in the polyurethane
to an elongation of ca. 20 %. After that a distinct plateau chain. In these temperature areas the movement of the
follows until break of the specimen (figure 4). The Hook methylene groups of the chains as well as that o f the oxy-
behavior in the range of low elongation is due to the struc­
ture o f the polyurethane based on a long-chained polyether
and short-chained diols. The relatively broad plateau range
is assumed to be caused by the rearrangement of the poly-
ether chains in the soft-segment and their sliding at each
other. The hard-segment phase does not play any signifi­
cant role. A reduction of the amount of short chain diol (P
224) results in further improvement of the elastic behavior
(figure 4).
The relaxation spectra of the polyurethanes made from

Table 3. Semi-rigid Sheets from Recycling Polyols.

P 221 P 222 P 223 P 224
Polyol (g) 135 135 135 135
Lupranol™ (g) 15 15 15 15
Lupranat ™ M20 A (g) 82.5 78.8 71 75
Tensile strength
(kN/mm2) 32.3 18.2 16.9 16.5 Figure 4. Strain-elongation measurements o f 4-mm-
Elongation at break (%) 22.8 18.2 72.8 78.7 specimens made from polyols 223 or 224 and Lupranat
Shore A hardness 95 95 91 95 M 20 A.

6 / Langenstrassen, Huth, Pohl, Schmidt, Behrendt, Goering, Ivanyi

 Table 4. Polyols and Polyurethanes Made o f Them. structured in the temperature range beyond room tempera­
P 15 P 134 ture (figure 6). It can be assumed that the composition of
Soft foam (g) 850 740 the polyether polyols made from soft foams is nearly the
Diproylene glycol (g) 290 119 same in both cases. Obviously, the lower isocyanate index
Polyethylene glycol 400 (g) - 119 in the sample made of P 134 as well as the use of a single
Di-n-butylamine (g) 53 29.5 glycol in P 15 have an influence on the relaxation behavior
Hydroxyl number (mg KOH/g) 294 256 of the polyurethane. The glass transition of P 134 polyure­
Viscosity (mPas, 25 °C) 3,940 9,680
Amine number 17 46 thane is broadened compared to that of polyol P 15 and
Lupranat™ M 20 A (mixture 1:1 1.7: 1 increased by 30 K to 85 °C, other transition may be found
ratio) at about +50 °C and +130 °C. The relaxation spectrum
Isocyanate index 0.7 0.97 indicates better mixed phases and a limited phase segrega­
Additives: tion caused by the composition of the recycling polyol and
titanium dioxide (phr) 8 the isocyanate index.
castor oil (phr) 6 Figure 7 shows the relaxation spectra of a series of poly­
urethanes produced from recycling polyols with increasing
gen atoms o f the polyethers as can be seen from the amount of soft polyurethane foam by the same solvolysis
maxima of the storage modules. The transition area at -21 mixture of dipropylene glycol and di-n-butylamine (for
°C is caused by the onset of the segment mobility in the composition see table 3). The glass transition of the soft
polyether elastic phase and marks in this way the first a - segment phase at -45 °C appears explicitly indicating a
area of the phase-separated polyurethanes (for comparison: well pronounced phase separation of the weakly cross-
the transition areas of OH-functional high molecular linked polyurethane network based on the high molecular
weight polyether alcohols with an average molecular weight polyether. The glass transition region in the tem ­
weight o f ca. 5,000 reacted with 4,4’-diphenylmethane- perature range of 75°C exhibits features related to the
diisocyanate are at ca. -45 °C). In the range of approxi­ change in structure of the recycling polyol correlating to
mately 55 °C the main relaxation is found showing the their composition: the polyurethane made from recycling
glass transition of the polyurethanes from the low molecu­ polyol with the lowest amount of soft foam (P 221, 59.2
lar weight glycols (for comparison: the a relaxation of a weight-% of soft foam) shows a single glass transition at
polyurethane made of dipropylene glycol and methylene 76.5 °C with only little expressed shoulder at higher tem ­
bis(4-phenyl-isocyanate) is found at +78 °C). The relaxa­ peratures (at about 105 °C). The higher the content of soft
tion of the hard segments derived from the polyisocyanate foam in the composition the higher the amount o f high
could not be determined because of thermal degradation molecular weight polyether triol, but at the same time also
beyond 200 °C. It is established that the transition tempera­ that of polyureas resulting from the polyisocyanate-water
ture of hard segments of aromatic polyisocyanates and reaction during soft foam production. The glass transition
short chain glycols or water (polyureas) are in the range of of the polyurethane of the recycling polyol P 222 (64.5
240 °C. The loose ends (OH groups) existing in the poly­ weight-% of soft foam) is found at 82.8 °C, i. e. the in­
mer because of the low isocyanate index act as an inner crease in the glass transition temperature of 6.3 K suggests
softener and cause lower glass transitions and lower a better segregation into soft and hard phases of the poly­
strengths. urethane. The high temperature shoulder is found in the
Compared to the relaxation spectrum of the film made range of 115°C and is more pronounced as in the polyure­
from polyol P 15 that of the film from polyol P 134 is less thane of P 221. The polyurethane from recycling polyol P
223 (65.6 weight-% of soft foam) has its glass transition

Figure 5. DMA measurement o f a 4-mm-film o f the polyure- Figure 6. DMA measurement o f a 4-mm-Hlm of the polyure­
thane made from polyol P 15 and Lupranat M 20 A. thane made from polyol P 134 and Lupranat M 20 A.

Langenstrassen, Huth, Pohl, Schmidt, Behrendt, Goering, Ivanyi / 7

 rigid foams, duromer foams, sheets, adesives, coatings and
binders for composites containing natural raw materials
[4]. The main properties of the recycling polyols are their
adjustable hydroxyl numbers and viscosities. In general,
for machining viscosities lower than 10,000 mPas (25 °C)
are required which can be achieved by this procedure. The
polyols have to be homogenous and stable and their reac­
tivity has to be suitable to the special application. The
homogeneity is achieved by the composition of the sol­
volysis mixture, especially of the glycols. Low-molecular
polypropylene glycols, in particular dipropylene glycol, but
also propylene glycols of an average molecular weight of
up to 600 g/mole, have proved to be suitable for this pur­
pose.
The reactivity of the recycling polyols is essentially de­
Figure 7. Relaxation spectra o f polyurethanes produced termined by the amine content and the type of the hydroxyl
from recycling polyols according to table 3. groups. If a secondary aliphatic amine like di-n-butylamine
is used only a low residue of it may be contained in the
polyol so that the reactivity may be too high for some ap­
broadened and split into two maxima at 67.8 °C and 83.9 plications. It was found that the addition of acetylacetone
°C suggesting the appearance of a third phase in the poly­ in a certain ratio can mask the amine reactivity and control
urethane. Further, an explicit shoulder may be seen at the pot-time of polyurethane preparations. The properties
about 128 °C. These data let us assume that the dissolved of the polymers are not evidently influenced by the addi­
polyureas in the recycling polyols are reacted with the tion of acetylacetone.
polyisocyanate and thus incorporated by chemical bonds The main area of application of these recycling polyols
into the polymer matrix. They are thus chemically bound was in the field of sheet materials [5]. During the develop­
reinforcing fillers in the nanometer range and exhibit prop­ ment of these polyols it became evident that the solvolysis
erties of reinforcing nano particles. The shift of the transi­ procedure developed provides polyurethanes with elastic
tion temperatures to higher ranges implies that the nano properties without any necessity of the addition of long-
particles support the phase segregation and form within chained primary raw materials. Hence, for the production
themselves hydrogen bonds of higher density with the of semi-rigid or elastic polymers it is sufficient to react
known higher bond energy in them. The breakdown of the only recycling polyols with commercial polymer isocy­
physical structure of the rigid polyurea segments occurs anates. Hardness and elasticity of the polyurethanes can be
because of these two reasons at higher temperatures. Fur­ adjusted only by vaiying the isocyanate index. This repre­
ther, beyond about 125 °C there is a plateau region in any sents an interesting method for the production o f coatings
of the polymers up to the end of measurement at 200 °C with different properties made from the same polyurethane
resembling the classical rubber plateau but depending on system.
the composition in its absolute values. The higher this The coatings for tanks, tubes or concrete surfaces devel­
plateau region is the less the rigid domains are broken and oped on the basis o f the recycling polyols, compared to
support the structure to higher temperatures; in this respect current commercial products parallel machined, have a
the content of the polyureas supports the absolute altitude better strength and an optimal elongation behavior. The
of this plateau. With higher polyurea concentration in the strength of the coatings manufactured as a film may be
recycling polyols, the value of the mechanical loss in the twice as high as that of a comparison product having at the
plateau region increases. same time a higher elongation and a broader application
The series of recycling polyols thus produced show a de­ range with regard to temperature. The coatings can be used
fined influence of composition on the mechanical proper­ without any limitation in the temperature range o f -20 to
ties. When mechanical or physical properties are known or +60 °C owing to a completely different structure of the
the table of them corresponds to a certain field of applica­ polyol component and the broad transition area in the stor­
tion the formulation of the recycling polyol may adjusted age module resulting therefrom. After reaction with pure or
directly to the desired properties. raw 4,4’-diphenylmethanediisocyanate, the recycling
polyol structure yields polyurethanes the segregation be­
DISCUSSION havior o f which can be controlled by the solvolysis mixture
and the formulation so that it is possible to produce final
The newly developed procedure of combined aminolysis products with a wide range of properties by the exclusive
and glycolysis o f polyurethanes allows the production o f use o f recycling polyols. Thus, new property spectra and
recycling polyols with wide range properties which are new application fields have been opened to recycling poly­
useful for a great number of applications. The hitherto ols made from polyurethane soft foams.
examined application spectrum includes rigid foams, semi­

8 / Langenstrassen, Huth, Pohl, Schmidt, Behrendt, Goering, Ivanyi

REFERENCES Hannelore Hath

1. Behrendt, G., and M. Pohl. 1998. ’Tw o Waste Streams After working as laboratory as­
- One New Technology - a New Family of Valuable sistant at BASF, Hannelore Huth
Products/4paper presentes at Purdue Industrial Waste studied chemistry and was lead­
Conference, West Lafayette, May 1998. er of the quality control labora­
2. Behrendt, G., and M. Pohl. 1999. “Verfahren zur Her- tory at a corundum producer.
stellung von Polyolen und diese Polyole,” German Of- After an education in environ­
fenlegungsschrift 199 17 932, April 16,1998 / October mental analytic she joined the
21,1999. Institut fur Kreislaufwirtschaft
3. Behrendt, G., and M. Pohl. 1999. “Verfahren zur Her- and is now the leader of the
stellung von Polyurethanen,” German Offenlegungs- laboratory o f polymer analytic.
schrift 199 17 934, April 16,1998 / December 9,1999.
4. Behrendt, G., Pohl, M., Lehrack, U. and J. Volk. 1999.
Svetlana I. Ivanyi
“Verbundstoffe und Verfahren zu ihrer Herstellung,“
German Offenlegungsschrift 198 17 541, April 16,
Cand. Ing. Svetlana Ivanyi studies chemical engineering at
1998 / October 28,1999; Pohl, M., Behrendt, G., Leh­
the University of Chemical Technology and Metallurgy at
rack, U. and J. Volk. 1999. “Polyurethan-Ceralith-
Sofia (Bulgaria). She worked at the Technische Fach­
Verbundwerkstoffe,“ presented at the 3rd Beckmann-
hochschule Wildau for six months for her practical work to
Symposium, Wismar, June 5,1999.
receive her diploma degree.
5. Behrendt, G., and H. Koch. 2001. “Entwicklung neu-
artiger Betonversiegelungsmaterialien,” R&D-project
M artin Pohl
of TAB GmbH and Technische Fachhochschule Wil-
dau, 2001.
After finishing his studies at Humboldt-Universitat Berlin,
Martin Pohl worked as an assistant in analytics. In 1996 he
BIOGRAPHIES joined the Institut fur Kreislaufwirtschaft and was the pro­
ject leader in polyurethane recycling and formulation until
Rainer LangenstraBen
2000.
After finishing his studies at Karl - Heinz Schmidt
Humboldt-Universitat Berlin,
Rainer LangenstraBen worked as After finishing his studies in
a research scientist at the Aca­ chemical technology, Karl-Heinz
demy of Sciences in organic, Schmidt worked in the appli-
physical and radiochemistry. In cational engineering department
1998 he joined the Institut fur of Synthesewerk Schwarzheide.
Kreislaufwirtschaft and is the After 1990 he worked as a mer­
project leader for development cantile employee of plastics pro­
of polymer recycling procedures. ducers. In 1998 he joined the
Institut fur Kreislaufwirtschaft
H arald Goering and is the project leader for poly­
urethane application techniques.
After his study of physics at
Berlin he worked at the Institute
of High molecular Compounds G erhard Behrendt
of the Academy of Sciences in
After completing his studies at
polymer physics and measure­
Humboldt-University Berlin,
ment devices development.
Gerhard Behrendt worked as a
Recently he is engaged at the
research scientist at the Aca­
Federal Agency of Material
Research and Testing, Berlin, demy of Sciences in poly­
urethane research. After 1990 he
and working on nanostructures.
worked as manager of a subsidi­
ary of an Environmental Engi­
neering Co. Since 1995 he has
been engaged as a professor at
Technische Fachhochschule Wildau and is the manager of
the Institut fur Kreislaufwirtschaft since 1995.

Langenstrassen, Huth, Pohl, Schmidt, Behrendt, Goering, Ivanyi / 9

 Improved Comfort Factor with Recycled Content by
Reformulation of Flexible Slabstock Foams

ROBERT VILLWOCK
Mobius Technologies, Inc.
900-A Golden Gate Terrace
Grass Valley; California 95945

ABSTRACT existing predictive method (which is often factory-specific)

while accounting for the effects of powder.
Finely ground scrap polyurethane foam (Mobius Pow ­
dered Urethane, or M -PU) is used commercially in the FOAM TRIALS
production of new slabstock. The ability to reformulate to
maintain density, hardness, and other properties is key to The experimental results presented here and used to
mechanical recycling. Here, we present results of recent formulate and illustrate the mathematical property models
flexible slabstock trials with powdered urethane, and relate were obtained from two sources. The first source was
them using simple mathematical models for the density, foam produced on Dow Europe’s pilot-scale slabstock ma ­
hardness, SAG factor, and maximum block temperature. chines at their International Development Center in Mey-
The models can be used to add powdered urethane to new rin, Switzerland. Details of these trials are presented in a
formulations, or to define practical limits for the use of separate paper [3], though some of the data are duplicated
recycled powder in foams, or to estimate the cost savings in Table 1. These data were used to formulate the models.
achieved by using powdered urethane. The models were tested against data obtained from a sec­
Comfort factor (or SAG factor) is generally improved by ond source— full­scale commercial slabstock trials. From
the addition of powdered urethane. When foams are for­ this source, what is presented here is limited to a demon ­
mulated to maintain density and 40% ILD, the increase in stration of the fit between model predictions and experi­
SAG factor is due to an increase in 65% ILD and a slight mental results. The data from the latter source were not
decrease in 25% ILD. used to formulate the model, but rather served as an inde­
pendent test of its usefulness.
INTRODUCTION
M ATHEM ATICAL M ODELS
This work is focused on the prediction of four key quan ­
tities for conventional slabstock foams made with Mobius Density
Powdered Urethane (M-PU): density, hardness, SAG fac ­
tor, and maximum block temperature. These quantities are Recycled powdered urethane occupies volume inside the
routinely predicted for foams without powdered urethane junctions of the foam that is normally occupied by new
in slabstock plants around the world in the course of daily polymer, but the powder does not generate any gas to blow
production. There are nearly as many predictive models as the foam. Also, the powder doesn’t contribute to the heat
there are foam plants. These range from simple rules of of reaction; rather, it absorbs heat. Both of these effects
thumb, to the valuable experience of skilled chemists and can be compensated by increasing the water and isocyanate
operators, to sophisticated spreadsheets and computer pro ­ in the formulation to produce more carbon dioxide and
grams. Examples of the latter include the FoamKit™ soft­ more heat. For foams with densities below about 21 kg/m3
ware [1], and the models presented by Schiffauer and co­ (1.3 lb/ft3), the exotherm may limit the amount of pow ­
workers [2]. Our goal here is not to replace these methods, dered urethane that can be safely added to the formulation
but to offer a flexible, simple, and physically meaningful in this way.
way to add the influence of recycled powder to an existing An existing density prediction, which ignores the pow­
model. That is, we aim to preserve the structure of the der in the formulation, may be a complicated empirical

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 11

equation in use at a foam manufacturer, or an answer given Figure 3 shows the excellent agreement between the
by FoamKit™, or a simple expression like: model and experimental data. The experimental data in
this figure include data from several full-scale trials and
P = (6.25 lb/ft3) / P ^ o (1) other data that were not used in setting the values of the
When there is powder in the formulation, the density softening coefficients.
will be increased by a factor of l/(l - w), where w is the
weight percent of recycled powdered urethane in the final SAG Factor
foam. That is, w = Ppo^dJiPj - P&&ioss)­ With powder, the
modified density prediction is: The SAG factor is the ratio of ILD at 65% deflection to
the ILD at 25% deflection. The predicted SAG factor is
p' = p / ( l - w ) (2) therefore derived from the individual ILD predictions.
Figure 4 shows that there is good agreement between the
where p is from Equation 1, or some other existing density model prediction and experimental data.
prediction. This form allows maximum flexibility because The data in Figure 1 and some of Table 1 show that SAG
factor increases when powdered urethane is added to the
existing experience is captured in the expression for p, for
which Equation 1 is just a simple example. Equation 2 formulation, even when overall hardness is decreased.
relies on the assumption that the specific gravity (not foam When hardness is not maintained, SAG factor increase is
simply due to the reduction of both 65% ILD and 25%
density) of the polyurethane material in the recycled pow ­
der closely matches that of the new polymer. ILD. However, when foams are reformulated to maintain
Table 1 shows the results of this approach applied to the hardness, such as shown in Table 3 and Figure 5, it is clear
data in reference [3]. The fit is good, with error within the that the SAG factor does show a real improvement. In that
range of typical experimental error, and uses no additional case, the 65% ILD increases, and the 25% ILD decreases
adjustable parameters. Rather than use Equation 1 for the slightly; the SAG factor improves from 2.4 to 2.9. These
density model, for the results in Table 1 we used a density changes indicate significant improvement in the actual
model commonly used by the foam -machine operators comfort of a cushion made from file reformulated foam.
where the trials took place.
Figure 1 shows the excellent agreement between the Maximum Block Tem perature
model and experimental data. Additional data are provided
in Figure 1 from other foam trials not shown in Table 1. When powdered urethane is added to a formulation,
there are two competing effects on the maximum block
Hardness temperature. First, in order to maintain density, water and
isocyanate are increased. This contributes additional heat
Without correcting a foam formulation by means of of reaction. Second, the powder is nearly completely inert,
higher index, polymer polyol, or crosslinkers, the addition does not contribute a significant heat of reaction, and pro ­
of recycled powder tends to soften the foam. This is illus­ vides additional heat capacity to the reacting mixture.
trated in Figure 2, which presents the selected data from We make the assumption that the powdered urethane has
Table 1 that were used to formulate a model of this soften­ a heat capacity that closely matches that of the new poly­
ing effect. Predictions of ILD results were accomplished mer and the reacting mixture. This allows a simple estima­
using Schiffauer’s model [2] rescaled to match the results tion of the maximum block temperature using the initial
from control foams without powdered urethane. That temperature (T0), an existing prediction (ignoring the pow ­
model takes into account levels of water and methylene der) of maximum block temperature (T), and the weight
chloride, isocyanate index, and percent solids from poly­ percent powder in the foam (w).
mer polyols. However, like with the density prediction
above, any reasonably accurate model for hardness could T = T (1-w) + T0w (4)
be used.
Figure 2 visually suggests the simplest way to empiri­ As was the case with the density model, no additional ad ­
cally account for the effect of powder—that is, to consider justable parameters are required.
that each part of powder added to the formulation linearly Figure 6 shows good agreement between predicted and
decreases ILD by some percentage. By analogy to Equa ­ actual maximum block temperatures within about 4°F. For
tion 2, the modified hardness prediction is: these predictions, the value of T was determined using
FoamKit, and the maximum block exotherm was calcu­
H' = H (1 ­ (3) lated from Equation 4.

where H is a suitable prediction of hardness (ignoring the

powder) such as Schiffauer’s model [2], and ks is an ad ­
justable parameter we define as the softening coefficient.
Values of K were determined by fitting to the data in the
first five columns of Table 1, and are shown in Table 2.

1 2 / Villwock
 Table 1. Experimental results [3] and model predictions for slabstock foams.
FORMULATIONS (pphp)
Mobius Powdered Urethane 0 15 20 25 30 0 20 20 20
VORANOL 3322 polyol (48) 100 100 100 100 100 100 100 100 -

VORALUX HL-106 polyol (43.7) - - - - - - - - 100

total water 4.00 4.70 4.80 5.00 5.20 4.00 4.51 4.51 4.51
Amine catalysts 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16 0.16
Silicone surfactants 1.00 1.00 1.00 1.00 1.00 1.00 1.10 1.10 1.10
Tin catalysts 0.20 0.25 0.25 0.25 0.30 0.20 0.25 0.25 0.20
VORANATE T-80 50.8 58.5 59.2 61.4 64.7 50.8 56.2 58.7 58.5

EXPERIMENTAL RESULTS
density (lb/ft3) 1.53 1.59 1.53 1.56 1.59 1.58 1.53 1.59 1.51
ILD 25% (lb/50 in2) 27.5 25.2 22.3 20.3 19.3 26.3 22.2 24.3 —

ILD 40% (lb/50 in2) 34.4 32.4 28.8 26.5 25.4 31.9 28.8 31.0 32.8
ILD 65% (lb/50 in2) 65.6 66.0 61.3 56.1 56.2 62.0 61.3 65.8 —

SAG factor (65/25) 2.4 2.6 2.7 2.8 2.9 2.4 2.8 2.7 2.7

MODEL PREDICTIONS
density (lb/ft3) 1.54 1.54 1.57 1.58 1.58 1.54 1.58 1.57 1.57
ILD 25% (lb/50 in2) 26.8 23.2 21.9 20.8 20.4 26.8 22.5 24.2 24.6
ILD 40% (lb/50 in2) 32.1 29.2 28.0 27.1 27.2 32.1 28.7 30.9 31.4
ILD 65% (lb/50 in2) 64.2 59.8 58.1 56.8 57.2 64.2 59.1 63.2 64.1
SAG factor (65/25) 2.4 2.6 2.6 2.7 2.8 2.4 2.6 2.6 2.6
maximum block temperature (°F) 322 319 314 312 312 322 306 308 308

Table 2 . Values o f the softening coefficient, ks.

property ks
ILD 25% (lb/50 in*) 6.67%
ILD 40% (lb/50 in2) 3.70%
ILD 65% (lb/50 in2) 2.56%

Table 3 . Model predictions for slabstock foams reformulated for constant density and constant 40% ILD .
Mobius Powdered Urethane 0 15 20 25 30
VORANOL 3322 polyol (48) 100 100 100 ___ ___

VORALUX HL-106 polyol (43.7) — — — 100 100

total water 4.00 4.70 4.80 5.00 5.20
Amine catalysts 0.16 0.16 0.16 0.16 0.16
Silicone surfactants 1.00 1.00 1.00 1.00 1.00
Tin catalysts 0.20 0.23 0.24 0.25 0.26
VORANATE T-80 index 110.2 116.9 119.5 110.2 112.2
density (lb/ft3) 1.54 1.54 1.54 1.54 1.54
ILD 25% (lb/50 in2) 26.8 25.6 25.1 24.6 24.1
ILD 40% (lb/50 in2) 32.1 32.1 32.1 32.1 32.1
ILD 65% (lb/50 in2) 64.2 65.2 65.8 68.2 68.8
SAG factor (65/25) 2.4 2.6 2.6 2.8 2.9
maximum block temperature (°F) 322 320 320 314 314
chemical-cost savings (%) - 4.0% 5.0% 3.0% 4.1%

Villwock/ 13
 Figure 1. Experimental results (actual values) vs. Figure 2. Effect o f powdered urethane on hardness of
model results (predicted values) for foam density (lb/ft3). a conventional 1.5 - pcf slabstock foam without reformu ­
lating to improve hardness.

Figure 3. Experimental results (actual values) vs. Figure 4. Experimental results (actual values) vs.
model results (predicted values) for foam hardness model results (predicted values) for SAG factor
(ILD, lb/50 in2).1
4

14 / Villwock
 Figure 5. Model results showing the effect o f powdere Figure 6. Experimental results (actual values) vs.
urethane on hardness o f a conventional 1.5-pcf slab- model results (predicted values) for maximum block
stock foam while reformulating to maintain hardness . temperature (°F).
The data are from Table 3. SAG factor ; the ratio o f 65
ILD to 25% ILD improves with the addition o f M-PU.

improve SAG factor, and realize significant chemical- cost

REFORM ULATION savings. For the calculation of chemical-cost savings, the
values in Table 4 were used. The cost of Mobius Pow ­
Powdered urethane takes up volume in the reacting dered Urethane has been adjusted down from $0.42/lb in
foam, but does not participate in the production of C 0 2 to our previous paper [5] to $0.25/lb to reflect the recent drop
blow the foam. The powder also contributes heat capacity in price for scrap foam. This recent price change is consis­
to the reaction without contributing heat of reaction. Be ­ tent with a recent market analysis [6] for which the prem ­
cause of this, without an increase of water, the resulting ises have not significantly changed today.
foam will be more dense and have a lower exotherm. To
maintain density, one could increase the level of alternative Table 4 . A ssum ed co s t o f ra w m aterials fo r th is w ork .
blowing agent, but this does not address the lower exo­ assumed cost ($/lb) [6]
component
therm and would cool the foam further. Fortunately, in ­
polyol 0.60
creasing the water level reduces the density and increases
polyol + 10% SAN 0.65
the temperature. A slight increase on the order of 0.01 to
powdered urethane 0.25
0.04 parts of water per part of powder is typically required.
TDI 0.90
Isocyanate requires a corresponding increase.
surfactant 2.50
Although the additional water provides some of the heat
amine 5.50
necessary to heat the added powder, the exotherm usually
tin 2.50
is lower than that of the unfilled formulation [4]. This con ­
tributes to the effect that foams tend to be less firm (as
measured by 25% or 40% ILD) when powder is added to M ETHYLENE CHLORIDE REDUCTION
the formulation. One can correct for this and maintain
hardness in several well known ways, including raising the The Mobius Process can offer a dual environmental
isocyanate index or increasing die amount of polymer benefit for low -density flexible polyurethane foams blown
polyol solids. with methylene chloride— recycling and reduction of me­
Table 3 and Figure 5 illustrate how the simple models thylene chloride emissions. When recycled powder is
presented here can be used to reformulate a conventional added to a foam formulation, foam density is generally
slabstock foam to maintain density, maintain hardness, maintained by adding additional water and isocyanate, not

Villwock / 1 5
by adding methylene chloride. As a result, the net amount Flexible Foam”, Proceedings of the API Polyurethanes
of methylene chloride used to make the same amount of Conference, Boston, MA, Oct., 2000.
foam is reduced when Mobius Powdered Urethane is used. 6. Jeff Jensen, “Real-World Economics of Polymethane
Exotherms are about the same because the powder removes Foam Recycling,” Proceedings of the Polymethane
heat from the reaction. Low -density foams with Mobius Foam Association, Arlington, VA, Spring 2000.
Powdered Urethane also save money because the powder
costs significantly less than the other chemicals used in NOTATION
foam manufacture. This savings remains even though the
reformulated foam has more isocyanate in the formulation. H foam hardness (ILD) with powder, lb/50 in2
For example, a 1.2-pcf foam reformulated with 15 pphp of H foam hardness (ILD) without powder, lb/50 in2
Mobius Powdered Urethane can provide a cost savings of ks softening coefficient, 1/pphp
about 4% while reducing methylene chloride emissions by P g a s loss parts of gas loss in a formulation, pphp
about 14% (about 900 pounds of reduction per million P h2o water content in a formulation, pphp
board feet of foam). Ppowder M -PU content in a formulation, pphp
Pt total parts in a formulation, pphp
CONCLUSIONS T maximum block temperature with powder, °F
T maximum block temp, without powder, °F
The addition of powdered urethane to a formulation gen ­ To initial temperature of foam formulation, °F
erally requires slight modification to the formulation to w weight % of powder in a foam, dimensionless
maintain density and hardness. The SAG factor is gen ­ p' foam density with powder, lb/ft3
erally increased. The maximum block temperature due to p foam density without powder, lb/ft3
the reaction exotherm can be kept within reasonable limits.
Typically there is a reduction in the emission of alternative
BIOGRAPHY
blowing agents such as methylene chloride. Significant
cost savings can be realized by displacing more expensive
R obert Villwock
chemicals with powdered urethane.
Simple models were presented that allow prediction of Robert Villwock received a B.S.E.
key foam properties when using recycled powder. These in Polymer Science and Engineer­
include density, hardness, SAG factor, and maximum ing from Case Western Reserve
block temperature. These models were presented in a way University and a Ph.D. in Chemical
that allows them to be used along side any existing predic ­ Engineering from the University of
tive model of foam properties in use at a specific location. California at Berkeley. As a Re­
This allows maximum flexibility and takes advantage of search Engineer for The Dow
the existing knowledge captured by those models. Chemical Company, he contributed
to the research and development of
REFERENCES new thermoplastics, including sty­
rene terpolymer foams, polyethylenes, and recycled plas ­
1. FoamKit is a trademark of TG Cellsoft Ltd. tics. He was a key part of the Applied Plastics and Materi­
2. R. Shiffauer, “Mathematical Property Prediction Mod­ als R&D group, which launched Dow’s first effort into
els for Flexible Polyurethane Foams— A Comparison plastics-recycling R&D in 1989. Since 1999, he has been
between Conventional Slabstock, High Resilience Director of Polymer Engineering with Mobius Technolo­
Slabstock and High Resilience Molded Foams”, Pro ­ gies, Inc.
ceedings of the 35th Annual Polyurethane Techni-
cal/Marketing Conference, Oct. 9-12, 1994, pp. 225- Mobius Technologies is the leading supplier of turnkey
238. recycling systems for the polyurethane foam industry. Lo ­
3. P. Berthevas, F. Aguirre, and J. Tu, “Recent Devel­ cated near the San Francisco Bay Area in northern Cali­
opments in the Recycling of Flexible PU Foams back fornia, Mobius maintains engineering, laboratory, pilot
into New PU Foam Products”, API Polyurethanes plant, manufacturing and administrative facilities at its
Expo, Columbus, OH, Sept. 30 - Oct. 3, 2001, in Grass Valley headquarters. The facilities are equipped to
press. handle all aspects of polyurethane recycling development,
4. S. Venkatasanthanam, “Utilization of Powdered Poly­ from initial laboratory trials to pilot plant demonstrations
urethane Foam Scrap as a Filler in Producing Polyure­ and commercial scale-up. Mobius maintains active re ­
thane Foam: Processing and Properties,” M.S. Thesis, search and development programs in size reduction, classi­
Illinois Institute of Technology, Chicago, IL (1995). fication, material handling, rheology, blending science, and
5. Herman Stone, Robert Villwock, and Biyan Martel, urethane material science. These resources are available to
“Recent Technical Advances in Recycling of Scrap their customers.
Polyurethane Foam as Finely Ground Powder in

16 / Villwock
Recent Developments in the Recycling of Flexible PU Foams
Back into New PU Foam Products

P. BERTHEVAS and F. AGUIRRE J. TU

Dow Europe S.A. The Dow Chemical Company
International Development Center 2301 North Brazosport Blvd.
13, rue de Veyrot— P.O. 3 Freeport, Texas 77541
CH -1217 Meyrin 2 (GE)
Switzerland

ABSTRACT displayed good processability on existing PU foam mold­

ing equipment.
Two different approaches have been evaluated by The
Dow Chemical Company to recycle flexible PU foams back INTRODUCTION
into new flexible PU products: The first is the room-
temperature pulverization process as developed and com ­ The Dow Chemical Company Inc. Midland Michigan,
mercialized by Mobius Technologies California. The sec­ and Mobius Technologies Inc., Grass Valley, California
ond is a solvolysis using dicarboxylic acid developed by signed a Technology Cooperation Agreement in March
Prof. Bauer of the Aalen University of Applied Sciences in 2000 to optimize the incorporation of recycled PU powder
Germany. into new PU foam formulations and evaluate the resulting
economics.
For the recycling of slabstock PU foam production trim
waste, only the pulverization process was used which pro­ Pulverization of flexible PU foam and the use of the re­
vides a commercially attractive alternative to foam flocks sulting PU powder in the production of new flexible foam
rebonding. The process allows the recycling of trim foam is not new. This approach has been described by a number
back into new slabstock foam by introducing finely ground of workers since the mid -90’s [1-8]. In most cases how­
PU powder into the polyol and adjusting the recipes. New ever, the average particle sizes were in the 200 to 300 mi­
slabstock foam is produced with recycled foam loading of crons range and only limited processing and foam
10 to 15% by weight. performance data, usually on hand-mixed foams, were
communicated.
Formulating approaches developed to closely match
original foam performance ­ including the use of specifi The cost for the early cryogenic pulverization process
cally tailored polyols, and liquid C 02 blowing - will be was considered to be too high for economically viable in­
reviewed, as well as the influence of the PU powder filled dustrial application. Finally, room temperature pulveriza­
foams. tion processes using 2-roll-mill type of equipment were
successfully used industrially at recycling slabstock pro ­
For molded PU foams used in car seats, both pulveriza­ duction wastes and trim scrap [9-15].
tion and solvolysis have been evaluated for recycling such The PU powders used in the study were produced by
foam back into automotive molded foams for car seats, Mobius Technologies from either commercially available
head-rests, as well as noise & vibration insulation pads for slabstock production scrap or foams produced at the Dow
use under car carpets. Europe pilot facilities in Meyrin, Switzerland. The foams
were pulverized by Mobius Technologies using their spe­
The recycled foams were used at concentrations of 5 to cifically developed 2-roll-milling, sifting and separation
10% by weight of the new molded foams. Foam perform ­ processes for which patents have been filed [14].
ance met typical OEM’s specifications for such parts, and

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 17

 The resulting PU powders are described in Table 1. The blends were kept under slow agitation (30 rpm) in
the machine tank and recirculated with a Moyno pump until
Table 1. PU powders used in the evaluations the required process temperature (23 °C) was reached.

PU powder # 1 Trim scrap from conventional slab foams For the highest powder level (30 pphp) the blend was
Sieved to achieve a top -cut of 125 microns
maintained overnight under the above-mentioned condi­
PU powder # 2 HR molded foam produced on Dow pilot
tions in order to achieve enough degassing.
Sieved to a top -cut of 125 microns
The PU powders pick up moisture during storage, and
PU powder # 5 HR slab foam produced on Dow pilot water content typically increases from < 0.5% when the
Sieved to achieve a top -cut of 125 microns powder is freshly produced to 1.5 - 2 % after prolonged
storage under typical warehouse conditions (20°C and 80%
PU powder # 8 Trim scrap from conventional slab foams
Sieved to achieve a top -cut of 75 microns
RH).
The water content of polyol/PU powder slurries was
checked using a Karl-Fischer titrator and the results were
This report summarizes the effect of such fine polyurethane
taken into account when adjusting the output of the addi­
powders on the processing and physical properties of flexi­
tional water stream so that total water (pphp) during the run
ble slabstock and molded foams produced on Dow Europe
corresponded to that indicated in the recipe.
pilot foaming machines at its International Development
Center in Meyrin, Switzerland.
Samples of the blends were kept under slow rotation on a
paint-rolling machine during 24 hours for degassing.
The experimental part and the results are split into two
Viscosity was measured at 25 °C using a Brookfield Vis­
sections, the first one covers the use of PU powders from
cometer Model LV.
recycled flexible foams back into slabstock foams and the
second the use of such powders into molded PU foams for
Although densities from 17 to 30kg/m3 were produced,
the automotive industry such as seat cushions, headrests
most of the evaluation was done with 25 kg/m3 density
and NVH foams.
foams as this represents the largest volume grade in the
global slabstock market.
The second part also compares this recycling approach
via pulverization to the solvolysis with dicarboxylic acid.
Foaming trials were carried out on a low-pressure Poly-
mech® machine with a polyol output of 20 kg/min and a
A - SLABSTOCK FOAMS
mixer speed of 4500 rpm and for liquid CO2 blowing trials
on a Hennecke UBT machine fitted with a Cardio ® sys­
The objective here was to define the required modifica­
tem, with a polyol output of 40 kg/min.
tions to the recipes containing PU powder, so that the re ­
sulting slabstock foam would match as much as possible
After 3 days the foams were tested using the test meth-
the characteristics of the corresponding unfilled reference
ods/Standards listed in Table 2.
foam, especially density, hardness and porosity.
Table 2.
EXPERIM ENTAL
Property Standard
Density Core DIN EN ISO 845 - 95
Voranol ® 3322, a 3500 MW , PO/EO mixed feed triol,
Hardness ,CFD DIN EN ISO 3386-1-98
Voralux ® HL 106 a 10% SAN copolymer polyol and Hardness, IFD ISO 2439 - 97
Voralux ® HL 109 A 15% SAN copolymer polyol were Tensile Strength ISO 1798097
used in this study. Elongation ISO 17980 - 97
The same standard blowing and gelling catalysts as well Tear Strength ASTM D 3574 - 95
as silicone sufactant were used throughout the trials. Resilience ASTM D 3574 - 95
The PU powders were dispersed in 100 kg of polyol at Air Flow ASTM D 3574 - 95
various loading levels (15, 20, 25 and 30 pphp) in an open Com pression Sets DIN EN ISO 1856 - 80
220-L drum using an “impeller-plus-ring” type disperser Dynamic Fatigue BS EN ISO 3385-95
(diameter 30 cm) at 420 rpm. Image Analysis LEICA

The resulting blends were kept under slow agitation (60

rpm) during at least 12 hours to achieve sufficient de ­ RESULTS
gassing. They were then transferred into the water-jacketed
stainless steel machine tank using a membrane pump. Reformulating to Produce Close M atches to the Origi ­
nal Foams.

1 8 / Berthevas, Aguirre, Tu
 The foaming trials indicated good processability al­ Foam hardness (40% IFD) decreases proportionally to
though the mass-flowmeter for polyol had to be recali­ the amount of PU powder in the formulation as shown in
brated due to the higher viscosity of polyol/PU powder Figure 2.
blends.

Polyol/PU powder blend viscosity increases rapidly

when powder loading exceeds 15 %.
It is interesting to note that the finer PU powder # 8
gives much lower viscosity than the coarser PU powder #
1, almost half at higher loading (> 15%) Figure 1.

Figure 2. Guide factor vesus PU powder content at 25-

kg/m3 density

SAG or Support factor increased from 2.4 for the un­

filled foam up to 3.0 for the foam with 20 pphp of PU
powder at 25 kg/m3 (Figure 3).
Figure 1. Voranol ® 3322 / PU powder blend viscosity

To achieve accurate metering to the mixing head and

controlled foam nucleation, hence reproducible results, it
is important to have an efficient degassing process.
Mobius have developed continuous mixing and de ­
aerating units to feed polyol/PU powder blends to a com ­
mercial-scale slabstock machine, but such units were not
used for these pilot trials.

A number of changes had to be made to the recipes when

the PU powders were added, in order to match the density
and foam porosity of the control foams, which was our first
objective:
the water level, pphp was increased in line with the PU Figure 3. SAG factor versus PU powder content at 25 kg /
powder content to blow the foam to the same density. m3 density

PU powder acts as a cell opener. The tin catalyst (pphp) Resilience decreased from 44% to 40% ball rebound for
had to be increased in proportion to the powder content in 20 pphp of PU powder and then remained constant up to 30
the formulations to maintain the same foam porosity / air pphp of PU powder (Table 3).
flow.

Beyond 20 pphp of PU powder in the formulations,

(12% on the foam), foam densification was detected, and
the surfactant was increased to avoid densification in the
blocks.

Foam Physical Poperties

According to results listed in Table 3, when comparing

PU powder filled foams with the corresponding unfilled
reference foams at similar density, permeability and TDI
Index, we can conclude that:

Berthevas, Aguirre, Tu / 1 9
 Table 3. Recycled PU powder in slab foam Table 5. Cardio ® CQ2 blown low density foams
Formulation (pphp)
Voranol 3322 100 100 100 100 100
Formulation (pphp)
Mobius PU powder # 1 0 15 20 25 30
(pph polyol/powder blend)
Water ( total) 4.0 4.7 4.8 5.0 5.2 Voranol 3322 89 0
Amine catalysts 0.16 0.16 0.16 0.16 0.16 Voralux HL 109 0 89
Silicone Surfactant 1.0 1.0 1.1 1.1 1.2 Mobius' Powder # 1 11 11
Stannous Octoate 0.20 0.25 0.25 0.25 0.30 Water ( total) 4.70 4.20
Voranate T-80 ( Index 110) 50.8 58.5 59.2 61.4 64.7 Amine catalysts 0.12 0.12
Foam properties Silicone surfactant 1.30 1.30
Density (d) kg/m3 24.4 25.5 24.4 25.0 25.4 Stannous octoate 0.18 0.20
Hardness, 40% IFD N 153 144 128 118 113 Liquid C02 2 2
Guide factor,(IFD40%/ d) 6.30 5.64 5.23 4.71 4.46 Air injected in polyol line l/mn 13 25
SAG factor ( IFD 65/25) 2.40 2.62 2.74 2.76 2.91 Shim mu 300 350
Hardness, 40% CFD kPa 3.8 3.7 3.2 2.9 2.9 Voranate T-80 [ Index ] [107] [112]
Air Flow scfm 4.3 2.5 3.6 3.4 2.9 Voranate T-80 [ parts] 55.8 51.8
Resilience ( ball rebound) % 44 41 40 40 40 Foam properties
Elongation % 173 182 213 224 202 Density kg/m3 17.7 19.6
Tensile strength 110 108 121 110 112 CFD 40% kPa 2.0 3.4
Tear strength N/m 368 365 405 460 387 Air Flow scfm 6.3 2.3
Compression set (75%) % 3.5 4.6 6.9 9.4 7.2 Resiliency (ball rebound) % 43 33
IFD Iso 40% N 78 123
Guide Factor (IFD 40%) 4.4 6.3
Mechanical properties, Tensile, Tear and Elongation SAG (IFD 65/25) 2.4 2.5
were statistically unchanged. Tear Str. N/m 391 335
Compression sets increased with PU powder loading but Elongation % 213 139
were still acceptable especially for PU powder not exceed­ Tensile kPa 90 97
Comp. Set 75 % % 4 9
ing 20 pphp (typically c.a 12% in the new foam).
Effect of PU Powder Particle Size
As it was found that the addition of PU powder produces
softer foams, the isocyanate index was increased or the
Two PU powders produced from conventional slabstock
polyol replaced by a 10% solids copolymer polyol Voralux
trim wastes were compared:
® HL 106 to recover hardness (Table 4)
The first one was pulverized and sifted to a top-cut of
125 microns (powder # 1), the second to a top-cut of 75
Table 4. Re-formulating to recover hardness .
microns (powder # 8).
Form ulation ( pphp)
Voranol 3322 100 100 100 They were evaluated in typical 4 pphp water conven­
Voralux HL 106 100
Mobius PU powder #1 20 20 20 tional slabstock formulations based on Voranol ® 3322, by
Water ( total) 4.0 4.8 4.8 4.8 replacing the polyol with mixtures of polyol and PU pow ­
Amine Catalysts 0.16 0.16 0.16 0.16 der, respectively 90/10, 85/15, 80/20 and 75/25.
Silicone surfactant 1.0 1.1 1.1 1.1
0.20 0.25 0.25 0.20
Interestingly, the polyol/PU powder viscosity was almost
Stannous Octoate
Voranate T-80 [ Index ] [110] [110] [115] [110] halved when using the finer PU powder as reported earlier,
Voranate T-80 ( parts) 50.8 59.2 61.9 58.5 which is very positive from the processing point of view.
Foam properties (Figure 1).
Density ( d ) kg/m3 24.4 24.4 25.4 24.1
Hardness, 40% IFD N 153 128 138 146 As for the resulting foam physical properties, the coarser
Guide factor,( IFD40%/d) 6.30 5.23 5.44 6.07 powder #1 gave slightly higher hardness but mechanical
Hardness, 40%CFD kPa 3.8 3.2 3.4 3.6 characteristics such as elongation, tensile strength and tear
SAG factor ( IFD 65/25) 2.40 2.74 2.71 2.75
3.6 3.7 2.7
strength were slightly higher with the finer powder # 8 as
Air Flow scfm 4.3
Resilience ( ball rebound) % 44 40 40 37 well as resilience, as illustrated in Table 6.
Elongation % 173 213 181 182 These trends will be checked further by evaluating PU
Tensile strength kPa 110 121 105 141
powder with top -cut at 45 microns.
Tear strength N/m 368 405 322 420
Compression set ( 75%) % 3.5 6.9 5.7 6.7

Similar trends were observed for all foam densities.

Soft foam with densities as low as 17.7 kg/rmwere produced

using the Cardio ® liquid C 02 blowing process and 19.6
kg/m3 hard grade using Voralux HL 109 copolymer polyol
(Table 5).

20 / Berthevas, Aguirre, Tu
Table 6. Effect of PU powder particle size on slab foam Table 7. PU powder from conv. slab foam vs PU powder
properties from HR Slab foam.

Formulation Formulation (pphp)

(php of polyol/PU powder blend)
Voranol 3322 100 100
Voranol 3322 100 90 90 80 80
Water ( Total) 4.8 4.8
Moblus PU powder #1 10 20
(125 mu) Amine catalysts 0.16 0.16
Moblus PUpowder # 8 10 20 Silicone surfactant 1.1 1.1
(75 mu) Stannous Octoate 0.22 0.22
Water ( total) 4.0 4.0 4.0 4.0 4.0
Moblus PU powder # 1 20
Amine catalysts 0.16 0.16 0.16 0.16 0.16
Moblus PU powder # 5 20
Silicone Surfactant 1.0 1.0 1.0 1.0 1.0
Stannous Octoate 0.20 0.20 0.20 0.20 0.20 Voranate T-80 ( index =110) 59.2 59.2
Voranate T-80 [lndex=110] 50.8 50.0 50.0 49.1 49.1 Foam properties
Density kg/m3 23.1 23.3
Foam properties
Density kg/m3 24.4 23.5 22.9 22.3 23.7 Hardness, 40% IFD N 130 123
Hardness, 40% IFD N 153 134 125 115 105 Guide factor,( IFD40%/density) 5.63 5.28
Guide factor,( IFD40%/density) 6.3 5.7 5.5 5.1 4.4 SAG factor ( IFD 65/25) 2.7 2.7
SAG factor ( IFD 65/25) 2.4 2.5 2.5 2.7 2.7 Hardness, 40% CFD kPa 3.27 3.02
Hardness, 40% CFD kPa 3.8 3.0 3.0 2.8 2.5 Air Flow scfm 3.5 3.3
Air Flow scfm 4.3 4.6 4.5 4.1 4.5 Resilience ( ball rebound ) % 41 40
Resilience ( ball rebound) % 44 41 42 38 40 Elongation % 185 161
Elongation % 173 143 173 144 179
Tensile strength kPa 112 96
Tensile strength kPa 110 94 118 98 107
Tear strength N/m 400 353
Tear strength N/m 368 344 427 313 329
Compression set (75%) % 3.5 3.1 3.1 4.7 3.9
Compression set (75% ) % 4.9 5.1

Morphology of the PU Foams Containing Recycled PU

Effect of the Type of Foam Used to Produce the Powder powder

PU powder # 1 derived from conventional slabstock trim Observations on thin slices of the produced PU foams at
scrap was compared to PU powder # 5 which was derived c.a 25 kg/m3 with increasing PU powder loadings using a
from HR slabstock foam but ground to the same particle Leica® optical microscope indicated that at up to 20 pphp
size with a top-cut at 125 microns. As can be seen in table- not much distortion to the original foam structure was ob­
7, no significant differences could be noticed between the served although cell counts indicated a slightly higher pro ­
two resulting 4.8 pphp water blown conventional slabstock portion of coarse cells.
foam filled with 20 pphp of PU powder, with the possible
exception of mechanical properties, elongation, tensile and Cells counting was done using the Leica ® Q500MC Im­
tear strength which were lower for the foam containing the age Processing and Analysis System linked to the micro­
HR foam pow der. scope via a video camera. The data collected was further
The TDI based HR slabstock formulation used to pro ­ processed using an Excel ® program The results are in
duce the powder and its foam properties were similar to Table 8.
typical TDI based HR molded foam for car seats which
indicates that such seat foams could be recycled into PU Table 8. Cell structure vs PU powder loading
powder for processing back into conventional slabstock Cell Size Range (mm3) 0 pphp 20 pphp 25 pphp
Mean Value PU powder PU powder PU powder
foam. 0.000 —0.104 31.9 29.1 26.8
0.104 - 0.504 59.1 60.4 59
0.504 - 2.000 8.9 10.5 14.1

Cells Count 551 505 452

(Unit Area)

The trend was more pronounced at higher loading (25

pphp) with a clear indication that PU powder was then ac­
cumulating at the joints or Plateau borders (Figure 4)

Berthevas, Aguirre, Tu / 21
 B - PULVERIZATION AND CHEMOLYSIS ARE
USED TO RECYCLE FLEXIBLE PU FOAM BACK
INTO AUTOMOTIVE MOLDED FOAMS.

The need to find ways to recycle PU foams after their use­

ful life e.g in cars has been recognized by the Polyurethane
Industry early on, both in the US via the creation of the
Polyurethane Recycling and Recovery Council (PURRC)
in 1990 [16,17,19], as well as in Europe by the creation of
the Recycling Task Force within ISOPA [18,20], and in
Japan by JURA, the PU Industry Association.
Since then, various routes have been investigated by all
players in the industry, some using mechanical processes,
either flocking and rebonding, or comminuting the old PU
foam to a fine powder and adding it to the new foam for­
Figure 4. Cell structure with 25 pphp of PU powder mulation [1-15] or using some sort of chemolysis process
to break down the old PU foam and introduce the chemo-
Formulation Costs lysate into new PU formulations e.g for flexible or semi­
rigid molded foams [21-28].
Foam costs were calculated taking the cost for the Mo-
bius PU powder as the sum of market value of production Many problems were encountered however regarding both
foam scrap (e.g. for rebonding) plus the cost to pulverize it, the processing side and the resulting foam performances.
taken as 17 c/lb. [11] The chemolysis processes selected, i.e hydrolysis, glycoly­
The market value of production foam scrap for rebonding sis or aminolysis in general worked well for selected foams
was varied from 0 to 50c/lb.To demonstrate the effect of not containing SAN coming from copolymer polyols used
polyol price on this approach to recycle scrap foam back in original foam recipes. However, the resulting recyclates
into new foam, the polyol price was varied from 50 to 60 contained large amounts of short chain polyols and were
c/lb. hence not suitable for using directly back in flexible foams.

Cost calculations indicated that foams with 20 pphp of Further separation had been worked out e.g with the “tan­
PU powder and formulated to produce the same density (25 dem chemolysis” process [21] leading to a flexible polyol
kg/m3) and hardness as the corresponding reference foams, fraction c.a 1/3 which could be used back in flexible foam
were lower in cost than the reference foams, and depending formulations, and a larger fraction, c.a 2/3 made of short
on prevailing market and operating conditions, this differ­ chain aromatic rigid polyols which could be used in rigid
ence can vary between 0 and 6% as illustrated in the at­ foams.
tached Figure 5. However, in real life post-consumer PU foam wastes such
as car seats, SAN is almost always present as SAN type
copolymer polyols have been used for the last 30 years to
produce TDI based HR molded foams and hot molded
foams, and in some cases smaller amounts were also used
in MDI based foams, which makes it extremely difficult to
use above chemolysis routes.

A different process, solvolysis with dicarboxylic acids was

more recently claimed to be able to recycle SAN contain­
ing foams, giving an homogeneous recyclate under the
form of a stable dispersion of oligomers and SAN in a new
polyol used in the process.
F ig u re 5. Influence of polyol and scrap foam price on This process was developed and patented by Prof. Dr.
slabstock production cost savings (10% PU powder, 25 Gunter Bauer of the Aalen Technical University in Ger­
kg/m3) many [23,28] and is being used today by the company Re-
gra Recycling GmbH in Pirmasens, Germany.

22 / Berthevas, Aguirre*Tu
EXPERIMENTAL Table 10. Recyclates Obtained by Solvolysis of HR
Molded Foams for Car Seats
Most of the work covered here for molded foams also in­
1) Reactor Loading
volved flexible slabstock and molded foams derived PU
powders using Mobius Technologies pilot pulverization From TDI based From MDI based
p lan t, and adding the PU powder to the formulated polyol HR molded foam HR molded foam
blend for the production of new molded foams for seats, Parts Parts
head-rests and NVH pads . PU foams 100 PUfoam 100
Dicarboxylic add 20 Dicarboxylic add 20
Stabilizer 4 Stabilizer 4
Solvolysis was only used for car seat foam recycling, which Process additives 4 Process additives 4
Vorano® CP-6001 72 Voranol® CP-6001 72
for this study were Dow pilot plant produced MDI based 200 200
foams not containing SAN from copolymer polyols, and 2) Analytical data
TDI based foams containing SAN. (Table 9).
Solvolysate #1 Solvolysate #2
The two recycling approaches for car seat foams were Hydroxyl number 40 ± 10 45 ±10
(mg KOH/g)
compared for processability and final foams performances (active Hydrogen)
versus OEM specifications as well as for economics. Add number 3 ± 0,5 3 ±0,5
(mg KOH/g)
Viscosity mPa*s 4900 ± 500 7600 ±500
Table 9. Seat molded foams solvolysed Aromatic Amines % 0,08± 0,02 0,07 ±0,02

Type of HR foam TDI MDI MDI Molding polyols used were the 6000 mw triol Voranol®
CP 6001, the high functionality polyol Specflex® NC 632
Formulations (pphp)
Polyol Voranol CP-6001 10 100 100 an the 40% SAN copolymer polyol NC 700. Other polyols
Polyol Specflex NC 632 53 used as cell opener / processing aid were Voranol® CP-
Copolymer Polyol Specflex NC 700 37 1421 and Voranol ® 4053.
Cell opener Voranol CP-1421 1 Isocyanates were TDI Voranate® T -80 and Puromeric
Cell opener Voranol 4053 2.5 2.5 type MDI Specflex® NE 112
Water 3.50 3.60 4.20 The chemolysates were added to the polyol side of the
DEOA 1.20 0.50 0.50 TDI based HR molded foam for car seat, replacing 10 parts
Amine catalysts 0.55 0.40 0.65 of polyol Specflex® NC 632 and 10 parts of copolymer
Surfactants 1.00 0.80 0.80 polyol Specflex® NC 700 by 20 parts of the chemolysate
Voranate T-80 ( I = 100) 42.5 as indicated in Table 11.
Specflex NE 112 (I =100) 61.5 72.5
Overall molded density, kg/m3 32 45 39

The foams were solvolysed by Regra Recycling GmbH

after optimization of the process by Prof. Bauer at the
Aalen Universisty of Applied Sciences using Voranol ®
CP-6001 as the carrier polyol as indicated in Table 10 .The
resulting recyclates (Table 10) were added to the polyol
blend to produce new TDI based car seat foams.

Berthevas, Aguirre, Tu / 23
Table 11. Solvolysates from seat foams are used Diethanolamine, silicone surfactant and amine catalysts
back in molded seat foam . were then added using the same mixer as described in the
first part for slabstock.
Formulations ( p p h p )
Polyol, Specflex NC 632 65 55 55
Copolymer Polyol, Specflex NC 700 35 25 25 The TDI based HR molded foams for car seats were run
Solvolysate # 1 { FromTDI HRfoam) 0 20 0 on a high pressure Admiral machine equipped with a
Solvolysate # 2 ( From MDI HR foam) 0 0 20 Krauss-Maffei Duplex MK - 12-12/16 UL mixing head. The
Water ( Total) 3.6 3.4 3.4 filters before the pumps were removed to avoid plugging
DEOA 1.1 1.5 1.5 with the PU powder.
Amine catalysts 0.30 0.50 0.35
The isocyanate temperature was 20°C and pressure 150
Silicone surfactant 1.0 1.0 1.0
TDI ( Index) (95) (94) (93) bars. The pressure in the line was maintained at the se­
TDI, php 40 39 39 lected 160 bars by changing the polyol injector when the
Recycled PU foam, %of the final foam 0 7.3 7.3 PU powder was used.The polyol stream temperature was
Processing &reactivity increased from 21 to 30°C.
Polyol Temperature, oC 27 27 27 The square aluminum mold (420x420x100mm) was kept at
Gel Time, sec 70 77 75
65°C.
Rise Time, sec 160 180 160
Free rise density, Kg/m3 29 33 33
The mold release agent used was Klueber 41-2013
Flowability in the mold Good Good Good Total output was 300 or 350 g/sec.
Demold Time, mn 5 5 5
Overall Molded density kg/m3 35 36 36 The MDI based head-rest and NVH molded foams were
Crushing Force, 50% deflection N 725 185 375 processed on a “Lance” type HP molding machine Cannon
Curing factor (IFD 50/Hot IFD 50) 0.47 0.47 0.44 HE 60 described earlier.
Foam Properties Spacs
For the runs with PU powder, the “polyol” injectors were
(vs Ford Specs) changed from 1 to 2 mm.
Overall Molded density kg/m3 35 36 36
Hardness IFD25% N 88 82 84 Mold temperature was kept at 60°C +/- 2°C.
SAG factor ( IFD 65/ IFD25) >2.5 3.5 3.5 3.7 For head-rests foams, an aluminum square mold
Hysteresis loss % 21.4 22.4 21.9 (300x300x100 mm) was used.
Density core kg/m3 33 35 35 For NVH pads, a flat square “Langer” mold
Hardness CFD 40% kPa 2.2 2.1 2.3 (852x865x22 mm) was used.
Resilience, ball rebound % 69 67.5 70.5
The release agent in both cases was Klueber 918191K.
Elongation % > 100 103 90 100
Tensile strength kPa > 120 109 97 110
Tear strength N/m >220 242 214 206 Foam testing was done according to standard flexible
Dry Compression set, 50% CT % < 10 5.5 5.2 6.1 foam testing methods listed in the first part of this paper
Heat Aged Force loss, CFD 40% % <25 1.4 15.3 NA (Table 2), plus automotive molded foam specific testing
Humid Aged Force loss, CFD 40% % < +/- 30 -11.2 -10.0 -6.8 methods as listed in Table 12.
Dynamic fatigue Height loss % <5 3.3 2.9 2.8
Dynamic fatigue ,40% IFD loss % <25 20.8 23.0 19.6
Table 12 Specific seat foam tests used
Flammability mm/mn <100 DNI DNI DNI

Hardness IF D , SAG factor, Hysteresis Loss ISO 2439-97

The foams were produced on a high pressure Cannon HE Hardness CFD ISO 3386-1-98
60 “Lance” type machine equipped with a Krauss-Maffei
Elongation & Tensile strength ISO 1798
MK 12-18 UL mix-head, pouring in an aluminum square
Tear strength ISO 8067-95
mold (420x420x100mm) maintained at 65oC using as mold
Dry Compression set ISO 1856-80
release agent Acmosil 39 6153. The polyol stream was
maintained at 27oC when the recyclate was added as the Heat & Humid Aged Force loss ISO 2440-97

difference in blend viscosity was negligible. Isocyanate Dynamic fatigue, height & Force loss ISO 3385-95
stream was kept at 25oC. Flammability MVSS 302
Stream pressures were respectively 190 and 180 bars for Fogging Reflectometric
polyol and isocyanate. Airflow resistivity( NVH) Norsonic Nr 940
Total output 300 g/sec. Insertion Loss ( NVH) Petite Cabine
The PU powders used are the ones described in Table 1
Damping factor & E-modulus( NVH) BMW test rig
in the slabstock part. (modified DIN 53426)

The PU powders were dispersed in the polyol blend first

as done for slabstock. The water content of the blend was
then measured, and more water added to adjust to the level
indicated in the recipes.

24 / Berthevas, Aguirre, Tu
RESULTS Table 13. Recycled PU powder is used in TDI based
seat molded foam
From the processing standpoint, it should be noted that
PU powder addition to the polyol Formulations (pphp)
Polyol, Specflex NC 632 65 65
Copolymer Polyol, Specflex NC 700 35 35
Water ( Total) 3.6 4.1
DEOA 1.1 1.1
Amine catalysts 0.30 0.35
Silicone surfactant 1.0 1.0
Mobius PU powder # 8 0 10
TDI ( Index 9 5 ) 40 44
PU powder, % on fin al foam 0 6.2
Processing characteristics
Polyol Temperature oC 21 30
Reactivity Same Same
Free rise density kg/m3 29 29.5
Crushing Force, 50% deflection N 880 695
Hot IFD 50% N 115 105
Foam properties Specs
Figure 6. Water level effect on viscosity in the blend con ­ (vs Ford Specs)
taining 20% of PU Powder Mobius #8 Overall molded density kg/m3 35 35.4
Core density kg/m3 34.5 35

blend increases the blend viscosity, with water addition Hardness, CFD 40% kPa 3.28 3.37
Hysteresis Loss % 24.9 26.4
increasing it further until the water content reaches c.a 1.5
SAG factor ( IFD 65% / IFD 25%) >2.5 3.0 3.3
%. Thereafter the viscosity stays constant (Figure 6). Resilience, ball rebound % 66.5 67.5
Elongation % >100 103 100
For this reason, when foaming the resulting recipes, the Tensile strength kPa >120 122 142
polyol stream temperature often has to be increased to re­ Tear strength N/m >220 242 246
duce the viscosity for improved pumpability and reduce Dry Compression set, 50%, CT % <10 6.2 6.3
Heat Aged Force loss,CFD 40% % <25 7.1 8.6
pressure in the lines, and the 1mm “polyol” injector was
Humid Aged Force loss,CFD 40% % < +/-30 -2.9 4.1
replaced by a 2 or 3 mm injector. Dynamic Fatigue,Height loss % <5 3.2 3.0
Dynamic Fatigue, loss of IFD40% % <25 24.9 27.7
On the other hand, addition of the recyclates from sol­ Flammability m/mn <100 DNI 70
volysis only had limited impact on processing and same Fogging % >80 95.3 99.3
injectors and stream temperatures were used.

Regarding the molded foam characteristics versus the

corresponding virgin foams of same density, one can draw
the following conclusions:

As can be seen in Table 11, foam characteristics meeting

the Ford specifications for car seats, could be obtained with
both solvolysates when 20 parts were used which are
equivalent to 7.3% of old foam in the new foam. The only
values not meeting the specs were elongation, tensile and
tear strength as the formulation selected was already bor­
derline in that respect .Addition of the solvolysates did not
change much these characteristics .

With the addition of recycled PU powder (Table 13) at a

level of 10 pphp or 6.2% on the new foam most Ford seat
specifications could be met with this 35 kg/m3 foam, with
the exception of hardness loss after dynamic fatigue
which was already borderline with the non-filled foam.

Berthevas, Aguirre, Tu / 25
 This was also noticed for the MDI based head-rest foam
(Table 14). Table 15. Recycled PU powder is used in NVH
molded foam
Table 14. PU powder is used in MDI based head -rest
molded foam Formulations ( pphp)
Polyol Voranol CP-6001 100 100 100
Cell opener Voranol CP-1421 4 4 4
Formulations ( pphp) Water ( Total) 4.5 4.7 4.8
Voranol CP-6001 100 100 100 Amine Catalysts 0.60 0.60 0.60
Voranol CP-1421 4 4 4 Silicone surfactants 0.50 0.70 0.50
Water ( Total) 4.53 4.9 4.7 Mobius PU powder # 8 . 0 15 20
Amine Catalysts 0.75 0.80 0.80 Specflex NS 540 ( Index 7 0 ) 49.7 53.4 54.3
Silicone surfactants 0.30 0.30 0.30 PU powder, % on final foam % 0 9.0 11.6
Mobius PU powder # 8 0 15 20 Processing characteristics.
Specflex NE 112 ( Index) (85) (8 5) (95) Polyol Temperature oC 25 35 35
Isocyanate Temperature oC 25 29 30
Specflex NE 112 ( pphp) 61.9 65.5 73.9
Reactivity Same Same
PU powder, % on final foam 0 7.9% 9.8%
Demold Time sec 48 48 42
Processing characteristics
Foam properties
Polyol Temperature oC 25 36 36 Overall Molded density kg/m3 49.7 49.5 50.8
Isocyanate Temperatue oC 25 30 30 Airflow resistivity [r] N.s/m3 450 472 656
Cream Time sec 10 5 5 Hardness, CFD 40% kPa 4.0 4.7 4.1
Gel Time sec 60 46 53 Elongation % 80 78 80
Rise Time sec 92 63 78 Tensile strength kPa 47 55 44
Demold Time sec 150 150 150 Compression set, 50%, CT % 6.3 8.8 9.1
Overall Molded density kg/m3 43.3 43.3 43.3 Flammability, mm/m DNI 68 130
ii
Crushing Force, 50% deflection N 460 345 325 Humid Aged properties
Curing factor, 43 39 30 Elongation % 119 112 103
(IFD 50 / Hot IFD 50)
Tensile strength kPa 47 55 44
Foam properties Specs
Hardness loss, CFD 40% % 55 57 59
(Ford Specs) Thickness loss % 2.1 2.2 2.3
Core density kg/m3 41.6 41.3 41.7
Compression set, 50%, CT % 8.6 13.0 13.4
Hardness, CFD 40%, kPa 3.7 3.2 3.6 BMW acoustic test results
SAG Factor, ( IFD 65/ IFD 25) > 2 .5 3.1 3.8 4.2 (Modified DIN 53426)
Hysteresis loss % 28.5 35.0 31.2 E-modulus kPa 75 91 82
Resilience, ball rebound % 62 61 60 Damping factor 0.103 0.120 0.112
Elongation % 80 75 66
Tensile strength kPa 95 83 91 With 20 pphp of PU powder or 11.6% on the foam,
Tear strength N/m 129 121 151 processing was difficult due to the high viscosity.
Dry Compression set, 50% CT % <10 4.9 7.3 8.1
Insertion loss obtained in the “Petite Cabine” automotive
Heat Aged Force loss, CFD 40% % <25 5.8 7.6 6.0
Humid Aged Force loss, CFD 40% % < +/-30 45.9 30.6 30
acoustic test used by Renault and PSA Group, did not indi­
Dynamic fatigue, Height loss % <5 2.5 2.7 3.7 cate statistically significant changes when 15 pphp of PU
Dynamic fatigue, 40% IFD loss % <25 25.9 26.8 34.8 powder was used. Differences in foam airflow resistivity
Flammability mm/mn < 10 0 58 82 78 had a larger influence. (Figure 8)

However, the required characteristics being less stringent

in this respect for head-rest foams, it would appear from
these results that a PU powder level of 8 to 10% on the
final foams should be feasible for this application.

For MDI based NVH molded foams (Table 15), both the
foam mechanical properties, durability i.e Compression-set
and acoustics properties are acceptable at PU loading lev­
els of 15 pphp or 9% on the foam.

Figure 7. Petite Cabine, (5 kg/m2 mass layer) + NVH

Foam, 22mm thick.

26 / Berthevas, Aguirre, Tu
 REFERENCES
The results obtained in terms of both processability and
final seat foam performances have been summarized in the 1. H. Stone, S. Lichvar and F. Sweet, 1995. “ Commer­
attached Table 16 with the addition of costs as given by the cial Potential for Recycling of Finely Ground Foam in
respective suppliers, Mobius Technologies [11] and Regra Flexible Polyurethane Foam” .
Recycling (Feb 2001). Proceedings o f the Polyurethane 1995 Conference (SPI)
These are the costs to pulverize or solvolyse clean flexi­ . pp. 244­252.
ble PU foam. 2. D. Gibala, R.J Cain, and M.C Salsamendi 1995. “ Ef­
Any cost to recover the foam, wash it and transport it fect of Post-Consumer Automotive Seating Granulate on
must be added which is the same in both cases. In the case TDI-Based Flexible Foam Properties “ . Proceedings o f the
of car seat dismantling to recover the foam these costs can Polyurethane 1995 Conference ( SPI).
be substantial [29]. Lower costs are achieved by recovering pp-261-269.
foam from ELV shredding residues (ASR’s), but the cost of
cleaning for decontamination has to be added. 3. D. Sinopoli and P. Farkas, 1996. “ Ecostream™ Pro­
cess in Automotive Polyurethane Foams”. Proceedings of
Table 16. Pulverization versus Solvolysis to recycle car the Polyurethanes Expo *96
seats pp. 68-76
Pulverization Solvolvsis
4. C. Mitchell, T. Patel, S. Venkatasanthanam, and F.
Processability Acceptable Better
(in new molded foams) Shutov, 1996.
Max. Recycle content “ Pulverized Polyurethane Foam Waste as a Reinforcing
(olf foam in new foam) c.a 7% c.a 7% Agent for Virgin PUR Foams “. Proceedings o f the Poly­
Performances in car urethanes Expo *96. pp. 105-110.
seat foams Acceptable Acceptable
Capital investment
(1500MT/yr) $ 2 MM $ 0.7 MM 5. Bryan M artel, 1997 / ‘Reclamation of Urethane Foam
Processing cost from Automotive Seats” , Proceedings o f the Polyurethane
(2 shifts/day, 200 days/yr) 17c/lb 16c/lb (0.77DM/kg) World Congress *97, Amsterdam, The Netherland ,pp.588-
597
CONCLUSION
6. K.L Swisher, F.A. Shutov, 1998. “Mechanical
As a conclusion, one can say that the Mobuis pulveriza­ Properties of Cellular Polyurethanes Filled with
tion process is an economically attractive recycling alterna­ Polyurethane Disperse Filler”. MD -Vol 82, Porous,
tive to flock rebonding for slabstock production and trim Cellular and Microcellular Materials ASM E1998. pp-53-
scrap when the market price for such foam is low (c.a. < 30 60
c/Ib) as is the case today. 7. R. Villwock, B. M artel, P. Berthevas, 1999. “Recent
Recycled foam contents are typically 10 to 15% in the Technical Advances and Economic valuations in automo­
new foam. tive seat Recycling”. Annual Recycling Conference ,SPE
This same technology, but also solvolysis with dicarbox- Recycling Division, Dearborn ( MI), Nov ‘99.
ylic acids can be used to recycle flexible molded foams.
For car seats, to meet performance requirements, the re ­ 8. F. Shutov, 1999. “Recycling of Polyurethane Foam
cycled foam content should be limited to about 7%. For Waste Derived from Used Automotive Seats” . Proceed­
other less critical applications such as head-rests and NVH ings o f the 15th International Conference on Solid Waste
pads, levels as high as 10% may be used. Technology and Management**. Dec *99
Philadelphia, PA, USA. pp. 3D/1-3D/6
ACKNOWLEDGEMENTS
9. E. Weigand, W. Rasshofer, M. Hermann, G.F Bau­
The authors wish to thank their many Dow colleagues mann and M. Nakamura
who had a role in generating the data presented in this pa ­ 1993. “Recycling of Polyurethanes Put into Practice”.
per, in particular Hercule de Freitas, Philippe Osche, Alain Proceedings o f the
Bleton, Gaby Chassot, Jean-Claude Niederer, Jean Cour- Polyurethanes World Congress 199. pp- 211-217
tial, Daniel Cadolle, Chantal Roethlisberger, as well as
Mobius Technologies, Regra Recycling and Prof. Bauer. 10. 7. Wagner, K. Peltzer, J. Wirth, 1994. “ Process for
Powdering Polyurethane, Polycarbamide and/or Polyure­
thane-Polycarbamide Materials in a Rolling Mill “. EP
® SPECFLEX, VORALUX and VORANOL are Trade­ 0711 221 B1 Filing date 18 july 94.
marks of The Dow Chemical Company

Berthevas, Aguirre, Tu / 27
 11. Jeff Jensen , 2000, “ Real-world Economics of Poly­ 22. K.W Krosen, J.F.G Hopper, D. Thorpe, A. Kirk and
urethane Foam Recycling”, PFA Conference 2000. G. Parrinello, 1992
“Developments in the Recycling of Polyurethanes and
12. H. Stone , R. Villwock and B. M artel, 2000, Design of Foams for Easier Recycling “. Proceedings o f
“Recent Technical Advances in Recycling of Scrap Po- the SPl-34th Annual Polyurethane Technical/Marketing
luyurethane Foam as Finely Ground Powder in Flexible Conference, pp. 577-582.
Foam”.Proceedings of the Polyurethanes Conference
2000, pp 3-10 23. Prof.Dr G. Bauer, 1995, “ Process for the production
of isocyanate reactive polyol dispersions and their use”
13. M.P. Neuray, J. Wirth, H.M Sulzbach, 2000, Patent D E 195 12 778 C l ,Filing date April 5, 1995.
“Methods for the Recycling of Polyurethane and Polyure­
thane Composites “ Proceedings of the Polyurethanes Con­ 24. B. Naber , V. Neiss and M. Gassan, 1995, “ New
ference 2000,pp. 37-48 Polyols Made by Glycolysis from PUR and PIR Rigid
Foam Scrap and Their Applications”
14. R. Villwock, B. Martel, H. Stone, 2000. “Polymeric Proceedings of the Polyurethane 1995 Conference, pp.
Foam Powder Processing Techniques and Powder pro ­ 287-290.
duced Thereby” .
US patent Appl. # 60/172,081 dated 23 Dec. 2000 . 25. D.A. Hicks, C.B H em el, A.C Kirk, R.J Stapleton, &
A.R Thompson, 1995,
15. R. Villwock , B. Martel , P. Berthevas, 2001. “Re­ “Recycling and Recycled content for Polyurethane
cent Technical Advances and Economic Evaluations in Foam” Proceedings of the Polyurethane 1995 Conference,
Automotive seat Recycling” . Proceedings of the Interna­ pp. 279-286
tional Automobile Recycling Congress , March ’01 , Ge­
neva . 26. B.Naber, V.Neiss, M.Gassan ,1996. “ The Applica­
tion of Polyurethane-and Polyisocyanurate-Glycolysates”
16. A.C Kirk , 1992, “ PURRC Flexible Foam Task Proceedings of the Polyurethane Expo’96, pp. 89-93.
Group Contributions to the Recycling/recovery of Flexible 2000, pp.
Polyurethane Foam”. Proceedings of SPI-34th Annual
Polyurethane Technical/Marketing Conference, 1992, 21. K.K You, D.T Durocher , P.Ch. Kierkus an T.L
pp . 572-576. Fishback, 1997,
“Chemical Recycling of Polurethanes and Applications
17. G. J. Wasiczyk and T.J Cerabona, 1993 , “ Develop­ for the Recyclates”
ing a viable Polyurethane Waste Management Plan : A Proceedings of the Polyurethanes world Congress ‘97, pp
PURRC Update “. 598-602
Proceedings of the Polyurethanes World Congress 1993,
pp. 196-204. 28. G. Bauer , M. Kugler , R. S. Chakrabarti, 1999,
“Recycling of PUR Flexible Foams: - Solvolysis with
18. C. Bastian ,1993, “Polyurethane Recycling in Per­ Carboxylic Acids fills the Gap “.Kunstoffe 89 ( 1999) 2,
spective” pp. 85-88
Proceedings o f the Polyurethanes World Congress {93,
Vancouver, 29. F.E. Mark, A. Kamprath, 2000 “Recycling & Re­
pp 205- 210 covery Options for PU seating Material - ELV Treatment
Today and in Future. An LCA Type Cost Analysis” Society
19. J. Quinlan, 1994, “ Recycling Polyurethanes in the o fAutomotive Engineers Conference 2/2/2000.
Ws “
Proceedings from the SPI-35th annual Polyurethane
Technical/Marketing Conference, ’94. pp 537-543 .

20. C. Bastian and G. Strobbe, 1997, “ Polyurethanes

Recycling- a Joint Industry Initiative” . Proceedings o f the
Polyurethane World Congress ’97 pp. 582-587.

21. H. R. van der Wal, 1992, “New Chemical Recycling

Process for Polyurethanes”. Proceedings of the SPI -34th
Annual Polyurethane Technical/Marketing Conference, pp.
560-564.

28 / Berthevas, Aguirre, Tu
 BIOGRAPHIES

Paul Berthevas

Paul Berthevas graduated from

the “Ecole Nationale Superieure de
Chimie de Paris”, in 1968.
He worked for the past 28 years
on Polyurethane Chemistries and
Application Technologies and
contributed to the development of
Copolymer Polyols and their use in
flexible polyurethane foams and co­
authored a number of papers on the
subject.
He also developed chemistries and application technolo­
gies for the production of knife coatable, heat curable mi-
crocellular polyurethane frothed foams and holds patents
on this application technology.
He then became Global Technology Manager at Dow
Chemical with responsibility for molded polyurethane
foams and is today Senior Development Associate respon­
sible for the flexible foam recycling project. He is located
at Dow Europe’s IDC in Meyrin / Geneva, Switzerland.

Fabio Aguirre

Fabio Aguirre, chemist. Graduated

from Universidad Industrial de
Santander, Bucaramanga, Colombia, in
1989.
He Joined Dow Latin America in
polyurethanes TS&D Bogota in 1989
and was involved in development
projects for shoe-sole elastomers, rigid
foam and flexible foams for the area of
Latin America. Currently he is member of the flexible foam
group at Dow Europe’s polyurethanes IDC located in Mey­
rin, Switzerland.

Johnson Tu

Johnson is a Sr. Development

Specialist in the Research and
Development Group of The Dow
Chemical Company in Freeport,
Texas. He received a BS Science
from the Chengkun University in
Taiwan in 1983. He joined Dow in
1990 and has held several positions in
R&D/TS&D and New Business
Development.
 SESSION B
AUTOMOTIVE LOW V.O.C. EMISSIONS

TECHNICAL SESSION ORGANIZERS

Jim Tobias, Chair

AIR PRODUCTS & CHEMICALS, INC.

Jim Lambach
BAYER CORPORATION

Allan James
DOW CHEMICAL CANADA INC

TECHNICAL SESSION MODERATOR

Don Schomer
BAYER CORPORATION

TECHNICAL SESSION VICE - MODERATOR

Jim Tobias
AIR PRODUCTS & CHEMICALS, INC.
Non-VOC Catalysts for HR Molded Foams Used in Automotive
Seating Applications

MITSURU SAKAI, ATSUSHI ISHIKAWA, MASAYOSHI MORII and KAZUHIKO KIUCHI

Kao Corporation
Performance Chemicals Research Laboratories
1334, Minato Wakayama 640-8580
Japan

ABSTRACT even slightly. The catalysts are generally tertiary amines

and/or organometals. Especially tertiary amine catalysts
We reported on the fundamental characteristics of are essential for controlling the balance o f water-isocyanate
KAOLIZER No.25 (KL -No.25), which is a reactive amine and polyol-isocyanate reactions.
catalyst, in the previous paper[l]. In this paper we In Japan, one o f main uses of PU foam is in automotive
describe the results on KL-No.25 applied to practical seat cushions, which is produced by Hot or Cold molding
automotive foams. process at equal levels. It is predicted that Cold molded
These days, automotive seat cushions contents have been foam (HR foam) production will be more in the future.
changing from Hot molded foam to HR molded one from Although TEDA has been mainly used as a catalyst for
the viewpoint o f energy saving and cost reduction. At the HR molded foam; at present a reactive amine catalyst is
same time, volatile organic compounds (VOC) in the cars drawing attention from the viewpoint of VOC reduction,
have been an important factor in the problem o f safety and since the reactive amine catalyst with an active hydrogen
health. group, which reacts with - NCO in the molecule, is taken in
Triethylenediamine (TEDA), which is volatile, is PU foam not to vaporize in the air.
generally used as a catalyst for HR molded foam because it However, it is generally said that the use o f reactive
has strong catalytic activity and gelling action. We have amine catalyst for HR molded foams causes some problems
found that KL-No.25 also has strong gelling action and can as follows.
replace TEDA. We have compared the characteristics of 1. Physical properties:
KL-No.25 with those of TEDA used in HR molded foam Foam strength and wet-set performance are influenced
systems (back and front seats). Furthermore, we have because many typical reactive amine catalysts on the market
noted the relation between the gelling/blowing index and the have only one active hydrogen group resulting in the
catalytic activity o f typical reactive amine catalysts. termination of polymerization.
Japan has been introducing the replacement o f TEDA in 2. Moldability:
some HR molded foam makers. We expect KL-No.25 will Cell defect and a poor flowability are being observed.
be useful for HR molded foam in automotive seat cushions The reaction o f a reactive amine catalyst with -NCO slows
in the world. down in process of the reaction because of decreasing the
freedom of the catalyst.
INTRODUCTION In the previous paper we reported about some excellent
characteristics o f KL-No.25; for example low odor, anti­
These days, problems such as the ozone layer depletion, vinyl staining, anti-fogging and good heat stability o f foam.
health concerns and environmental impacts are increasingly In this report we will describe the results of practical test of
important issues. These are also important problems for KL-No.25 as a catalyst for HR molded foam. At the same
polyurethane (PU) industry to tackle, and an approach from time, we are going to report some valuable data that KL-
the side o f amine catalysts is required from catalyst No.25 is a useful amine catalyst for VOC reduction.
producers.
The amine catalysts are indispensable for production of
polyurethane foams and for changing the foam properties

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3 2001 , 33

EXPERIMENTAL cylindrical cup. Cream time, gel time and rise time were
recorded continuously after the mixing. Foam physical
Catalysts properties were evaluated by the methods o f the standard
ASTM or JIS.
The list o f conventional tertiary amine catalysts and
reactive amine catalysts evaluated in this paper are shown in Measurement of free amine in PU foam
Table 1.
Polyurethane foam was prepared using an aluminum
Hand mix Evaluations mold 200 x 200 x 70 mm in size. Just after demold, test
specimens (50 x 50 x 50mm) were cut from PU foam, and
Polyol formulations were blended with all o f polyols, put in room temperature and then in an oven at 70°C.
chain extenders, catalysts, surfactants and water as a Two specimens of each foam system were taken at specified
blowing agent for 3 minutes using a mixture for laboratory. time intervals, and were placed in a 500ml beaker, adding
Premixed formulations and Sumidule VT-80 50ml ion exchanged water. The specimen was compressed
(TDI/MDI= 80/20) were maintained in water bath controlled 20 times with a rod (surface area 25cm2), to elute the free
at 23°C, and either o f them were weighed together into a amine in PU foam in the water. 25ml was then taken using
polycup in the proportions o f the intended Index for each a hole pipet and was titrated with 0.0IN HC1.
foam as quickly as possible. Immediately, the mixture was
blended for 5 seconds with a Homo Disper, rotating at 6000 Corrosion test with polycarbonate (PC)
rpm. The mixed foam was poured into an aluminum mold
measuring 200 x 200 x 70 mm or 300 x 300 x 100 mm, and The test was carried out by the method in Figure 1. A
heated at 60 - 65°C. polyurethane foam sample o f lOg and a PC plate (40 x 35 x
3mm) were hung in a covered glass bottle containing lOOg
Machine Evaluations ion exchanged water, and placed in an oven at 80°C for 100
hours. After heating, the haze value of PC plate was
Machine evaluations were carried out using a high- measured.
pressure machine HK-270. Polyol formulations were
prepared according to the same method as the hand mix, but Sample for DTA
the scale was large. The foaming conditions are the
following. The reactive amine catalysts were made to react with
pure MDI in acetone, and then the acetone was evaporated
- Temperature of the raw materials: 23°C under reduced pressure at room temperature to get the
- Mold . 400 x 400 x 100mm aluminum sample.
- Mold temperature: 63°C
- Demold time: 6 minutes RESULTS AND DISCUSSION

Foam Evaluation Hand mix Evaluations

The processing properties of polyurethane foam were There are many kinds o f seat cushions in a car, for
measured by the free-rising foam produced using a example the driver’s seat, passenger seats and back seats.

Table. 1 C atalyst
Abbreviation Chemical composition Kao’s product name
Conventional amine catalysts
TMHDA N, N,N’,N’-Tetramethyl-1,6-hexanediamine KAOLIZER No.1
DMCHA N,N-Dimethylcyclohexylamine KAOLIZER No. 10
TMEDA N,N,N’,N’-Tetramethyl-1,3-ethylenediamine KAOLIZER No. 11
BDMAE 70%Bis[(2-dimethylamino)ethyl]ether 30%Dipropylene glycol KAOLIZER No. 12
MMP 4-Methylmorphorine KAOLIZER No.21
TEDA 33%Triethylenediamine 67%Dipropylene glycol KAOLIZER No.31
Reodtiyfijaminfi catalysts
KL-No.25 6-Dimethylamino-1-hexanol KAOLIZER No.25
DMEEE 2-{2-[2-(Dimethylamino)ethoxy]-ethoxy }ethanol KAOLIZER N0.23NP
DMEE 2-[2-(Dimethylamino)ethoxy]ethanol KAOLIZER No.26
MHEDA N,N,N’-Trimethyl-N'-hydroxyethyl-1,2-ethylenediamine
MHPDA N,N,N’-Trimethyl-N’-hydroxyethyl-1,3-propanediamine

34 / Sakai, Ishikawa, Morii, Kiuchi

 Table 2. Form ulation fo r HR foam
Formulation Front seat Back seat
System name F-TE F-25 B-TE B-25
Polyol (OHV=28mgKOH/g) 85 —
Polyol (OHV=35) 85
Polymer polyol (OHV=28) 15 <- «­ «"
Triethanolamine 2 — <- <­
Silicone type surfactant 1 <- 1.2 —
Water 3 «- 4.8 «-
BDMAE 0.1 «- «- <—
TEDA 0.25 0.6
KL-No.25 0.25 0.6
Figure 1. M ethod o f Corrosion test with PC Isocyanate (TDI/MDI=80/20) INDEX=105 INDEX'*105

This time we selected a typical front system whose molded bubbles, but TEDA little.
density was high and a typical back seat o f which the density Moldability o f KL-No.25, which is the most sought after
is relatively low, and evaluated some characteristics as a characteristic at this evaluation, was as good as TEDA
catalyst for HR foam. except for crashing. Especially the skin condition and the
We examined the both systems with hand mix first. cure o f slit’s burr o f molded foam were better in KL-No.25
Formulations and results are shown in Tables 2 and 3. Wet than in TEDA. Although KL -No.25 has slightly higher
set and foam strength o f KL-No25 and TEDA were almost crashing level than TEDA, it is machine-acceptable level.
the same in both front and back seat systems. In addition, Tensile strength of KL-No.25 was better than that of
we found that KL-No.25 has two different characteristics TEDA, but wet-set was reverse. Other properties showed
from TEDA: softness and good air permeability. Since no the same tendency as hand mix.
distinct defect was observed, we proceeded to machine As above, the system using KL-No.25 has shown
evaluation. approximately equivalent results as TEDA in reactivity,
moldability and physical properties, and we can say that
Machine Evaluations KL-No.25 is a candidate to replace TEDA, with a minor
modification o f polyol component if necessary.
We carried out machine evaluations about the back seat
system because a low-density system is more sensitive Measurement of VOC
performance. The result is shown in Table-4. The
reactivity of KL-No.25 was slightly fast in comparison with We indirectly measured VOC by two methods.
that o f TEDA, and KL -No.25 generated some healthy

Tabie 3. R esuit o f hand m ix

Systems F-TE F-25 B-TE B-25
Free foam Reactivity C.T (sec) 9 9 8 8
G.T 83 84 52 56
R.T 148 165 95 94
Core density (kg/m3) 35.5 34.2 23.6 23.7
Molded foam
Overall density (kg/m3) 59.7 60.1 34.8 34.9
Core density (kg/m3) 57.6 57.8 33.6 33.5
Hardness (F) 54 48 53 50
Physical Ball Rebound (%) 67 66 62 63
properties Tensile (kPa) 100 100 140 135
Elongation (%) 110 125 120 120
Wet set (%) 17.5 16.9 24.8 25.0
Air Permeability (cc/cm2/sec) 155 181 51 62

Sakai, Ishikawa, Morii, Kiuchi / 35

 Table 4. Result o f Machine evaluations
Systems B-TE B-25
INDEX too 105 110 100 105 110
Free foam Reactivity C.T (sec) 5 6 5 5 5 5
G.T 64 54 57 50 50 54
R.T 90 81 100 68 70 83
Max height (mm) 247 252 252 262 273 269
Core density (kg/m3) 23.7 24.3 24.2 23.0 22.6 22.4
Molded foam
Moldability Flowability 9 9 9 9 9 9
Skin n n n n n n
The four comer’s cell state 4/4 4/4 3/4 4/4 4/4 4/4
Cure of the slit’s burr 3/4 3/4 3/4 4/4 4/4 4/4
The level of crashing 3 3 3 3-4 3-4 3-4
Physical Overall density (kg/m3) 35.0 35.0 35.4 34.8 35.4 35.4
properties 25%ILD (N/314cm2) 75 87 95 59 69 81
Ball Rebound (%) 55 57 54 56 55 56
Tensile (kPa) 136 126 132 132 157 142
Elongation (%) 120 110 120 135 135 130
Wet set (%) 25.1 24.6 24.7 27.1 27.1 27.2
Air Permeability (cc/cm2/sec) 48 41 35 117 117 73
Flowability & Skin: g; good m; mid n; not so good
The Four Comers’ cell state & Cure of the slit’s burr: good parts / measuring parts
The level of crashing: 1; weak 5; strong

(1) Measurement of free amine in PU foam cannot be called a true non-VOC catalyst. We adopted a
corrosion test with PC, which is used as a material for the
One method is a measurement o f free amine catalyst instrument panel and the cover o f dome light in a car.
which remains in PU foam. As Figure 2 shows, free KL- Corrosion of PC is critical under the circumstances o f high
No.25 doesn’t remain in the foam after demold, indicating temperature and humidity as in Japan summer
that all KL-No.25 have reacted with -NCO. On the other Table 5 shows the result. Corrosion o f PC was observed
hand, TEDA remains in the foam to generate TEDA in conventional non-reactive tertiary amine catalysts. In
gradually at room temperature and rapidly at 70°C, reactive amine catalysts, some gave much corrosion
unexpectedly, and KL-No.25 and DMEEE didn’t at all.
(2) Corrosion test with polycarbonate (PC) This difference o f corrosion in the reactive amines is
thought to be due to heat decomposition.
I f a part o f reactive amine catalyst with - NCO pyrolyzes
under heating and releases free amine in an air, the catalyst
Table 5. R esult o f corrosion with PC
Catalysts Dosage (parts)
0.5 1.0
Reactive
KL-No.25 0.1 0.1
DMEEE 0.1 0.2
DMEE 78.2 71.3
MHEDA 92.8 81.3
MHPDA 88.9 67.4
Conventional
TEDA 3.4 87.8
TMHDA 81.0 81.3
DMCHA 58.7 58.4
Figure 2. R esult o f m easurement o f free amine in PU foam BDMAE 77.5 73.5

36 / Sakai, Ishikawa, Morii, Kiuchi

 The result in corrosion test with PC may be explained by reaction o f the hydrosis o f PC is especially promoted under
the pyrolysis behavior. We prepared standard samples by an alkali condition [2]. It is considered that the product of
reacting typical reactive amine catalysts and pure MDI, and KL-No.25 or DMEEE reacted with -NCO is stable to heat,
carried out the differential thermal analysis (DTA). but the corresponding products o f DMEE or MHEDA is
Figures 3 -6 show DTA chats. Notice that DMEE and not stable and caused the dissociation o f amine compounds.
MHEDA with bad corrosion have distinct one endothermic
and one exothermic peak, and KL-No.25 and DMEEE with Thus KL-No.25 reacts with -NCO completely and does
good corrosion have no distinct exothermic peaks. The not generate VOC at all in working environment and a car.

Figure 3. DTA chart o f DMEE Figure 4. DTA chart o f MHEDA

Figure 5. DTA chart o f KL -No.25 Figure 6. DTA chart o f DMEEE

Sakai, Ishikawa, Morii, Kiuchi / 37

 Table 6. C atalytic activity
Temperature 30°C 50°C
Gelling Blowing Gelling/Blowing Gelling Blowing Gelling/Blowing
KL-No.25 1.740 0.495 3.51 3.679 0.826 4.45
DMEEE 1.358 1.026 1.32 2.861 2.101 1.36
DMEE 1.115 1.267 0.88 2.367 1.913 1.24
MHEDA 2.043 2.136 0.96 3.961 3.355 1.18
MHPDA 2.008 0.759 2.65 4.565 1.363 3.35
TEDA 2.276 0.383 5.04 4.211 0.860 4.90
BDMAE 2.119 10.727 0.27 2.890 9.927 0.29
Non-Catalyst 0.155 0.053 2.93 0.290 0.054 5.38

Catalytic activity of reactive amine catalysts 2. S. Homma; Polycarbonate Resins Handbook o f THE
NIKKAN KOGYO SfflMBLN,LTD., 1992, p.338-361.
Finally, we explain about the position o f catalytic activity
o f KL-No.25 in reactive amine catalysts, which is of BIOGRAPHIES
importance for selecting the best catalyst and adjusting the
catalyst’s dosage. The catalytic activity was estimated Mitsuru Sakai
using the method o f Fakas. The solvent used was a Mitsuru Sakai joined Kao Corporation
mixture o f toluene and dioxane at a ratio o f 85:15, and the in 1973 from Sasebo Technical College.
isocyanate used was a mixture o f TDI and MDI with a ratio He is currently developing
o f 80:20. The result is shown in Table 6. It is noted that polyurethane foams and amine
KL -No.25 has the strongest gelling action in typical reactive catalysts.
amine catalysts. That is to say, KL-No.25 is the most
suitable catalyst for substituting TEDA from the viewpoint
o f catalyst activity.

CONCLUSION

We have studied the practicability of KL-No.25 as a non- Atsushi Ishikawa

VOC catalyst for HR molded foam and come to the Atsushi Ishikawa joined Kao
following conclusions. Corporation in 1984 after receiving
-Moldability. KL-No.25 is as good as TEDA. B.S. and M S. degrees in Applied
- Physical properties: Foam strength and wet-set o f KL- Chemistry from Kanazawa University
No.25 is almost the same as that of TEDA. In addition, in 1982 and 1984 respectively. He is
KL -No.25 gives soft foam and good air permeability. currently developing polyurethane
-VOC: KL-No25 reacts with -NCO completely and foams and amine catalysts.
doesn’t generate VOC at all in working environment and
a car.
-Catalytic activity: KL-No.25 is the strongest gelling
action among typical reactive amine catalysts.
We expect that KL-No.25 will be widely adopted as a non- Masayoshi Morii
VOC catalyst for HR molded foam.
Masayoshi Morii joined Kao Corporation in 1962.
REFERENCES After development works in the automotive foam
application, He is currently developing
1. M. Sakai; A. Ishikawa; M. Morii; K. Kiuchi; Proceedings polyurethane foams and amine catalysts.
o f the api Expo’99 International Technical Conference,
1999, p.259-266.

38 / Sakai, Ishikawa, Morii, Kiuchi

Kazuhiko Kiuchi
Kazuhiko Kiuchi joined Kao
Corporation in 1982 after receiving B.S.
and M.S. degrees in Applied Chemistry
from Gunma University in 1980 and
1982 respectively. He is currently the
group leader in the Polyurethane
Research Division.

Sakai, Ishikawa, Morii, Kiuchi / 39

New Silicone Surfactants for TDI Based HR Molded
Foams— Optimization of Bulk Stabilization and
Processing Latitude

T. BOINOWITZ, G. BURKHART and W. BUNTING

J. KLIETSCH Goldschmidt Chemical Corp.
Goldschmidt AG 914 East Randolph Road
Goldschmidtstrafie 100 Hopewell, Virginia 23860-2458
45127 Essen
Germany

ABSTRACT INTRODUCTION

Silicone surfactants are essential additives for the Transportation of people and goods has become more
production of high resilience (HR-) polyurethane molded and more important due to increased globalization. Diverse
foam. The structural design of a silicone surfactant should polyurethane components used to make interior parts of
match the different requirements of the various automobiles, contribute to higher comfort, safety and
formulations in order to optimize the balance between design. This paper will focus on automotive seat cushions
both chemical stabilization and cell openness. This in turn based on TDI technology. In comparison to MDI systems,
helps to insure production safety and consistency. usage of TDI allows the production of lower density foam
cushions due its higher NCO number. To achieve desired
This paper focuses on straight TDI 80 technology, being hardness, polymer-filled polyols are used. The lower foam
predominately used in North America. The TDI density, higher filler content, and lower functionality of
technology is used to provide high quality foams at lower TDI foams are the main reasons why they require more
densities. From the surfactant manufacturer’s point of physical stability from the surfactant.
view these foams, based on formulations including TDI in
combination with polymer polyols, are so-called non­ This paper describes the development strategy for a
stable systems. Besides cell regulation, these systems surfactant which fulfills all requirements of foam
require additional physical stabilization provided by the producers. In detail these are: a) high bulk stability, b)
silicone surfactant. Simultaneously, the silicone surfactant good flowability and c) sufficient crushability. Beside these
must have a broad processing latitude leading to foams processing demands, mechanical properties of the foam,
which are crushable with low force to open up the cells. which are indirectly influenced from the surfactant by cell
In addition, the automotive industry has increased its structure and tightness, should fulfill market demands. Very
interest in low emanation (FOG/VOC) foams. recently automotive producers added emanation limits to
Goldschmidt analyzed the composition of emanations their list of specifications, mainly in Europe. Consequently
from typical foam grades. Consequently silicone Goldschmidt included minimization of volatiles in
surfactants were optimized with respect to the lowest surfactants in the development criteria.
contribution to volatiles.

In this paper three new experimental products will be

introduced, which were specially designed for North
American type TDI formulations. They are the outcome of
a large screening program making systematic variations of
key structural parameters. The advantages of these
products, concerning processing and bulk stability, will be FIGURE 1. General chemical structure of HR surfactants
demonstrated. Their impact on cell structure, skin
appearance, and mechanical properties will be shown. For this screening program the parameters a) siloxane
chain length (n and m, see FIGURE 1) b) number of

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 41

organic modifications (m) and c) length and d) polarity of Formulations
pendant organic group (R) were varied systematically.
The properties of tested surfactants were evaluated in
EXPERIMENTAL different formulations (TABLE 1) stressing bulk stability,
crushability or flowability.
Foam Preparation
Formulation 1 2 3 4 5
The foams were produced either using a handmix Arcol 1342 "
clear polyol 90
laboratory procedure or a high pressure 7 component
10 WF75 "
machine with an output of 10 kg/min. Generally the 40% SAN 10
components were used at a temperature of 21°C. In the E 848 "
handmix procedure a propeller-blade strirrer with a clear polyol 60 45 40
diameter of 6 cm and 4 blades was used. The components E 852
43% SAN " 40 55
without the isocyanate were mixed for 1 min. at 1000 rpm NC 63 0 21
in a 2-liter polyethylene beaker. The isocyanate was added clear polyol 45
and subsequently mixed for 7 seconds at 2500 rpm. The NC 700 2)
40% SAN 55 60
mixed blend was poured into an aluminum mold of
40x40x10 cm ( - 500 g foam). This square mold, called a
Water 4,2 4,0 3,7 4,0 4,2
DEOA, pure 1,2 1.5 1,3 1,5 1,0
test block mold, was used mainly in the horizontal TEGOAMIN® 0,1 0,1 0,1 0,15 0,15
position. To stress flowability of the foams the mold was b d e 3)
tilted at a 20° angle from horizontal. For some TEGOAMIN® 0,3 0,3 0,3 0,25 0,25
experiments a commercial automotive front seat mold TA-33 3)
( - 700 g foam) was also used. The mold temperature was 49,4 49,2 46,6 51,2 51,9
TDI 80
<101> <105> <108> <110> <110>
65°C in all cases. Demold time was 4.5 min for all SAN 4% 17% 22% 24% 24%
formulations. Both molds were sprayed with GORAPUR® n Product from Bayer
LK 888 release agent shortly before pouring. 2)Product from DOW Chemical
3)Product from Goldschmidt
Physical Foam Property Testing
TABLE 1. Used TDI formulations
Physical property testing of the polyurethane foams was
done according to standard procedures described in the
literature. Force to crush values (FTC) were measured by RESULTS AND DISCUSSION
ten indentation cycles exactly 45s after demolding. After
the 10th FTC cycle foams were mechanically crushed and 43 new laboratory products were synthesized for this
an 11th cycle was recorded which represents a completely project. The parameters siloxane backbone length, number,
crushed foam. For CLD, sag-factor and hysteresis the test polarity, and molecular weight of organic modification
procedures described in DIN 53577 were used. were varied over a broad range.
In the following paragraphs the strategy and sequence of
Determination of Emanations the screening procedure will be discussed. In the tables,
values of the best three candidates will be summarized,
The emanations were analyzed according to Daimler- which are also named as experimental products:
Chrysler test method “PB VWT 709”. This method EP-K-52, EP-K-53 and EP-K-54.
describes a coupling of a thermodesorber (TDS 2 from
company Gerstel, Mttlheim/Germany) with a gas Processing Latitude
chromatography/mass spectrometry system (GC/MSD,
Hewlett Packard HP 6890/HP 5973). The VOC Broad processing latitude from the surfactant producer’s
desorption conditions were 30 min at 90°C. The column perspective means a low influence of stab. cone. var. on
for gas chromatography was an Optima 5 MS, 60 m, foam tightness. This guarantees a safe process considering
dF: 0,5 /zm. differences in raw materials, as well as deviations of tools
This method identifies all chemical substances which and process parameters within their tolerances.
are volatile at a temperature of 90°C. To quantify the Formulation number 1 is a critical blend which tends to
emanations, all peaks were related to a toluene standard. high FTC values due to overstabilization, because of low
Figures were given as ppm [mg emanations/ kg foam] in filler content. Consequently it is a valuable formulation to
toluene equivalents. evaluate the impact of surfactant concentration variation on
The FOG -data were obtained with a sampling FTC. As the lowest surfactant concentration, we chose the
temperature of 120°C for 60 min. The concentrations were amount of stabilizer which lead to foams without voids.
determined as hexadecene equivalents.

42 / Boinowitz, Burkhart, Klietsch, Bunting

The latitude was evaluated by adding 50% and 100% to
the original level. Product Level Flow defects* Flow defects*
[PPhp] _ Formulation. 1 Formulation 2
EP-K-52 0,5 2
Product Level FTC, FTC, FTC, Cells/
0,7 2 1
1st. 10“’ 11th cm
1,0 1,5
cycle cycle cycle
1,3 1,5
EP-K-52 100% 1544 123 113 11
EP-K-53 2
150% 1703 143 115 11,5
0,7 2 1
200% 1830 174 115 11,5
1,0 1,5
EP-K-53 100% 915 121 110 11
1,3 1,5
150% 1409 126 108 11,5
11,5 1,5
200% 1564 150 104
EP-K-54 0,7 1,5 1
EP-K-54 100% 1295 145 119 11
1,0 1,5
150% 1615 161 115 11,5
1,3 1,5
200% 1625 188 110 12
Comp. 1 1,5
Comp. 1 100% 1190 135 114 11
0,7 2 1
150% 1049 155 117 11,5
1,0 2
200% 1708 182 108 12
11 1,3 1,5
Comp. 2 100% 961 130 105
145 101 11,5 Comp. 2 0,5 3,5
150% 1300
1374 169 103 11,5 0,7 3 1
200%
1,0 1,5
1,3 1,5
TABLE 2. FTC values (N) using formulation 1 at a) * 1 = good, 2 = slight, 3 = medium, 4 = worse
minimal surfactant level, b) + 50% and c) + 100%
TABLE 3. Visual inspection of flow defects (subsurface
voids) using different surfactants in formulation 1 and 2
The surface and subsurface appearance were good for
all foams. The major criteria was the ratio of FTC, 10th Of the ten best candidates from the processing latitude
cycle at 200% to FTC at 100% surfactant level, which is screening only 3 also provide sufficient stability to support
flow of foam. Unsuitable surfactant structures resulted in
1.41 for EP-K-52 voids demonstrated in FIGURE 2.
1.24 for EP-K-53 a) 4 = worse b) 3 = medium
1.29 for EP-K-54
1.35 for Comp. 1
1.30 for Comp. 2

Of all surfactant candidates tested, 3 could not stabilize

the foam and 15 substances exhibited a poor processing
latitude, i.e. ratios above 2.0

Flowability

During expansion of the foam in the mold the surfactant

has to provide physical stability to the system to overcome
shear forces especially in the foam areas close to the mold
surface. The mechanical stress is constantly growing with
the degree of conversion (time or molecular weight),
because of increasing viscosity. Therefore the contribution
of the surfactant to flowability can be studied best in
complicated (e.g. sharp borders), steep molds. Also a
higher mould temperature increases chemical reaction rate
and consequently the shear forces. If the surfactant does
not provide enough stability, subsurface voids will be
visible even without cutting the foam.

The candidates were tested in formulations 1 and 2 in FIGURE 2. Subsurface defects in molded front seat
the commercial automotive mold. Foam cushions were cushions using formulation 1. Different surfactants at 0,7
visually inspected on a light table (TABLE 3). pphp.

Boinowitz, Burkhart, Klietsch, Bunting / 43

Bulk Stability out. The results are summarized in TABLE 5. Every foam
was made twice to get statistically relevant data.
As mentioned above, TDI based formulations are
generally not self stable, i.e. the surfactant has to provide IF D C LD
Level [%] 25% 40% 65% Height 25% 40% 65% Sag
additional physical stability. With unsuitable surfactants, [mm]__ factor
collapse in the whole center of the molded part can occur Comp. 1
100 90 134 270 98.3 1.6 2 4.35 2.7
or at least severe defects close to the surface. Sometimes 100 95 140 291 97.8 1.65 2.1 4.4 2.7
they are so severe that skin peels off upon demolding. 150 94 140 286 97.8 1.65 2.05 4.35 2.6
150 101 143 293 97 1.8 2.3 4.65 2.6
Bulk stability can be tested best in formulations with 200 100 142 287 96 1.75 2.2 4.75 2.7
high filler content. Formulations 3 and 4 could not 200 92 134 268 94.8 1.67 2.1 4.5 2.7
differentiate between different surfactants. Crushability 95 139 283 96.9 1.69 2.13 4.50 2.67
EP-K-52
and cell structure as well as skin appearance were similar. 100 96 141 286 98 1.65 2.1 4.35 2.6
Lowering the DEOA level in formulation 5 to 1,0 pphp 100 104 150 304 98 1.8 2.2 4.65 2.6
150 93 138 281 95.8 1.65 2.05 4.45 2.7
made the system less stable. Venthole collapses could be 150 97 141 286 96.8 1.75 2.2 4.65 2.7
observed at surfactant levels of 0,5 pphp. FIGURE 3 200 97 142 285 95 1.75 2.2 4.65 2.7
200 97 141 282 96.5 1.75 2.15 4.5 2.6
shows this phenomenon with 2 (top) and 4 (bottom) 97 142 287 96.7 2.15 4.54
1.73 2.65
rating. EP-K-53
100 105 152 317 99.8 1.85 2.25 4.75 2.6
100 104 150 309 97 1.9 2.35 4.8 2.5
150 105 153 308 96.5 1.95 2.35 4.85 2.5
150 102 148 296 97 1.8 2.25 4.7 2.6
200 100 144 290 97.3 1.85 2.3 4.75 2.6
200 94 141 278 93.3 1.7 2.2 4.5 2.6
102 148 300 96.8 1.84 2.28 4.73 2.56
EP-K-54
100 104 152 307 98.5 1.9 2.35 4.85 2.6
100 102 147 297 98.5 1.8 2.25 4.75 2.6
150 100 148 294 99 1.85 2.3 4.75 2.6
150 102 150 300 99.5 1.85 2.3 4.8 2.6
200 98 146 294 96.5 1.85 2.3 4.85 2.6
200 102 148 296 96.8 1.8 2.2 4.65 2.6
101 148 298 98.1 1.8 2.3 4.8 2.6

TABLE 5. CLD and IFD of foams made using three

different surfactant levels. In bold: Average values

FIGURE 3. Typical collapse induced from flow through In conclusion, foams made with EP-K-53 and EP-K-54 are
central venthole. Rating: 2 (top) and 4 (bottom) significantly harder, then pads made with competitive
product 1 or EP-K-52, although no visual differences in
Surfactant Level Formulation 5 cell structure can be observed.
[pphp] Venthole stability
(1 to 4*) Em anation Referring to Daim ler - Chrysler
EP-K-52 0,5 2,5
EP-K-53 0,5 2,5 Recently some of the European automotive producers
EP-K-54 0,5 2 have included emanation limits to their specifications for
Comp. 1 0,5 2 foam cushions.
Comp. 2 0,5 3
* 1 = good, 2 = slight, 3 = medium, 4 = worse

TABLE 4. Visual inspection of venthole stability using

different surfactants in formulation 5

M echanical Properties of the foams

The mechanical properties are mainly determined by the

functionality and molecular weight of raw materials, foam
density, filler content, isocyanate index as well as catalyst
package. The surfactant has an indirect impact on
mechanical properties of the foams via cell size and
regularity, and air flow.
However, hardness measurements from foam pads made FIGURE 4. Test setup for heating of automotive interior
with three different surfactant concentrations were carried and measurement of volatile compounds via GC/MS.

44 / Boinowitz, Burkhart, Klietsch, Bunting

Specifically Daimler-Chrysler and VW are interested not with the standard amount and with double the amount of
only in the amount of volatiles released from the surfactant. These surfactants allow a manufacturer to
polyurethane foam, but also in the chemical composition overcome deviations of cushion quality caused by day to
of the emanation spectra. day variations of raw materials and foaming process
parameters. In addition, the 1st cycle FTC should be low
Three years ago Goldschmidt implemented the Daimler- to allow for easy demolding.
Chrysler test method to determine VOC and FOG of
foams. TABLE 6 summarizes the total VOC values and Flowability was tested in a commercial automotive seat
the analysis of major components. Foams were made with mold. If the surfactant does not provide sufficient stability,
formulation 1 and a surfactant level of 1 pphp. subsurface voids will become visible, especially near the
“A” surface of the part (bottom surface of the mold).
Substance EP-K-52 EP-K-53 EP-K-54 Comp. 1
TA-33 203 193 201 212 From the surfactant supplier perspective, TDI based
TA -BDE 153 140 144 152 systems can be classified as not self-stable. They need
Dinitril 27 26 25 27 additional physical stability, which has to be provided by
BHT 7 6 6 7 the surfactant. This was evaluated in formulations with high
Siloxanes 37,4 17,2 22,7 45,3 filler (SAN) content. Venthole stability was recognized as a
Total 511 481 532 501 valuable parameter to differentiate between candidates
tested.
TABLE 6. VOC values in ppm according to Daimler-
Chrysler test method “PB VWT 709”. In addition to processing related criteria mentioned
above, the surfactant should have a low contribution to
The results indicate that today the amine catalyst package emanations. We evaluated and minimized the impact of
contributes the major share to VOC emanations. Recently surfactant contributions to VOC according to Daimler-
amine suppliers developed non fugitive catalyst packages, Chrysler.
which are optimized concerning emissions.
In conclusion with EP-K-52, and EP-K-53 and EP-K54,
The dinitril component is the initiator for SAN three surfactants were selected which are superior to
polymerization. competitive products as well as Goldschmidt products
concerning the set of all four criteria.
However, surfactant suppliers have also made
significant efforts to develop superior surfactants with a
lower contribution to emanations. Goldschmidt now offers
a large variety of surfactants of the LF (= low fogging)
series. The usage of EP-K-52, -53 and -54 will allow foam
producers to fulfil the requirements of today’s automotive
industry.

SUMMARY

This paper describes the development of surfactants for

TDI-based automotive seating systems. 43 new
experimental structures were evaluated in 5 different
formulations, stressing the different requirements of the
surfactants.

In particular these requirements are:

a) Broad processing latitude,
b) Sufficient support of flowability
c) High bulk stability and
d) Low contribution to emanations.

Processing latitude was tested by measurement of force to

crush (FTC) values of foams made with different
surfactant levels. The indicator for broad processing
latitude was a low value for the ratio of FTC (10th cycle)

Boinowitz, Burkhart, Klietsch, Bunting / 45

BIOGRAPHIES

Tammo Boinowitz Georg Burkhart

Dr. Tammo Boinowitz, born in Dr. Georg Burkhart joined

1966, studied Chemistry at the Goldschmidt AG in 1981 after
University of Hamburg. He receiving his Ph.D. in chemistry
received his Ph.D. in Physical at the University of Bochum,
Chemistry from the University working mainly at the Max-
of Duisburg. He joined Planck-Institut of Coal research.
Goldschmidt AG in 1995, He has been involved in both
being involved first in R&D of polyurethanes research and
Polyethers. Today he is technical service and since 1986
responsible for development he is the leader of
and technical service of Goldschmidt’s polyurethane
additives for flexible slabstock laboratories.
and molded foams

William Bunting Juergen Klietsch

William Bunting received a BS Juergen Klietsch joined

and an MS in Chemistry from Goldschmidt AG in 1969. For
Rensselaer Polytechnic Institute more than 20 years he is
in 1977 and 1980, respectively. involved in product
He joined the molded HR foam development and technical
applications group of Union service for high resilience foam.
Carbide Corp. in So. He co-authored several patents
Charleston, WV in 1984. After and publications.
spending three years at Union
Carbide and ten years at
Carpenter Co. in Richmond,
VA he joined Goldschmidt
Chemical Corporation in 1997. He is now a Sr. Technical
Service Representative responsible for technical support
of all automotive applications, in addition to microcellular
and other elastomer-related technologies.

46 / Boinowitz, Burkhart, Klietsch, Bunting

Elimination of Amine Emissions from Polyurethane Foams:
Challenges and Opportunities

F. M. CASATI, J. M. SONNEY, H. MISPREUVE R. HERRINGTON and J. TU

and A. FANGET The Dow Chemical Company
Dow Europe S.A. 2301 North Brazosport Blvd.
Intematioanl Development Center Freeport, Texas 77541
13, rue de Veyrot—P.O. 3
CH-1217 Meyrin 2 (GE)
Switzerland
polyol that help reduce and in some cases eliminate cata­
lysts while improving process. Comparative data will be
presented to illustrate the utility of these new polyols in
ABSTRACT
helping the polyurethane industry continue to reduce vola­
tile emissions.
Polyurethanes are a unique class of plastic in the sense that
independent manufacturers buy liquid polyols and isocy­
INTRODUCTION
anates and then react them in the presence of catalysts, sur­
factants and other additives in widely varying processing
Flexible polyurethane foams are well-established industrial
schemes to make the final commercial product. In some
products with important technical and economic advan­
manufacturing scenarios, there is a growing concern about
tages. These materials have found their way into our daily
worker exposure to volatile organic chemicals (VOC) such
lives through their use in the endless array of consumer
as the amine catalysts. Emissions from manufactured arti­
products. To complement this wide-ranging success, sig­
cles that contain polyurethane components are also a con­
nificant development efforts continue today to improve
cern.
targeted performance features like comfort, durability, re­
sistance to moisture and temperature [1-3]. Improvements
The polyurethane industry is continuing its legacy of suc­
in the environmental impact of these products are also be­
cessfully reducing the amount of volatile organic chemicals
ing gained through reductions in the amount of VOC’s re­
released from its operations. Examples of this include the
leased during their manufacture. In some manufacturing
total elimination of the classical chlorofluorocarbon blow ­
scenarios, there is a concern about worker exposure to the
ing agent from flexible foams and the replacement of vola­
vapors of traditional amine catalysts. Amine emissions
tile antioxidants with less migratory ones. Today, many of
from manufactured articles containing polyurethane com­
the classes of ingredients commonly used in preparing
ponents is a growing debate due to potential fogging, PVC
polyurethanes are available in low-fogging or low VOC
(PolyVinyl Chloride) staining and PC (Poly Carbonate)
grades. The next big challenge being addressed by the in­
degradation [4-6].
dustry is the elimination of amine catalyst vapors. This is
important for both VOC and odor concerns.
The industrial hygiene and safety issues related to handling
and use of amine catalysts in polyurethane production are
The catalyst suppliers have responded by introducing a
complex. It begins with the need to purchase one or more
variety of new, non-fugitive (sometimes called reactive)
of these additives in separate containers that must be off­
amine catalyst compositions. Their success at reducing
loaded, warehoused, transported to the formulation room
amine emissions while providing catalytic activity is widely
and in many cases physically handled by workers assigned
reported in the literature. Difficulties associated with reac ­
the task of blending up a formulation. Most amine catalysts
tivity profile adjustment, higher usage levels and losses in
have exposure and odor properties such that personal pro ­
certain foam physical properties suggest that an improved
tective equipment is required [7]. Empty working and
level of technology is still needed. This paper will present a
review of work The Dow Chemical Company is doing to shipment vessels need to be properly handled and prepared
for disposal or shipment back to the vendor.
develop flexible foam grades of polyether

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 47

Due to the exothermic nature of the reactions and the extra Table 1. Challenges to catalyze polyurethane reactions
heat applied in some curing scenarios, freshly prepared Solubility parameters (including water) ~~
polyurethane products are always at an elevated tempera­ Steric hindrance__________________________
ture. With the classical non-reactive type amines, odorous Difference of NCO reactivity between isomers
vapors are released and in the absence of proper ventila­ Acidity__________________________________
tion, worker exposure can occur [8-10]. When captured by Viscosity build-up_________________________
an exhaust system, these amine vapors can represent a sig­ Effect of temperature on catalyst efficiency
nificant contribution to the total VOC’s being released by
the manufacturing plant. The first requirement for good catalytic activity is that the
catalyst molecule be highly compatible with the mixture of
Even after cooling down and undergoing fabrication into reaction raw materials. Molecular structure of the catalyst
finished products, trace amine odors are sometimes still and the resulting steric hindrance will impact how easily
detectable. Amine catalyst vapors have been identified as the molecule can approach the reactants and facilitate the
one component of the new car smell [11]. desired reactions [20]. The catalyst package needs to be
effective when isocyanates of varying isomer content and
The catalyst suppliers have responded by introducing a reactivity are used. Trace acidity, common in some raw
variety of new, non-fugitive (sometimes called reactive) materials, impact the performance of the catalyst(s). Mo ­
amine catalyst compositions. Examples of their success at lecular mobility of the catalyst should be high so that it can
reducing emissions while providing catalytic activity are continue to exert its effect as the system reacts and builds
recorded in the literature [12-18]. Difficulties associated viscosity. Since the reactions are exothermic, the catalytic
with reactivity profile adjustment, higher usage levels and effect should not be overly influenced by a rise in tempera­
losses in certain foam physical properties suggest that an ture.
improved level of technology is needed.
All of these requirements are more easily met with the con­
In this paper, we continue to report the previously an­ ventional, fugitive type amines rather than the newer reac­
nounced activities at Dow to develop active polyether tive type amines. To address these challenges, the reactive
polyols [19]. Comparative data will be presented to illus­ amines are normally used at significant higher concentra­
trate the utility of these new polyols in helping the polyure­ tions, hence sometimes are not fully bound in the polymer
thane industry continue to reduce volatile emissions. network [21]. In addition, most of the reactive amines are
mono-functional and can be considered as polymer chain
BACKGROUND terminators (which generally detracts from the ultimate
properties of the polymer). Dow active polyols are useful to
Most commercially manufactured polyurethane foams are overcome these challenges.
made with the aid of at least one catalyst. Many classes of
compounds have been investigated and the amines and the EXPERIM ENTAL
organo-metallics have been found most useful. Various
A list of all raw materials used with this study is given in
combinations of catalysts are used in order to establish a
Tables 2, 3 and 4 (SPECFLEX *, VORANOL *,
balance between the chain propagation or gelling (isocy­
VORANATE *, ISONATE * are trademarks of The Dow
anate with polyol) reaction and the blowing reaction (iso­
Chemical Company).
cyanate with water). Any given catalyst molecule will
normally have some activity for both reactions and it is Table 2. List of raw materials
common to find synergistic performance when using two or SPECFLEX NC 630 Polyether polyol
more catalysts in a formulation. The polymer formation SPECFLEX NC 632 u

rate and the gas evolution rate must be balanced so that the VORANOL CP 6001 u

gas is entrapped efficiently in the gelling polymer while the VORANOL CP 1421 u
u
foam cells develop sufficient strength to maintain their VORANOL 3137 A
structure without collapse or shrinkage. Catalysts are also SPECFLEX NC 700 SAN copolymer polyol
important for assuring completeness of reaction or “cure” SPECFLEX NE 150 MDI isocyanate
in the finished foam. Some of the challenges to success in VORANATE T-80 TDI isocyanate
designing a catalyst package for a polyurethane production
reaction are listed in Table 1.

48 / Casati, Sonney, Mispreuve, Fanget, Herrington, Tu

 Table 3. List of catalysts shapes were used to evaluate the processing characteristics
Trademark Supplier of both control and experimental foam formulations.
DABCO NE-200 Air Products
DABCO NE-1060 “
Foam Testing
u
DABCO 33 LV
DABCO BLV “

DABCO VP 137 u Foam physical properties were determined in accordance

DABCO T i “ with ASTM, ISO, NF and major OEM’s (Original Equip­
TOYOCAT RX-20 Tosoh ment Manufacturers) testing procedures.
TOYOCAT RX-21 “

PC-CAT HPI Nitroil

PVC Staining Tests
DMAMP-80 Angus
NIAXA-1 Crompton
JEFFCAT ZF-10 i Huntsman Accelerated aging tests at elevated temperature were car­
ried out in closed containers in the presence of a PVC foil.
(DABCO* is a trademark of Air Products & Chemicals Inc; A foam sample of 50x 50 x 25 mm was cut from the core of
TOYOCAT * is a trademark of Tosoh Corporation; NIAX * is the foam part which had been isolated just after production
a trademark of Crompton Corporation; JEFFCAT * is a and was placed at the bottom of a 1 liter glass jar. A piece
trademark of The Hunstman Corporation; PC-Cat * is a
of gray PVC skin, reference E 025 373A0175A, supplied
trademark of Nitroil Europe Handels GMBH; DMAMP-80* is
a trademark of Angus Chemical Company) by Bennecke-Kaliko (Germany), was hung by the rim of
the jar, which was then sealed. Aging was carried out at
Table 4. List of Dow Experimental Active Polyols 115°C for 72 hours. After cooling, the PVC sheet discol­
OH number Functionality EO capping oration was measured using a Minolta Chroma Meter CR
Polyol A 56 3 no 210. The smaller the change in color, the lower the A E
Additive A 56 <2 yes measured in this test.
Polyol B 32 >3 yes
Polyol C 32 >3 yes
RESULTS AND DISCUSSION
Foam Preparation Dow newly developed active polyols were tested in various
conditions with different technologies. The purpose of this
Foam formulations are indicated with each section of the work was to assess their performance and to define how
paper. Bench scale molded foams were produced in a they can be formulated to reduce and in some cases elimi­
30x30x10 cm test block mold heated at 60°C using the nate amine catalyst use. Initial work had been concentrated
standard hand-mix procedure [22]. Free-rise foams were
on the performance of reactive amines offered by the vari­
poured in a card-board box. ous suppliers and on their combination with Dow’s active
polyols. In most cases it was found that these products gave
Slabstock foams were either produced on a POLYMECH
good compatibility with Dow active polyols and that, by
or on a VARIMAX machines. For MDI HR molding high
allowing a reduction of their concentration in polyurethane
pressure impingement mixing machines (KRAUSS-
formulations, overall processing was improved.
MAFFEI and CANNON) were used to pour foams in a
40x40x10 cm standard test block mold and in a variety of Main results of this study are presented hereafter with vari­
production seat molds. ous technical options. These are concentrated on flexible
foams since these are the focus of this research program.
Foams for the processability studies in low-density TDI- Each foam technology will be reviewed and discussed
based HR molded foams were prepared in a Hi-Tech Engi­ separately.
neering Model RCM -40 high pressure, impingement mix­
ing foam machine. The machine was retrofitted with an SLABSTOCK FOAM
Admiral brand, L-type (10/14) mixing head. Two process
streams (isocyanate and polyol masterbatch) were used. Dow’s experience with slabstock foam has shown that most
Nominal throughput of the machine was 10 kg/min (22 common density range can be produced without the addi­
pounds/minute) at the component ratios necessary to main­ tion of traditional amine catalyst. Experimental Polyol A
tain a 105 isocyanate index for all foams made during this can be used alone to provide a smooth rise profile for low
study. density (16kg/m3) formulations using all-water blowing.
Rise rate of such formulation is identical to that of tradi­
Molded Foam Processing Tests tional amine catalyst blown formulations (Figure 1). Addi-

The protocol for testing foam processability factors have

been described in detail [22] and will be further discussed
below. A variety of heated aluminum molds of various

Casati, Sonney, Mispreuve, Fanget, Herrington, Tu / 49

tive A was used, together with Polyol A in formulations, Table 5. Slabstock foam: Formulation and foam
especially those with low water level, where more blowing physical properties______________________________
is necessary. Figure 2 illustrates rise rate comparison of 3.7 Foam formulations A B C D
PHP (parts per hundred parts of polyol) water formulations TDI INDEX 115 115 110 110
with the new active Polyol A. With tuning of additive A, a VORANOL 3137A 100 100
rise rate similar to that of traditional amine catalyst can be Exp.POLYOL A 100 99.3
reached. Exp ADDITIVE A 0.7
WATER 6.5 6.5 3.7 3.7
DABCO DC-5160 1.2 1.2 1 1
DABCO BLV 0.12 0.15
DABCO T-9 0.15 0.15 0.21 0.21

Reactivity
Cream Time (sec) 13 13 12 12
Rise Time (sec) 90 91 101 103

Foam physical properties

CORE DENSITY (kg/m3) 16.8 17.6 26.9 26.6
90%COMPRESSION SETS CD (%) 9.7 11.2 2.4 2.8
TEAR STRENGTH (N/m) 227 227 315 315
Figure 1. Rise profile of 6.5 water formulation ELONGATION 9 BREAK (%) 68 66 107 109
TENSILE STRENTH (Kpa) 85 91 90 92
IFD HARDNESS (N)
Mean Load at 25% Deflection 201 211 203 193
Mean Load at 65% Deflection 381 399 366 354
Mean Load at 25% Return 110 114 141 135
Mean Modulus 1.89 1.89 1.80 1.83
RESILIENCY (%) 21 32 41 42

Foams made with experimental Polyol A were tested for

PVC staining after production in order to detect any amine
emission. No staining was observed. In conclusion, it is
feasible to consider that a slabstock foamer may in the fu­
ture produce foam by just mixing polyols, water, surfactant
Figure 2. Rise profile of 3.7 water formulation and isocyanate.

Foams made with these new experimental polyols have M DI HR M OLDING

identical physical properties with that of traditional amine
blown foams, as can be seen in Table 5. An emphasis has recently been put on elimination of amine
emission in MDI based HR molding while maintaining
good humid aging resistance. Severe tests, such as BMW’s,
VW-Audi’s, Volvo’s, Opel-GM’s are already challenging
when using conventional catalysis, as it has been described
in previous articles [1,23]. The situation becomes more
difficult when using reactive amine catalysts for the follow­
ing reasons:
1) The catalyst level has to be increased in the formulation
to maintain equivalent reactivity. Since these reactive cata­
lysts are tied up in the polymer, they tend to disrupt it
somewhat.
2) The gellation profile is changed when strong, fugitive
catalysts such as TEDA (Triethylenediamine) are replaced
with reactive amines. Hence the polymer build-up is
slowed down or at least made differently.
3) With the high concentration of tertiary amines in the
polymer chain there is always a possibility of retro-
reaction, or chemical degradation, upon foam exposure to a
combination of heat and moisture. Indeed polymer mor-
50 / Casati, Sonney, Mispreuve, Fanget, Herrington, Tu
phology changes induced by accelerated aging are already T-80 HR MOLDING
reported with conventional foam systems [24].
For introduction to the TDI-based HR molded foam mar­
The present work was focused onto the use of Active ket, we targeted Active Polyol C that would allow foam
Polyol B together with reduced amount of conventional producers to lower their added catalysts levels by 50%. We
amines and/or reactive amines in order to meet VW -AUDI reasoned that this would significantly reduce worker expo­
’s technical (T2524-03) and environmental (PVC staining) sure and lead to foam processes and products with much
requirements. This program is still underway at the time of less volatile emissions.
writing this paper. Initial results presented in Table 6 con­
firm that HACS’s (Humid Aged Compression Set) values Improvement, or at least equivalence, in both foam
of < 15 % are easily met and PVC staining is minimized. processing and physical properties is important for the
More challenging is the HALL (Humid Aged Load Loss), commercial success of any new product. Data
or CFD (Compression Force Deflection) change at 40 % demonstrating the performance of a molded foam grade of
deflection after humid aging, specifications of (-25 < active polyol in these areas will be presented using the
HALL % < + 10) when fugitive amines are removed. Foam formulations given in Table 7. Note that for comparison
density of the tested foams was purposely reduced vs the purposes, we included a formulation based on the control
52 kg/m3target to assess the limits. polyols but at 50% reduced catalysts. At nominal packing
levels, this formulation typically gives molded foams at a
Table 6. MDI HR molding core density of 32 kg/m3 (2.0 lb/ft3).
Formulation A A B B C C
VORANOLCP 6001 64 64 49 49 64 64 Table 7. Formulations for low-density TDI HR molded
SPECFLEX NC632 34 34 34 34 foams.
VORANOLCP 1421 2 2 2 2 2 2
Exp. Polyol B 49 49 Control Control Active
Water 3.7 3.7 3.7 3.7 3.7 3.7 Component At Full At 50% Polyol C
-
Niax A1 0.12 0.12 0.06 0.06 Catalysts Catalysts At 50%
Dabco 33-LV 0.35 0.35 0.27 0.27 Catalysts
Catalyst A 0.20 0.20 SPECFLEX NC 630 62.5 62.5
Catalyst B 0.50 0.50 Exp. Polyol C 62.5
Diethanolamine LF 0.60 0.60 0.60 0.60 0.60 0.60 SPECFLEX NC 700 37.5 37.5 37.5
Tegostab B-8715 LF 0.50 0.50 0.50 0.50 0.50 0.50 Niax Y-10184 1.2 1.2 1.2
SPECFLEX NE-150
Diethanol Amine, pure 1.6 1.6 1.6
Index 85 105 85 105 85 105
Dabco 33-LV 0.35 0.17 0.17
Reactivity Niax A-1 0.08 0.04 0.04
Gel time (s) 52 57 60 66 74 78
Water 4.2 4.2 4.2
Properties
Molded density 45.3 45.0 45.0 45.2 45.1 45.6 VORANATE T -80 Index 105 105 105
kg/m3
40 %CFD (Kpa) 3.3 6.6 3.8 7.8 3.7 6.8
Airflow(cfm) 1.2 0.9 2.3 1.8 2.4 1.8 Masterbatch Viscosity, 3450 3700 3600
50 %CS (%) (CD 3.2 4.6 5.4 8.0 6.4 7.4 mPa s at 25°C
Humid aging
50 %HACS (%) 7.5 11.9 7.5 10.7 8.9 9.5
CFD change 4" 18.4 20.7 23.7 26 24.5 30.1 Also listed in Table 7 is the resultant masterbatch viscosity
cycle (%)
PVC staining
for each formulation. The masterbatch is that premixture of
Ae 5.5 1.2 0.3 2.9 0.3 0.7 ingredients that will be combined with isocyanate in the
mixing head in order to cause a foam producing reaction to
TEGOSTAB* is a trademark of Goldschmidt A. G.
begin. High viscosity here makes it difficult for the foam
machine to accurately meter, mix and dispense the foam
All foam physical properties were measured according to
producing chemicals. A practical limit for many commer­
VW’s PV 3410 test methods. Catalyst A is N,N,N’-
cial foam machines is taken as 5000 mPa s. The formula­
Trimethyl N’-Hydroxyethyl-bis-Amino Ethyl Ether and
tion based on the new Active Polyol C is comparable to the
Catalyst B is 2-Hydroxypropyl Imidazole.
two control formulations and well below the stated limit.
Dual-Hardness molded parts have been produced with ac­
Foam Processing
tive polyols and based on formulations similar to the one
given in Table 6 without any difficulties. Square pads were
The ability to benchmark important factors related to how
also made with 100 parts of Polyol B and no catalysts with
well a foam formulation will process on foam making
demolding times between 4 and 5 minutes, depending on
equipment is helpful in the development and commercial
the isocyanate index. Foam properties were found similar
introduction of new products. With Dow’s in-house facili­
to the ones shown in Table 6 with Formulation C.
ties and testing capabilities, candidate products undergo
rigorous testing for the following list of processability fac­
tors.

Casati, Sonney, Mispreuve, Fanget, Herrington, Tu / 51

 Table 8. HR foam processability factors all three foam formulations from Table 7 perform excellent
Masterbatch Viscosity________________ in this test.
Free-Rise Bulk Stability______________
In-Mold Flowability___________________
Vent Stability_______________________
Rise Profile_________________________
Shrinkage__________________________
Force-to-Crush______________________
Cure______________________________

Foam Stability

The production of defect-free molded foam parts requires

that the candidate foam system possess an adequate level of
stability through each stage of the foam reaction. The dif­
ficulty of maintaining adequate stability has increased as
foam density has decreased and part designs have gotten
thinner and more complex in shape. Figure 4. Comparative free-rise stability performance.

Free-Rise Bucket Foams

An easy way to measure overall stability of the bulk foam

is the ffee-rise bucket test illustrated in Figure 3.
Figure 5. In-mold flowability test.

In this test, the mixture is dispensed into one end of a 38.1

x 76.2 x 6.3 cm (15 x 15 x 2.4 in) heated mold and pur­
posely forced to flow over three rounded humps on its way
to the vent holes at the opposite end of the mold. This
stresses the growing foam and allows for comparison of
candidate polyols on the basis of area percent voids on the
show surface of the foam pad. Data for the current study
are presented in Figure 6.

Figure 3. Free-rise bucket test for bulk stability.

In this test, a 22.7-liter (five imperial gallons) plastic

bucket receives a shot from the foam-dispensing machine.
The shot size is sufficient to totally fill the bucket with an
approximate 10-centimeter (3.9-inch) crown of foam above
the top of the bucket. Typical data recorded would include
time to reach the top of the bucket, maximum and final
foam height and an internal foam stability rating.

The internal stability rating is obtained by cutting the cured Figure 6. Comparative flowability performance,
foam in half (parallel to the direction of rise) and ranking
the severity of voids on a 1 to 10 scale. A perfect foam In this test, the control foam with the full level of added
would receive a 10 rating while a collapsed foam would be catalysts performs well below our internal development
rated at 1. This grading is easily done by visual comparison standard of no more than 10 area percent voiding. Drop-
of the candidate foam to a series of similar foams with pre- ping the catalyst levels of the control foam by 50% resulted
viously established ratings. The data in Figure 4 show that in the formation of voids over forty percent of the foam’s

52 / Casati, Sonney, Mispreuve, Fanget, Herrington, Tu

show surface. Active Polyol C at a 50% reduced catalysts Foam Tightness
loading performed better than the fully catalyzed control.
The overall moldability of a foam is often said to be better
Vent Stability with a tight formulation than a more open one. One conse­
quence of running a tight formulation is the danger of foam
Another common location for voids in a molded foam pad shrinkage if the cell-walls are not adequately opened in a
are those areas immediately below or adjacent to the vent foam crushing event. The concern here is that the foam
devices. The data in Figure 7 were obtained by inspection survive that crushing event without losing physical dimen­
and measurement of voids appearing below the vent holes sions or bursting apart. If the foam is too tight, the internal
in a simple 38.1 x 38.1 x 11.4 cm (15 x 15 x 4.5 in) heated stress forces during crushing may exceed the tensile
test block mold. The vents for that mold were 1.6 millime­ strength of the fresh foam resulting in a physical splitting
ters (1/16 inch) in diameter and were arranged in a tradi­ apart of the foam pad as it goes through the crushing de­
tional five-hole pattern on the lid of the mold. vice. This gross splitting results in a foam pad that must be
scrapped. Comparison of relative shrinkage and force-to-
crush performances are often useful in the development of
new molded foam technologies.

Shrinkage

Our test scenario for measuring shrinkage in a standard test

block of molded foam is illustrated in Figure 8.

Figure 8. Shrinkage testing.

In this test, foams are prepared in the normal manner, care­

Figure 7. Polyol C vent stability results. fully removed from the mold and allowed to stand un­
crushed overnight. The foam pad is then cut through the
The development standard we use is that no foam formula­ area of maximum vertical shrinkage, measured for mini­
tion should demonstrate more than 5 square centimeters of mum thickness and a percent shrinkage number calculated.
void area per vent. Above this level, the voids become eas­ Comparative data for the present work is presented in Fig­
ily detectable without cutting into the foam. For the present ure 9.
study, the control foam at full catalysis just meets this crite­
rion. At the 50 percent reduced catalysts level, the control
formulation is well into the unacceptable range of perform ­
ance. Active Polyol C shows an improved level of per­
formance.

Also shown in this figure are the rise or mold fill times for
each formulation. This number was recorded as the time for
the first extrusion to be visible through the center vent hole.
Twenty-four seconds is very typical for the fully catalyzed
control formulation and that is matched nicely by the for­
mulation containing Active Polyol C. The practical concern
here is that any new formulation component not cause a Figure 9. Comparative shrinkage performance.
faster rise time since most commercial foam plants have
some limitations on how fast a mold lid can be closed and At the full catalyst levels, the control formulation performs
locked. Also, the rise time should not be so slow as for the at a shrinkage level that we consider to be an upper limit.
other control formulation, so that the foam won’t still be With numbers above 50%, we have found that some com ­
rising (and still be weak) as the molds go bumping along mercial foamers have trouble running thicker foam pads
down the conveyor line. through their crushing equipment. Shrinkage performance
of experimental Active Polyol C is quite acceptable.

Casati, Sonney, Mispreuve, Fanget, Herrington, Tu / 53

Force-To-Crush Cure

A useful laboratory test for comparing relative foam tight­ For a new polyol to be commercially viable, it must con­
ness is the so-called Force-to-Crush test. Our version of tribute to an adequate level of foam cure at several critical
this test scenario is illustrated in Figure 10. points in the manufacturing process. Table 9 presents some
practical examples of why foam cure is an important per­
formance feature.

Table 9. Practical examples of foam cure.

Judged At Affecting
Mold Fill Bulk Stability and Flash
Part Demold Station Green Strength and Fingerprinting
Crusher Station Hot Set
Shipping Station Short Term Static Fatigue

The extent of cure at the moment of mold fill directly af­

fects the amount of polymer that escapes the mold through
the venting devices and parting line. If this is excessive,
valuable material is wasted and additional pad trimming
and cleanup labor may be needed.

At the demold station, the foam must have enough green

strength to resist tearing and fingerprinting due to human
handling. These imperfections can usually be repaired off­
line at some additional expense.

Another measure of cure comes after the foam leaves the

crushing device. A poorly cured foam may densify and lose
a few millimeters of its height dimension. Wrinkles may
also appear on the exterior of the pad, further detracting
Figure 10. Force-to-crush test. from its cosmetic appearance. Similar events may occur at
the shipping station when multiple pads are compressed
In this procedure, a foam is prepared in the normal manner, into a shipping container for delivery to the seat assembly
cured for the desired time and then removed from the mold plants. A foam with inferior cure may exhibit a permanent
(usually the standard test block mold) with care. The un­ loss in physical dimensions when removed from such ship­
crushed foam pad is then quickly placed in the force meas­ ping containers. The cure testing scenario used for this
uring machine and deflected to 50% of its original work is shown in Figure 12.
thickness. Initial peak load and load decay are the basic
data available. Figure 11 presents the initial peak load data
for the present study. Foam based on the Active Polyol C
falls nicely between the two control foams and all foams
tested were well below any level of concern.

Figure 12. The hot set cure test.

In this test, foam pads made in the normal way are dis­
pensed into a 38.1 x 76.2 x 6.3 cm (15 x 30 x 2.5 in) heated
mold. After a 6 minute cure cycle, the foam pad was de-
molded and immediately roller crushed at a gap setting of
1.27 cm (1/2 in). The pad was then placed in the indenter
jig where four indenter feet regionally deflect the foam to
Figure 11. Comparative force-to-crush results. various residual thicknesses. Four replicate pads are tested
at different deflection times and a final cumulative percent

54 / Casati, Sonney, Mispreuve, Fanget, Herrington, Tu

hot set calculated. Data for the present study are presented tion values shown are taken from General Motors
in Figure 13. Engineering Standards Document GM6293M.

Figure 13. Comparative cure performance. Figure 15. Compression set performance.

The control foam based on the high-functionality polyol, Historically, compression set specifications have been
SPECFLEX NC 630 at a full catalyst level shows its typi­ among the most difficult to meet. All three foam systems
cal outstanding cure performance. Lesser, but still very easily met the requirements of both the regular and humid
adequate, levels of cure were seen with reduced catalyst aged specifications. Considering that compression sets usu­
levels in both the control and Active Polyol C based foams. ally respond to improvements in airflow, it seems logical
that even better performance for the Active Polyol C based
Foam Physical Properties foams will be possible with formulation fine tuning.

All foam pads were thoroughly tested for physical proper­ The remaining foam physical properties were consistent
ties according to G.M. test methods. Of particular interest with those expected for a 32 kg/m3 (2.0 lb/ft3) core density
are the load-bearing and compression set performances of molded foam based on TDI. Figure 16 presents a summary
the new Active Polyol C versus the control foams. Data of foam physical properties wherein the performance of the
comparing the 65% IFD (indentation force deflection) per­ Active Polyol C based foam is compared to that of the fully
formance are presented in Figure 14. catalyzed control foam.

Figure 14. Comparative loadbearing performance. Figure 16. Performance features of active polyol versus
fully catalyzed control foam.
Foams of this density grade and copolymer polyol level
will generally give a 65% IFD in the range of 300 New­ In this format, the data is intended to show, for example,
tons. Reducing the catalyst levels in the control foam gave that both compression sets are improved, as is the tear
the expected increase in loadbearing. Foams based on the strength. The tensile strength of the experimental foam was
new Active Polyol C were approximately 5% harder than 7% worse than the control but both foam systems were at
the fully catalyzed control foam. least 150% over the minimum specification requirement for
foams of this density. Elongation and resiliency perform ­
Compression set data are presented in Figure 15. The fig­ ances were identical. Airflow in the experimental foam was
ure shows the results for testing both the regular and the 17% lower indicating a need for further formulation fine
humid aged 50% Cd set. These foams fall within the Gen­ tuning.
eral Motors Class C foam group and the typical specifica-

Casati, Sonney, Mispreuve, Fanget, Herrington, Tu / 55

This formulation work is underway and current develop­ 5. Wilson A.S.; Gerrard D.L.; Bowley H.J. “Problems of
ments are confirming that total elimination of bis- (2- Interaction Between Plasticised PVC and Polyurethane
Dimethylaminoethyl) ether, causing halovision (7), should Foam” (1987). Presented at PVC’ 87, Brighton, UK.,
be feasible. April 1987
6. Christffeund A.; Huygens E.; Eling B. “Amine-Free
CONCLUSION Catalyst Systems for Automotive Instrument Panels”
(1991), Proceedings of Polyurethane World Congress,
The present paper has demonstrated that amine emissions Technomics: Lancaster, PA., 272-277
can be drastically reduced and even eliminated from flexi­ 7. Polyurethane Amine Catalysts: Guidelines for Safe
ble polyurethane foams through the use of new Dow active Handling and Disposal; Technical Bulletin AX173,
polyols combined or not with reactive catalysts. November 2000; Alliance for the Polyurethane Ind-
stry: Arlington, VA.
Further development of active polyols and adjustment of 8. Dernehl C.U. (1966), “Health Hazards Associated with
formulations are underway at Dow with the aim to give Polyurethane Foams”, Journal of Occupational Medi­
several options to the polyurethane industry to eliminate cine, 8/2, 59-62
amine emissions. Volatile and odor tests, not reported here, 9. Belin L.; Wass U.; Audunsson G. and Mathiasson L.
are being carried out routinely to make sure these programs (1983), “Amines: Possible Causative Agents in the
are giving the desired outcomes. Development of Bronchial Hypperreactivity in Work ­
ers Manufacturing Polyurethanes from Polyisocy­
The next step is the reduction of other volatile and fogging anates”, British Journal of Industrial Medicine, 40,
components in overall formulations in order to support the 251-257
polyurethane industry’s strive to very low emissions and to 10. Bugler J.; Maddison P.; Hope S.J.; Mills I.G.; Nutt
meet new stringent OEM’s specifications. A.R. and Wright M.D. (1992) Tertiary Amine and
Ethanolamine Catalysts in the Polyurethane Foam In ­
ACKNOWDLEGEMENT dustry: New Methods of Detection and Results of Per­
sonal Sampling in Nine UK Flexible Foam Factories”,
The authors wish to thank their many Dow colleagues who Cellular Polymers, 11/3, 171-200
had a role in generating the data presented in this paper. 11. Reed D. (1999), “VW aims to cut amines”, Urethanes
Special thanks are extended to Alain Bleton, Lance Coo­ Technology, 16/4, 3
per, Richard Elwell, Olga Milovanovic, Chris Noakes, 12. Boinowitz T.; Burkhart G.; Schloens H.H. (1999)
Olivia Renevey, Alan Schrock, Mark Sonnenschein, An­ “Emanations from Flexible Slabstock Foams - Signifi­
toine Storione, Ross Polk, Eve Taylor, Mary White, and cant Reduction of Emissions is Now Possible”, Pro ­
the Meyrin and Freeport foam testing laboratories person­ ceeding of Polyurethanes Expo 99; Technomics:
nel. Lancaster, PA., 283-287
13. Zimmerman R.L.; Renken T.L. (1988): “Low Odor
BIBLIOGRAPHY Reactive Amine Catalysts for Polyurethane Foams”,
Proceedings of the SPI-31st Annual Techni-
1. Broos R.; Casati F.M.; Herrington R.M.(1999), “En ­ cal/Marketing Conference, Technomics: Lancaster,
durance of Polyurethane Automotive Seating Foams PA., 36-40
under Varying Temperature and Humidity Condi­ 14. Mercando L.A.; Kniss J.G.; Tobias J.D.; Plana A.;
tions”, Journal of Cellular Plastics, Volume 36 - May Listemann M.L.; Wendel S.; “Non-Fugitive Catalysts
2000, 207-245 for Flexible Polyurethane Foams”, (1999) Proceeding
2. Casati F.M.; Broos R.; Herrington R.M. ; Miyazaki Y. of Polyurethanes Expo 99; Technomics: Lancaster,
(1997), “Tailoring the Performance of Molded Flexi­ PA., 103-134
ble Polyurethane Foams for Car Seats”, Proceedings of 15. Kometani H; Tamano Y.; Masuda T.; Van Maris R.;
the 1997 SPI Polyurethane World Congress”, Tech- Gay K.; (2000) “The Investigation of Amine Emis­
nomics: Lancaster, PA., 402-420 sions from Polyurethane Foam”; Proceeding of Poly ­
3. Casati F.M.; Herrington R.M.; Broos R . ; Miyazaki Y.; urethanes Conference 2000; Technomics: Lancaster,
Phan Thanh H and Cadolle D.; (1998) “Automotive PA., 11-21
Seating Foams: Improving their Comfort and Durabil­ 16. Rothe J.; Cordelair H.; Wehman C.; (2000) “New
ity Performance”, “Proceeding of Polyurethane the Catalysts for Low VOC in Flexible Slabstock Foam”,
1998 SPI Polyurethane conference”; Technomics: Proceeding of Polyurethanes Conference 2000; Tech ­
Lancaster, PA., 417-431 nomics: Lancaster, PA., 109-116
4. Williams K.P.; J. Gerrard, D.L. “The Degradation of 17. Sikorski M.; Wehman C.; Cordelair H. (1999) “New
Poly(Vinyl Chloride) Studied using Fourier-Transform Additive Solutions for Low VOC in HR Molded
Ramam Spectroscopy” (1990), Eur. Polym. J., 26 (12)- Foams”. Proceedings of Polyurethanes Expo 99;
1355-8 Technomics: Lancaster, PA., 135-144

56 / Casati, Sonney, Mispreuve, Fanget, Herrington, Tu

18. Mapleston P; “Riding Comfort is a Focus of Automo­ Ronald M. Herrington
tive Foam Systems” (2001), Modem Plastics, March
2001, 30-31 Ron Herrington is a Development
19. Broos R.; Sonney J.M.; Phan Thanh H.; Casati F.M.; Associate in the Polyurethane
(2000) “Polyurethane Foam Molding Technologies for Products Technical Service and
Improving Total Passenger Compartment Comfort”, Development Laboratory of The
Proceedings of the Polyurethane Conference 2000, Dow Chemical Company in Texas.
Technomics: Lancaster, PA., 341-353 He received a Bachelor’s Degree in
20. Malwitz N.; Manis A.; Wong S.W.; Frisch K.C.; Chemistry from the University of
(1986) “Amine Catalysis of Polyurethane Foams”, Houston in 1970. He joined Dow in
Proceedings of the 30th Annual Polyurethane Techni- that year and has since gained wide
cal/Marketing conference, 1986, Technomics: Lancas­ experience in the fields of slabstock
ter, PA., 338-353 and molded flexible foams.
21. Nutt A.R.; Skidmore D.W. (1987) “Recovery of
Chlorofluorocarbon 11 by Activated Carbon Scrub­ Jean-Marie Sonney
bing on a Polyurethane Foam Slabstock Plant”, Cellu­
lar Polymers, Volume 6, 1987, 62-77 Dr. Jean-Marie Sonney graduated
22. Herrington, R., De Genova, R., Casati, F. and M. from the Swiss Federal Institute of
Brown (1997), “Molded Foams, Chapter 11”, In Technology in Lausanne
Flexible Polyurethane Foams; Herrington, R. and K. (Switzerland) and received his Ph.D.
Hock, eds., The Dow Chemical Company, Form No. degree in physical organic chemistry
109-01061, 11.1-11.41 from the same institution in 1979.
23. Herrington R.M. and Klarfeld D.R. (1983) “Humid After a post-doctoral research
Aged Compression Set Phenomena in All Water fellowship at the University of
Blown HR Molded Foams”, Proceedings of the 1983 California, Santa Cruz, he joined the
SPI Polyurethanes Conference, Technomics: Lancas­ Geneva­based Research and
ter, PA., 177-182 Development group of BP Chemicals in 1981 and was
24. Brasington R. and De Roeck H. (2000) “Accelerated transferred to the Dow Chemical Company in 1989. During
Testing for Durability Performance of Automotive these years, he had various responsibilities in the field of
Seating Foam”, Proceedings of the API 2000 confer­ Analytical Chemistry, Quality Assurance, EH&S and Qual­
ence, Technomics: Lancaster, PA., 267-279 ity Management. He is currently a project engineer in the
Development Group for Molded Foams and Specialties
BIOGRAPHIES with responsibilities for NVH applications.

Francois M. Casati Johnson Tu

Frangois M. Casati graduated from ICPI (F), now CPE- Johnson Tu is a Sr. Development
Lyon (Ecole Superieure de Chimie Specialist in the Research and
Physique Electronique de Lyon), in Development Group of The Dow
1967. He has over 30 years of Chemical Company in Freeport,
experience in Polyurethanes, Texas. He received a BS Science
Industrial Amines and Biocides, with from the Chengkun University in
S.N.P.E., Recticel, Abbott Taiwan in 1983. He joined Dow in
Laboratories, BP Chemicals and The 1990 and has held several positions in
Dow Chemical Company. During R&D/TS&D and new Business
that time he has held various Development.
positions in Manufacturing, Marketing and Research &
Development. He is currently a Product Development
Leader, R&D, for the Flexible Foam business of Dow.

Casati, Sonney, Mispreuve, Fanget, Herrington, Tu / 57

Henri Mispreuve Alain Fanget

Dr. Henri Mispreuve graduated from Alain Fanget joined the Geneva-
the State University of Mons based Polyurethane Research and
(Belgium) in 1975 and received his Development group of Union Car­
Ph.D. in Physical Organic Chemistry bide Europe in 1978 and was then
from the same University in 1979. In transferred to BP Chemicals in 1979
1984, he joined Dow Europe S.A. in and to The Dow Chemical Company
Horgen (Switzerland). He is in 1989 where he worked in various
currently responsible for Technical PU moulding and ACES
Service and Development of applications.
Polyurethane Slabstock Foams at the company’s Interna­ He is currently a Technical Service
tional Development Center in Meyrin/Geneva (Switzer­ Specialist in the Automotive Thermosets R&D Group of
land). the Meyrin International Development Center, with specific
responsibilities in seating and NVH moulding applications.

58 / Casati, Sonney, Mispreuve, Fanget, Herrington, Tu

 SESSION C
TESTING

TECHNICAL SESSION ORGANIZERS

Sue Clift, Chair

BAYER CORPORATION

Ken McDaniel
BAYER CORPORATION

Allan Schrock
THE DOW CHEMICAL COMPANY

TECHNICAL SESSION MODERATOR

Steve Robbins
AIR PRODUCTS & CHEMICALS, INC.
Monitoring of Polyurethane Foam Cure

STEPHEN A. JONES, KEITH W. SCOTT and ERIC A. SHEARD

BRYAN G. WILLOUGHBY Lescon, Inc.
Rapra Technology Ltd. 551 Sunset Road
Shawbury, Shropshire, SY4 4NR Deer Isle, Maine 04627
United Kingdom

ABSTRACT rheometers, the sample chamber of the VNC is not

enclosed, and the solidification of the sample is not a
The curing of polyurethane foam involves complex problem. The concept and operating principles of the
changes - changes from mobile liquid to rubbery, or even Rapra VNC have already been reported [1,2].
rigid, product are accompanied by changes in density and Whilst the Rapra VNC offered the benefit of a continuous
sample size. All these take place under conditions where trace as a record of cure, its original configuration offered
there may be rapid changes in temperatures. The ability only a single trace - effectively the change in needle
to profile and record such events has traditionally required damping as cure progressed. However by bringing the
skilled, often time-consuming observation. However VNC under computer control, so that the monitoring
where non-foaming systems are concerned, technologists frequency can be changed at will, then a second output is
have had access to a variety of instrumentation that can obtained if the resonance condition can be maintained
monitor changes in viscosity or the development of elastic throughout the cure. This is the operating principle of the
character. This facility is now available for foamed Scanning VNC (SVNC), which can provide a two-
polyurethane, with a development of the Rapra Vibrating dimensional profile of the rheological changes during
Needle Curemeter (VNC). The development involves the cure. The application of this instrument to the
integration of the existing instrument with detectors for characterisation of a variety o f the curing systems, PU and
rise and exotherm. Continuous response/time plots are non-PU, has already been described. [3,4,5]
thus produced; a response for the development of visco­ The same outputs are, in principle, accessible when the
elastic properties being added to those for height and VNC or SVNC are applied to foaming cures - systems
temperature. The result is a unique fingerprint for the which have hitherto provided a severe challenge to
cure which is amenable to storage and archiving. This conventional rheometers. Traditionally there have been
paper will show the outputs for foaming PU systems and problems when attempting to monitor a sample of
discuss some of the insights accessible with this powerful changing geometry, and when the sample is within a
tool. cavity which may itself provide a significant heat sink.
The latter effect means that the actual process of
INTRODUCTION monitoring becomes intrusive to a process which is
critically dependent on self-generated heat for cure.
During the cure of a polyurethane, what may start off as a Such disadvantages are not experienced with the VNC,
relatively free-flowing liquid may finish as a solid with which places no special constraints on the geometry of the
recognisable elasticity. As far as dense - as distinct from sample and hence can accommodate volume changes
foamed - materials are concerned, this change in during cure.
character is amenable to monitoring by classical In effect the only change required to allow the VNC to be
rheometers although the cost of the instrumentation may adapted for the monitoring of foams is to allow sufficient
preclude their use for routine work. Furthermore the room for the sample rise. A practical configuration has
prospect of the sample setting solid within the monitoring the vibrator mounted on an 80x50x50 cm alloy frame, and
cavity monitoring liquid polymer cures. Unlike classical utilises a lightweight non-metallic “needle” of up to 50cm

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 61

 Figure 1. The Foam SYNC

in length. However in configuring a dedicated “Foam Tegostab B8408 - polyether-modified polysiloxane

VNC”, the opportunity has been taken to incorporate surfactant (ex TH. Goldschmidt AG)
additional outputs into the monitoring system. These are Distilled water
measurements of foam rise and temperature. The former Solkanel41b - HCFC blowing agent (ex A-Gas)
is obtained via an infrared detector, and the latter by an Suprasec 5001- Diphenylmethanediisocyanate (ex
embedded thermocouple. In this way four outputs are Huntsman Polyurethanes; supplied by Brian Jones &
available, all of which can provide essentially continuous Associates Ltd)
responses.
The standard geometry described above allows samples to MIXING PROCEDURE
be monitored in conventional plastic buckets of up to 25
litres in size: the basic layout is shown in Figure 1. The same basic rigid polyurethane foam formulation and
mixing procedure was used throughout this work. To 150
MATERIALS USED grams of Arcol 1030, 1.5 grams of Tegostab 8404 and
2.85 grams of water were added and mixed. To ensure
The balance between the competing chemical reactions thorough mixing a propeller blade rotating at
can have a marked effect on the resulting polyurethane approximately 3,000 rpm was used. 2.25 grams of PC
foam. A small change in the formulation can alter the Cat DMCHA and 27 grams of blowing agent were added
balance of these reactions and hence the polyurethane and again mixed. To this 204 grams (105%) isocyanate
product. By studying a basic polyurethane system, and was added and mixed for approximately 5 seconds. The
making alterations to this formulation we will show how full polyurethane formulation was then transferred into
the Foam VNC provides a valuable insight into the the Foam VNC bucket and the data collection
polyurethane cure / foaming process. commenced.
Raw materials used for this series of experiments are All the data, except Formulation II-S, presented within
given below:- this paper was collected using the Foam VNC in “Fixed
Arcol 1030 propylene oxide polyol equivalent weight Frequency” mode. That is the needle was forced to
140; tri-functional (ex Arco Chemie Netherlands; vibrate at a single frequency throughout the data
supplied by Brian Jones & Associates Ltd) acquisition period and not following the natural increase
PC Cat DMCHA - dimethylcyclohexylamine (ex Nitrol in resonance as the foam cures. Although this has the
Europe Handels GmbH) effect of ‘losing’ the frequency trace, the rate of data
PC Cat NP90 - 70% Bis (2-dimethylaminoethyl) Ether in acquisition is up to 20 times faster in fixed mode. Data
Dipropylene glycol (ex Nitrol Europe Handels GmbH) for all trials was collected over a 30 minute period.

62 / Jones, Scott, Willoughby, Sheard

 Figure 2. Height v Time with varying levels of blowing agent

INFLUENCE OF HCFC BLOWING AGENT LEVEL The temperature probe was located 3 cm from the base of
the vessel and inserted 5cm into the area where the foam
It should be recognised from the outset that there are two would form. Formulation I, containing the lowest level of
blowing agents present, one (C 02) a by-product of system blowing agent can be seen to have attained the highest
chemistry and the other (HCFC 141b) a physical blowing maximum temperature of 139°C at 6 minutes 6 seconds.
agent. The latter of these was varied in three formulations Formulation II reached 131°C at 6:12 and Formulation III
studied here. Formulation I was prepared with a 50% 126°C at 6:39. As only a finite amount of energy is
deficiency of blowing agent; Formulation II the normal released by the exothermic polymerisation reactions there
level and Formulation III had a 50% increase. Figure 2 is less energy available to heat the foam and hence a
shows the obtained Foam VNC plot for height (mm) lower maximum temperature is observed.
versus time (seconds). It can be seen, as would be The rate of temperature rise is also slower with the greater
expected, that increasing the level of blowing agent level of physical blowing agent. O f course, the lower the
produces a greater rise and therefore less dense foam. temperature at any point then the slower are the reactions.
Of course rise is routinely studied as a means to profiling Thus the interdependency of the ingredients is being
foaming PU cures. However, the Foam VNC allows other displayed: i.e. increasing the level of physical blowing
continuous outputs to be obtained, and in Figure 3, the agent acts as an energy sink and retards the polymer­
temperature/time profiles for these same cures are shown. forming chemistry.

Figure 3. Temperature v Time with varying levels of blowing agent

Jones, Scott, Willoughby, Sheard / 63

 Figure 4. Amplitude of Vibration v Time with varying levels of blowing agent

The effect on polymer formation can be seen in the needle amplitude. By comparing Figures 2-4 it can be
change in the amplitude of needle vibration with time. deduced that, in these examples, an increasing in the level
Figure 4 shows the changes for the three different foam of physical blowing agent:
formulations. The reduction in amplitude (shown as • increases the extent of foam rise, but
relative digital response) corresponds to an increase in • slows down the polymer reactions, and
damping. Each of the cures is characterised by a major • delays the build­up in foam properties.
increase in damping which starts at least 2 minutes into
the cure and continues for a further 2-2.5 minutes. This INFLUENCE OF CATALYST LEVEL
onset of this change coincides with the end of rise and
correlates with the model of Neff and Macosko for the Four formulations were studied with increasing levels of
development of physical properties within a water-blown dimethylcyclohexylamine catalyst. This amine is capable
foam [6]. The three profiles in Figure 4 reveal that this of catalysing both polymer formation (“gelling”) and C 0 2
change is occurring later in the cure as the physical production (“blowing”) with the balance in activity in
blowing agent level is increased. Thus a more complete favour of gelling [7]. The level of DMCHA catalyst
picture of the foaming process can be obtained from these added to each formulation was as follows
three complementary profiles of rise, temperature and

Figure 5. Amplitude of Vibration v Time with varying levels of DMCHA catalyst

64 / Jones, Scott, Willoughby, Sheard

 Figure 6. Height v Time with varying levels of DMCHA catalyst

Formulation IV 1.20 grams shows the development of polymer properties occurs

Formulation V 1.65 grams earlier the higher the catalyst level, and Figure 6 shows
Formulation II 2.25 grams the corresponding increase in the rate of blowing. Also it
Formulation VI 4.50 grams will be observed in Figure 5 that the profile o f property
The increased overall activity is reflected in the development with time changes as the catalyst level
temperature profile, where the peak exotherm was increases as the initial drop in amplitude is very much
reached earlier as the catalyst level was increased. The shaper for Formulation VI.
time required to reach the maximum temperature for each Figure 6 also shows that the maximum height recorded for
formulation was also recorded as:- these foams also increases slightly as more catalyst is
Formulation IV 132°C at 9 minutes 1 second added. This presumably reflects the comparative rates of
Formulation V 133°C at 7 minutes 17 seconds cure and blow. The point at which the amplitude traces
Formulation II 131 °C at 6 minutes 12 seconds shows a sharp decrease (i.e. when properties start to
Formulation VI 138°C at 5 minutes 9 seconds develop) is a diagnostic point in the cure profile. Only a
That the catalyst is effective with respect to both blowing small amount of expansion appears possible after this
and gelling can be seen in the amplitude and height point. A comparison of the data in Figures 5 and 6 shows
responses in Figures 5 and 6, respectively. Figure 5 that only about 30mm of foam expansion is available after

Figure 7. Amplitude of Vibration v Time with varying levels of water

Jones, Scott, Willoughby, Sheard / 65

 Figure 8. Height v Time with varying levels of water

the sharp decrease in amplitude is observed. The degree stoichiometry was kept the same throughout, so that
of blow at this point is therefore important and this will be additional water required corresponding additions of
seen to reduce slightly as the DMCHA level is reduced. isocyanate. As more water is added to the formulation
Thus reducing the catalyst level has a greater effect on (and more isocyanate), more carbon dioxide and urea will
blowing than on cure. This is an expected consequence of be produced, e.g.
the known selectivity of this catalyst. Hitherto this
selectivity might have been seen in a reduced overall rise H20 + — N =C=0 -----­ NH2 + C 0 2
(higher foam density) for a lower level of catalyst: the
Foam VNC allows the balance between blow and cure to ~~~NH2 + ------ N = C = 0 - > -----NH -CO -NH ----­ (Urea)
be observed as the foam is actually being made.
Increasing the generation of C 0 2 will result in a less
INFLUENCE OF WATER LEVEL dense foam being produced, as seen in Figure 8 for the
sequence Formulation VII > VIII > II. The formation of
Figures 7 and 8 show the needle amplitude and foam urea groups and the extra heat generated contribute to the
height profiles for three different levels of water in the greater rate of reaction observed in Figure 7 as the level of
formulation.. Formulation II contains 2.85 grams of water is increased. The maximum observed temperatures
water, Formulation VIII contains 2.25 grams and were:-
Formulation VII only 1.5 grams. The overall NCO/HO

Figure 9. Amplitude of Vibration v Time with varying levels of surfactant

66 / Jones, Scott, Willoughby, Sheard

 Figure 10. Height v Time with varying levels of surfactant

Formulation II 131°C at 6 minutes: 12 seconds

Formulation VIII 128°C at 6 minutes 46 seconds INFLUENCE OF SURFACTANT LEVEL
Formulation VII 126°C at 7 minutes 2 seconds
The characteristic point at which the amplitude starts to Silicone surfactants lower the surface tension of the foam
decrease (Figure 7) occurs earliest in the cure for the mix mix, improve the dispersion of nucleating air and stabilise
with the highest water content (Formulation II) and the rising foam until it is self supporting. By lowering the
profiles from this point display subtle differences in shape level of surfactant this should destabilise the foam and
which may reflect the manner in which urea groups are result in a more dense, probably unusable product.
formed and contribute ta the final polymer properties.
This process is considered further below.

Figure 11. Temperature v Time with varying levels of surfactant

Jones, Scott, Willoughby; Sheard / 67

 Figure 12. Four response traces for Formulation II in Scanning Mode

Figures 9 and 10 show the amplitude and height traces with implications for process control, the actual shapes of these
varying levels of surfactant. No significant differences in amplitude traces over this period may yet yield more
either the amplitude (cure) or height traces are observed by fundamental information about the changes occurring at
reducing the surfactant level from 1.5 grams (Formulation this critical stage in cure.
II) down to 0.75 grams (Formulation X). With no
surfactant added to Formulation IX it is not surprising to THE FOAM VNC IN SCANNING MODE
see the foam does not attain the same height as with the
surfactant present. Interestingly the transitional period as All the data presented within this paper has been collected
properties are developed occurs over a shorter time-scale using the foam VNC in “Fixed Frequency” mode to utilise
when no surfactant is present (Figure 9). It is clear from the greater data acquisition speed available. Figure 11
this, that the presence of surfactant is influential also on the shows data for Formulation II collected in full “Scanning”
nature of the polymer produced. mode. The extra trace can be seen for the shift in
It will also be seen that the initial drop in amplitude is resonance frequency which reaches a maximum value of
particularly sharp for Formulation IX, a feature which 563 hertz. The amplitude, frequency and height profiles
bears some similarity with Formulation VI discussed have been normalised and represented as percentage
earlier. Yet the two formulations are quite different: change to enable all traces to be shown on a single scale.
Formulation IX has no surfactant and only half the catalyst The resonance frequency trace reflects the changing
level of Formulation VI. However, one feature which the modulus (or hardness) in the product, a higher resonance
two mixes may share in common is the high temperature frequency reflecting a higher elastic modulus. Further
reached during cure: 141°C for Formulation IX and 138°C work investigating the resonance frequency plot is
for Formulation VI. These were the highest temperatures envisaged.
reached amongst the formulations studied here.
Whilst a higher exotherm might be an expected CONCLUSIONS
consequence of a higher catalyst level, the influence of
surfactant on this is perhaps less expected. The effect is The Scanning VNC has been combined with temperature
significant: Formulation XI attained a temperature of and rise detectors to provide simultaneous outputs to
141°C compared to 131°C for Formulation II (Figure 11). reflect the changes occurring during polyurethane foaming.
It may be that the influential parameter here is surface These outputs have been shown to provide a convenient
energy. Without surfactant, a foam of lower specific and useful method of assessing and comparing
surface area is obtained and consequently less energy is polyurethane foam cure. The Scanning VNC has already
consumed in blowing. Whatever the cause, the effect of proved to be an invaluable tool finding uses in areas such
increased temperature is readily recognisable in the as material development and quality control. With the
amplitude/time plot. The transitional period as properties incorporation of temperature and height profiles, the foam
begin to develop occurs over a much shorter time-scale as instrument should now provide similar benefits to the
the temperature increases. Whilst this has obvious polyurethane foam industry.

68 / Jones, Scott, Willoughby, Sheard

REFERENCES consultancy activities to polymer industry particularly in
the areas of cure monitoring, formulation development and
1. Willoughby, B. G., 1982, Rubber World, 187 (3) 26- flame retardancy.
33.
2. Willoughby, B. G., Scott, K. W. and Hands, D., 1990 Bryan G. Willoughby
“Cure Rationalisation using the Vibrating Needle Bryan Willoughby, a polymer chemist by training (BSc,
Curemeter” proceedings of the Rapra Seminar “Flow 1966; PhD.1971)
and Cure of Polymers”, Shawbury, U.K., 22-23 joined Rapra
March. Technology Ltd in
3. Willoughby, B. G. and Corfield, M., 1996, “The 1972 to service the
Scanning Vibrating Needle Curemeter - Development growing industrial
and Theory”, PMA Spring Meeting, Wesley Chapel, consultancy business.
FI., 16 April. Dr. Willoughby has
4. Sheard, E. A. and Willoughby, B. G., 1997 Adhesives initiated much
Age, 40(3), 44-49 supporting background
5. Jones, S. A., Scott, K. W., and Willoughby, B. G., “PU research, notably with regard to curing, ageing and process
Cure Mapping with the Scanning VNC”, PMA Spring emissions, and is the author or co-author of more than fifty
Meeting, FI., 4 April. papers and patents.
6. Neff, R. and Macosko, C.W., 1995, “A Model for
Modulus Development in Flexible Foam”, proceedings
of “Polyurethanes 1995”, 26-29 September.* Eric A. Sheard
7. Kometani, H., Tamano, Y., and Yoshimura, H., 1997
“The Characteristics and Role of Tertiary Amine Eric Sheard is the president of Lescon Inc., representing
Catalysts for Polyurethane Foams”, proceedings of the many raw material suppliers
Rapra Seminar “Catalysts in Polyurethane Foams”, and equipment manufacturers
Shawbury, U.K., 6 November. in the United States of
America. Eric obtained a BS
Ch Eng. from Newark College
BIOGRAPHIES of Engineering followed by
extensive post graduate studies.
Stephen A. Jones Eric has been the president /
Stephen Jones joined Rapra Technology Ltd in 1987 and chairman of the New York
has subsequently Rubber Group and the
obtained his first polyurethane Manufacturers
degree (BSc, Applied Association. Eric has lecturer, authored papers and lead
Chemistry) in 1994. seminars on subjects such as liquid elastomers, urethanes
Stephen has been and marketing.
involved in chemical
consultancy to the
polymer industry and
the development of the
VNC range of
instruments since 1994.

Keith W. Scott
Keith Scott joined Rapra Technology Ltd in 1982
immediately after obtaining his first degree (BSc,
Chemistry). He gained
a postgraduate degree
(Mphil, 1987) for a
thesis entitled
‘Synthesis and
Characterisation of
Polymer Supported
Catalysts and Wetting
Agents’. Keith has
been involved in wide
ranging chemical

Jones, Scott, Willoughby, Sheard / 69

Three-Way Thermal Conductivity Instrument Comparison

BRAD CANNEY, CRAIG DIXON and NANCY MATHIS

Mathis Instruments Ltd.
8 Garland Court
Fredericton, New Brunswick
Canada

ABSTRACT EXPERIMENTAL
Three thermal conductivity techniques have been Instruments
compared on polyurethane foam: guarded hot plate (GHP),
transient plane source (TPS) and modified hot wire Three instruments were used to measure thermal
(MHW). Trends and technique differences will be conductivity: Guarded Hot Plate (GHP), Hot Disk (HD)
discussed in the areas of contact resistance, homogeneity and TC Probe (TCP). The GHP method is steady state and
influence, variation through the thickness and mounting follows ASTM D518 [1]. The GHP results were
pressure. Strengths and weaknesses of the techniques will conducted independently and the results were received by
be discussed in the areas of test time, sample the authors. The other two instruments systems (TCP and
preparation/size, ASTM acceptance and calibration. HD) were interfacial and transient in operation rather than
the traditional steady state technique in GHP.
INTRODUCTION Both of the transient techniques are based on a principle
Thermal conductivity is a critical property in the that involves supplying a known amount of energy in the
insulation industry. It measures the ability of the material form of electrical current to an initially isothermal sample
to transmit heat based on the Fourier equation Q=kAAT/L via a contact sensor while monitoring the resulting
where: Q is the heat in W, k is the thermal conductivity of temperature rise by using the sensor also as a resistance
the material in W/mK, A is the cross sectional area thermometer. The sensors act as both a heat source and a
(m2)through which Q flows, L is the distance (m) through dynamic temperature sensor for measuring the thermal
which Q flows and AT is the temperature gradient (K). transport properties of a material. The sensors measure the
Insulations, by definition, are materials with low thermal temperature rise representing blocked heat flow from the
conductivity. These include not only polyurethane, but instrument sensor into the material tested. The lower the
polystyrenes, fumed silica, fiberglass, glass wool, cork and conductivity, the more heat is trapped and the higher the
a variety of other materials. In high temperature interfacial temperature rise.
applications, ceramic is the insulator of choice, although The heat from the element penetrates into the sample, but
the thermal conductivity at 50 mW/mK would not be never reaches the far side. The length of the test determines
considered low by those in the appliance or construction the penetration depth. Typically for polyurethane a
markets. On the lower end of the insulation spectrum, penetration depth of 2 mm would be expected during a 10
vacuum insulation panels can achieve conductivities as low second test. Results from testing correspond only to the
as 4 mW/mK. Over this range, it is key to be able to volume of material penetrated by the heat wave; therefore
characterize the thermal conductivity as a necessary design it is basically a surface measurement
property. The goal of this paper will be to compare
existing and innovative methods to measure thermal
conductivity.

,
POLYURETHANES EXPO 2001 SEPTEMBER 30-OCTOBER 3, 2001 71
Figure 1 (above): The transient techniques only penetrate
a few mm so the sensors must be in contact with the foam.

Figure 2 (right): The HD sample holder (right) stacks two

samples with the sensor between. The TC Probe requires
only one sample - shown with a surface weigh.

Materials Hot Disk, the technique incorporates a heating element that

is supported on a backing, which provides a rectangular one­
The sample set included 4 cardboard sandwiched dimensional heat flow. The heating element relies on
polyurethane squares. These were received in 300mm2 (1 interfacial contact with a solid flat sample having
ft2) by 50 mm (2in) formats that were sized for the GHP minimum dimensions of 5 mm by 25 mm.
testing that had occurred. The cardboard was 0.475 m. The The method requires a calibration with characterised
foam samples were cut into smaller sample sizes that were standard materials. The GHP thermal conductivity values
appropriate for the sample holder used in these for three of the four samples were used as reference
experiments and to facilitate efficient testing of their materials. This is a common technique for an end user to
exposed polyurethane faces (see Figure 1). When create calibration standards for a transient technique with
performing measurements, the Hot Disk sensor was fitted an established ASTM method.
between two sample halves, with each piece having a plane The timing options used for instrument calibration in
surface facing the sensor. The TC Probe™ utilized one of evaluating the foam samples were a 10 second test duration
this sample pair, also testing on the exposed PU surface and a 4 second test start time. The calibration plot of sensor
(see Figure 2). In contrast, the GHP measurements were response for the TC Probe™ using three foams as thermal
conducted with the entire sample with, the outer cardboard conductivity standards is illustrated in Figure 3. Upon
in contact with the hot and cold plates and therefor their completion of the calibration, the TC Probe™ was used to
contribution was incorporated into the results. measure the remaining unknown sample and the three
calibration samples.
Hot Disk

This instrument measures both thermal conductivity and

diffusivity, and from these two, volumetric specific heat
can be calculated. The standard software module was used
which is appropriate for materials with thickness typically
greater than 2 mm and thermal conductivity between 0.005
to 500 W/mK. Data points 10 to 200 were analyzed to
report the final thermal property results. Optimized testing
parameters included an 80-second test duration, 0.02 Watt
output of power to the sensor; a 9.734mm radius sensor
was used in these experiments. The “bridge” option was
used during the testing for optimum sensitivity.

TC Probe
Figure 3: TC Probe™ was calibrated with three samples of
The instrument is based on a transient modified hot wire polyurethane characterized by GHP. The error bars
(MHW) heat reflectance technique[2]. In contrast to the represent RSD for 5 tests in one location on the sample.

72 / Canney, Dixon, Mathis

 Table 1: Thermal Conductivity Results (W/mK)

Sample GHP TCP % diff HD Bridge % diff

A 0.0212 0.0212 0.08% 0.0304 43.19%
B 0.0213 0.0213 0.01% 0.0301 40.95%
C 0.0218 0.0217 -0.26% 0.0306 40.50%
D 0.0228 0.0223 -2.04% 0.0316 38.69%

RESULTS AND DISCUSSION

The second consideration regarding the bias is the
The results of the thermal conductivity testing with the contact resistance. HD measurements are not influenced
three instruments are presented in Table 1. The percentage by this effect as long as the contact pressure is maintained
difference in the columns is taken in reference to the GHP constant during the test time (80s) [4]. Location of the heat
result. The results are plotted in Figure 4. wave in the sample moves out from the sensor during the
It is clear from the table that a significant bias exists test. The heat is therefore resident in the boundary during
between the HD and the GHP. The graph shows that the initial test period (<0.5s). During the data
while, although a bias in the absolute value exists, the trend manipulation, the first data points are deleted and the slope
is the same and the results correlate well. Both the TCP of the data is examined. Interfacial contact resistance is
and HD have stated precision of 2% and accuracy of 5%. reflected in the data by an early time jump in the
The HD measurements in Figure 4 for are an average of 5 temperature difference, but that does not effect the later
tests which had precision better than 1% in all cases. The slope. A firm pressure is placed on the sample during the
potential of material and or technique differences will be measurement to mitigate this effect as well so that only a
examined. small amount of data from the start must be eliminated. By
The HD measurements are taken orthogonal to the GHP removing the contact resistance, the measured conductivity
measurements. It has been demonstrated previously [3] would be higher.
that thermal conductivity is anisotropic based on the A third contribution comes from the variation in the
direction of foam rise. To investigate this, the cardboard sample. If the core has a lower conductivity, the GHP,
was removed from the faces of the sample and the HD tests which measures across the entire profile, would average in
were repeated with the sensor in contact with that face. In those contributions, where the HD which only measures at
this way the heat was flowing in the same direction for the the surface would not. Again, this would generate a higher
two measurements. The cardboard to cardboard direction value for the HD measurement.
was referred to as facial and the results in table 1 were The TCP measurements calibrate out all of these affects
referred to as edge. The facial measurements were 10% by relating back to characterized values obtained by GHP.
lower than the edge testing. This would be consistent with The calibration, although an extra step, removes this bias.
the assumption that the direction of foam rise would be The advantages of the transient methods are two fold: test
parallel with the cardboard face. time and sample configuration.

Figures 4a and b: The results from the steady state (GHP) and transient methods (TCP and HD) correlate very well, but the
HD results (right) have a 40% higher value. This is due to three factors: testing orientation, contact resistance, and density
variation. Transient results had test times of 10 s (TCP) and 80 s (HD).

Canney, Dixon, Mathis / 73

CONCLUSIONS BIOGRAPHIES

The thermal conductivity of four samples of Brad Canney

polyurethane has been tested with three instruments with Brad is a Research and
different test methodologies. Development Scientist at
The TC Probe, which was calibrated with GHP Mathis Instruments Ltd. He
characterized samples, showed excellent correlation and holds a B.Sc. in Biochemistry
value agreement with the materials tested. Variation from from the University of New
GHP vales was under 2%; within the precision of the Brunswick, Canada. His
instrument. ongoing work will continue to
The Hot Disk, which did not require calibration, showed focus on new applications of
a correlation to the GHP results of 0.93, but had a bias of thermal conductivity testing
40%. This bias has been postulated to be due to a with transient techniques.
combination of measurement orientation, contact resistance
and material density variations.
The transient methods were evaluated to demonstrate Craig Dixon
that rapid thermal conductivity techniques are available Craig Dixon is the Applications
and provide relevant information to the polyurethane Specialist at Mathis Instruments
industry. They eliminate sample size requirements and Ltd. He holds a B.Sc. in
produce results in seconds rather than minutes. This brings Chemistry from the University
advantages for the R&D and QC areas. of New Brunswick, Canada. He
has focused on rapid transient
REFERENCES thermal conductivity testing of
polyurethane foam and VIPs
1. American Society for Testing Materials, Standard Test since 1997.
Method for Steady-State Heat Flux Measurements and
Thermal Transmission Properties by Means of the
Heat Flow Meter Apparatus, Designation: ASTM Nancy Mathis
C518-85, Philadelphia, PA, 1985. Nancy Mathis is the President
2. Mathis, N.E. ; Measurements of Thermal Conductivity of Mathis Instruments. She
Anisotropy in Polymer Materials, Ph.D. Thesis, Chem. holds a Ph.D. in Chemical
Eng. Dept., University of New Brunswick, Engineering from the University
Fredericton, N.B., Can., 1996. of New Brunswick, Canada.
3. Mathis, N.E., Chandler, C., “Orientation and Position She is the inventor and patent
Dependant Thermal Conductivity”, Proceedings of the holder for the technology
Society of the Plastics Industry’s Polyurethanes embodied by the TC Probe™
International Conference and Exposition, September thermal conductivity instrument
1999. Also accepted by Cellular Plastics, November distributed by PerkinElmer.
1999.
4. Cahill at University of Illinois at Urbana-Champaign.
(cf. Physical Review B 50 (1994) 6077 and J. Appl.
Phys. 81 (1997) 2590).

74 / Canney; Dixon, Mathis

Testing for MDI Emissions from Commercial Boards and
Polyurethane Glue

PREMAL P. PAREKH and BILL KAROLY

Huntsman Polyurethanes
286 Mantua Grove Road
West Deptford, New Jersey 08066

ABSTRACT

Huntsman Polyurethanes (HPU) conducted a series of aluminum and wood substrates cured with glue, MDI was
tests on aluminum and wood substrates cured with detected at 60 ppt during the first 8-hour sampling.
polyurethane glue and commercially produced composite Subsequent 8-hour sampling results were below the LOQ
boards cured with RUBINATE®*1*binder to evaluate MDI (<20 ppt).
emissions.
The studies were performed in HPU’s AIHA accredited INTRODUCTION
laboratory located in West Deptford, NJ.
Polyurethane glue was applied on wood boards or This study was conducted to determine the presence of
aluminum plates in different combinations (wood/wood, airborne concentration of MDI in finished commercial
wood/aluminum and aluminum/aluminum), sandwiched composite boards and during and after the application of
together and tested at two temperatures, ambient (21° C) PROBOND®*2*Polyurethane Glue to bond different
and 54° C. The wood boards and aluminum plates were substrates such as wood/wood, aluminum/aluminum and
tested for MDI emissions in a chamber that was flushed aluminum/wood.
continuously for 40 hours with air. During this time, the The glue was applied to two 6” x 5” x 0.25” aluminum
air in the headspace of the chamber was sampled plates, aluminum/wood plates and wood/wood plates
continuously for 40 hours, using Huntsman using a paint brush. After applying the glue, the plates
Sampling /Analytical Method I1024G revision 1.8. The were sandwiched together using a clamp. During this
filters through which the chamber air was drawn, were process of glue application and bonding, air sampling was
changed at every 8-hour interval for 40 hours. Five conducted for 30 minutes in accordance with the
consecutive 8-hour samples were taken under each test Huntsman method I1024G Rev 1.8(3). After 30 minutes,
condition. the test specimen was placed in a chamber at 21° C (room
Commercial composite board samples were cut to temperature), - 50% relative humidity (RH) air for
5" x 6" test specimen and placed inside the test chamber 40 hours, air was pulled through the chamber at a flow
rate o f - 1 L/min. The other test specimen was placed in
and tested at two temperatures, ambient (23° C) and at
the other chamber at 54° C, - 50% RH air for 40 hours.
90° C. Dry air (at room temperature, 23° C) or dry
Air was also pulled through the chamber at a flow rate of
nitrogen (at elevated temperature of 90° C) was flushed
- 1 L/min using a calibrated sampling pump (Gilian
continuously through the chamber for over 48 hours.
Sampler Model HFS 113A).
Both conditions were tested. During this time, the air in
The following finished commercial composite wood
the headspace o f the chamber was sampled continuously
boards were evaluated for MDI emissions:
for 48 hours, using Huntsman Sampling /Analytical
•particle board (100% RUBINATE binder)
Method I1024G. revision 1.8. The filters through which
the chamber air was drawn, were changed at every 8-hour • MDF board (100% RUBINATE binder)
interval for 48 hours. Six consecutive 8-hour samples •external soap released 100% RUBINATE binder
were taken under each test condition. Aspen board
The collected air samples were analyzed for MDI using
•Internal released 100% RUBINATE binder Pine
a High Performance Liquid Chromatograph (HPLC). The
limit o f quantification (LOQ) for the analytical method board
referenced above is 20 parts per trillion (0.2 pg/m3 of air •Siding OSB 100% RUBINATE binder Aspen board
sampled). No MDI emissions were detected at the LOQ • PF faced RUBINATE binder core Pine OSB board
(<20 ppt) for all composite boards tested. For the

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 75

 These boards were cut into 5" x 6" pieces. Air was 100% RUBINATE binder Aspen board, PF faced
drawn through the chamber at 23° C (room temperature) RUBINATE binder core Pine OSB board. These boards
for 48-hours and nitrogen at 90° C. The air was pulled were supplied by our customers.
through the chamber at a flow rate of - 1 L/min using a
calibrated sampling pump (Gilian Sampler Model HFS Derivatizing Agent
113A).
Quantitation o f MDI was achieved by pulling the Chemical Name: 1, (2-methoxy phenyl)
chamber air through a 13-mm glass fiber filter coated piperazine (1-2 MP)
with 1,2-methoyxy phenyl piperazine (1-2MP) Manufacturer: Fluka, Catalog # 65225
derivatizing agent and diethyl phthalate (DEP) as per Purity: 98% or better.
Huntsman method I1024G Rev. 1.8. The filters were
desorbed in dilute 1-2MP toluene solution and analyzed Emission Tests
for MDI using HPLC with FD and UV detectors
connected in series. The Glue tests were carried out at 21° C in air, - 50%
RH and at 54° C. in air,- 50% RH respectively. The
EXPERIMENTAL PROCEDURE composite wood boards were tested at 23 °C in air and at
90 °C in nitrogen. The test conditions were chosen to
Test Specimens cover two extreme temperature ranges. Prior to each test,
the test chambers were purged with air or nitrogen for at
A thin layer o f PROBOND® polyurethane glue was least a 24-hour period at both conditions. The
applied to four 6" x 5" x 0.25" aluminum plates using a atmospheres from the test chambers were monitored prior
paint brush. Then two plates were sandwiched together to to the test to ensure that the chambers were free of
prepare one test specimen. Two specimens were prepared residual MDI or other potential interferences. The test
for MDI emission testing at different conditions. Similarly specimens were placed in the test chambers and air or
wood/wood and wood/aluminum specimens were nitrogen was pulled through the test chambers and
prepared. through the two filters connected in series (front and back
Tlie finished commercial boards were also cut into filter assembly), at a sampling rate of - 1 L/min using a
sampling pump (Gilian Sampler Model HFS 113A).
5" x 6" pieces as test specimen.
The concentrations of MDI emitted from the specimens
were determined over a 40-48 hour period, with 5-6
Test Chambers consecutive 8-hour samples. Following each sampling
period, the filters were removed from their respective
Each test chamber consisted of a 2-piece silanized holders and desorbed immediately into separate vials. The
PYREX®^ glass container made by modifying a 12-L holders containing the filters were also rinsed and the
round bottom flask (bottom piece). The lid had several rinses were combined with the vials containing the
openings to allow for air flows, temperature respective filters to yield total MDI (filter + holder) value.
measurements, RH measurements, exhaust and sampling Upon completion of the entire 40-48 hour sampling
ports. During the testing, all unused openings were period of each test, the test specimens were removed from
closed except the one which was used for air changes. the chambers. The chambers were rinsed with a solution
Two test chambers were used simultaneously to conduct of 1-2 MP in toluene to remove and quantify any residual
the tests at different temperature conditions. These MDI that may have been present on the inside surface of
chambers were silanized with dimethyldichlorosilane the glass test chambers.
prior to each test to minimize surface reactivity of the
glassware with MDI. Sample Preparation/Collection/Analysis

Glue Prior to the start of each test, 13-mm glass fiber filters
were coated with - 2 mg of 1-2 MP and 1% DEP by
Commercial Product: PROBOND® Polyurethane Glue saturating each filter with - 0.1 mL of a solution of 1-2
Major Component: MDI, Manufacturer: Elmers Product MP and DEP in toluene (20 mg/mL of 1-2 MP). The
Inc. filters were allowed to dry in air. Each filter was loaded
into a Swinnex cassette. The cassettes were capped and
Commercial Boards placed in the refrigerator until used.
A total of two air samples (one per test specimen) were
Finished commercial boards tested were particle board taken during the application of the glue to the plates
(100% RUBINATE binder), MDF board within 6" from the application area. Air was drawn
(100% RUBINATE binder), external soap released through the filter paper housed in the Swinnex cassette
100% RUBINATE binder Aspen board, Internal released with a sampling pump at a flow rate o f - 1 LPM. The
100% RUBINATE binder Pine board, Siding OSB sampling durations were 30 minutes each.

76 / Parekh, Karoly
 Then the test specimens were loaded in the chambers, Pump A (Weak)
sampling trains (2 in total) were attached to the sampling
ports of chambers 1 and 2, respectively. The samples Acetonitrile 500 mL
were collected on two 1-2 MP coated filters loaded in two Water 500mL
separate cassettes connected in series to determine if there 1-Propanol 30 mL
was any breakthrough of MDI which would lead to Acetic Acid 1.1 mL
underestimation of MDI. The rear cassette was attached to Ammonium Acetate: 2.70 grams ( - 0.035 M)
the sampling pump and the front cassette was attached to
the exit line o f the test chamber. Pump B (Strong)
After sampling for 8 hours, the filters were removed and
desorbed as described above. The filter/toluene solution Acetonitrile 750 mL
were shaken for 1 hour, evaporated to dryness and Water 250mL
reconstituted in a mobile phase. An aliquot was 1-Propanol 30 mL
transferred to an autosampler vial for analysis by HPLC Acetic Acid 7 mL
using UV and FD detectors. Ammonium Acetate: 2.70 grams ( - 0.035 M)
An aliquot from each chamber rinse sample was also
transferred to an autosampler and analyzed by HPLC Injection Volume: 10 pL
using UV and FD detectors along with the filter samples Detector: FD Ex 246 nm: Em 367 nm; Gain 12
for each test. UV 242 nm; 2 AUFS
Data System: HP Chemstation v. 4.01
Standard Preparation/Detector Calibration
HPLC Gradient Conditions
Approximately 20 mg of 4,4' MDI (RUBINATE® 44
binder) standard was weighed and transferred into a 50.0 Time A B Flow
mL volumetric flask. The flask contents were diluted to (min) % % Rate (mL)
50.0 mL volume with dry toluene and mixed well. From
this flask 20 mL was transferred and stored in a vial fitted 0 75 25 0.5
with a PTFE lined septum. This vial was labeled as stock Isocratic first 4 min.
A (0.400 pg/pL). From stock A 5 mL was diluted to 25 4 75 25 0.5
mL in a volumetric flask with dry toluene, and mixed Isocratic first 4 min.
well. From this flask 20 mL was transferred and stored in 12 25 75 0.5
a vial fitted with a PTFE lined septum. This vial was Linear Gradient Portion
labeled as stock B (0.080 pg/pL). 13 25 75 0.5
Working standards over the range of 0.4 pg to 5 pg/vial Hold (1 minute)
were prepared by spiking appropriate aliquots from stock 13.5 75 25 0.5
A and B in a vial containing 1 mL o f dilute 1-2 MP Return to initial conditions
toluene solution plus 100 pL (2 mg of 1-2 MP + 1% DEP) 16 75 25 0.5
of concentrated 1-2 MP toluene solution. At least 3 Reequilabration
standards were prepared per each solution. The instrument
gave linear response to the 1-2 MP derivative of MDI Total Run Time: 16 minutes
over a concentration range of 0.1 pg to 10 pg for FD and
0.1 pg to 60 pg for UV, respectively. The desorption Filter Collection Validation
efficiency and check standard samples were also prepared
along with the standards. Average response factor ratios The filter collection method used for the determination
for MDI (MDI concentration/Peak height) were calculated of MDI in this study was evaluated to ensure that the MDI
and multiplied by the peak height responses generated derivative is stable under the test conditions. Two sets of
during sample analysis to determine MDI concentrations. 12 filters each coated with 1-2 MP and DEP were spiked
with known quantities of MDI. Using sampling pumps,
Instrument and Conditions for HPLC Analysis air was pulled through six coated filter cassettes from set
1 at 23° C for 8 hours at 1 L/min. Using sampling pumps,
HPLC System: HP 1050 LC System equipped with air was also pulled through the remaining six coated
Autosampler filters from set 1 at 90°C for 8 hour at 1 L/min. The set 2
Flow Rate: 0.5 mL/min filters were used as controls (non-exposed filters). Each
Column: Phenoshpere ODS-2, 3pm, 80A°, filter was then removed from the cassette and processed
150 mm x 3.2 mm i.d. vs the non-exposed filters. The collection efficiency of the
exposed cassettes (chamber air pulled through for
Mobile Phase: 8 hours) vs non-exposed was determined. The average
collection efficiency was - 91% (21°C) and -99% (54°C)

Parekh, Karoly / 77
indicating no loss of derivatization agent under the test The supporting data for each curve, in the form of a
conditions. response factors table, are found in Tables 1 and 2
respectively. The details for UV results are not listed
because all the samples were within the linearity range of
System Validation FD.

The system was validated to demonstrate that the

collection method for MDI would work at the test
conditions, if MDI were emitted from the bonded test
specimen. This was achieved by generating an MDI
atmosphere using pure 4,4' MDI in a PTFE dish placed in
the chamber and showing that the total MDI (collected on
the filter plus remaining in the dish) based on the mass
balance could be recovered after the test. The validation
recovery tests were conducted at 23°C in air and at 90°C
in nitrogen. The recoveries averaged to ~ >90% in both
the cases.

RESULTS AND DISCUSSIONS

Analysis of the samples collected during the application

of glue to the aluminum plates, wood/wood and Figure 1 Typical HPLC/FD Chromatograms o f 1) a
aluminum/wood detected no MDI, LOQ of <20 ppt. All MDIstandard at Concentration Level o f - 5 .0
chamber samples at room temperature condition showed pg, 2) a Blank Coated Filter Extract, and 3) a
no MDI at LOQ of <20 ppt. The first 8-hour chamber Sample Extract
sample at 54°C condition detected MDI at 60 ppt. The
subsequent four 8-hour samples did not detect any MDI at
LOQ of <20 ppt. See Table 1. The composite wood
specimens showed no MDI at <20 ppt (LOQ) for both
conditions. See Table 2. No breakthrough was observed in
the backup filter assembly.
Using the HPLC analytical instrument and conditions,
as described in this report, the derivatives of MDI were
resolved from the 1-2 MP, DEP and solvent front.
Chromatograms of a blank filter, a standard, and a sample
are shown in Figure 1.

Table 1. MDI levels for Glue Test from Wood/Wood, Wood/Alumlnum and Aluminum/Aluminum at 21°C and
54°C in air

Sampling Time/ 0 time* 8 hrs 16 hrs 24 hrs 32 hrs 40 hrs

Conditions
21°C, ~50%RH <20 ppt <20 ppt <20 ppt <20 ppt <20 ppt <20 ppt
54°C, ~50%RH <20 ppt 60 ppt <20 ppt <20 ppt <20 ppt <20 ppt

(ppt is parts per trillion)

*0 time = Sampling during the the glue application and bonding.

Table 2. MDI levels from all finished commercial composite boards at 21°C and 54°C

Sampling Time/ 8 hrs 16 hrs 24 hrs 32 hrs 40 hrs 48 hrs

Conditions
23°C in air <20 ppt <20 ppt <20 ppt <20 ppt <20 ppt <20 ppt
90°C in Nitrogen <20 ppt <20 ppt <20 ppt <20 ppt <20 ppt <20 ppt

78 / Parekh, Karoly
 (ppt is parts per trillion)

SUMMARY AND CONCLUSIONS

The PROBOND® polyurethane glue was applied with a BIOGRAPHIES

brush to bond two aluminum plates, wood/aluminum and
wood/wood together. MDI emissions were measured Premal P. Parekh
during application for a 30-minute period and after
application over a 40-hour period inside the chamber. Premal P Parekh is a Technical Associate in Industrial
Two chambers were used to conduct testing. Hygiene Department for Huntsman Polyurethanes. He
Finished commercial composite boards were also received his B.S. degree from St. Xavier’s College, India
evaluated for residual emissions of MDI. MDI was and M.S. (1988) degree in Chemistry from Stevens
monitored for a 48-hour period in the chamber. Institute of Technology, Hoboken, New Jersey. He
Air samples for MDI were collected on 13-mm 1-2 MP worked for Merck & Co in 1991. He worked for Rhone
derivatized filter. The filters were analyzed using HPLC Poulenc (1992) and Bristol Myers Squibb until 1994,
instrumentation equipped with ultra-violet (UV) and prior to joining ICI Polyurethanes, which later was
fluorescence detectors (FD) connected in series. Glued
acquired by Huntsman Corporation. Since joining ICI
substrates showed insignificant airborne levels of MDI
Polyurethanes, his
and wood composite boards showed no emissions above
responsibilities are supervising
the LOQ (<20 ppt).
and directing the IH lab to
REFERENCES handle product stewardship
samples. He is a member of the
1. RUBINATE® is a registered trademark of Huntsman American Industrial Hygiene
Association.
International LLC. The mark is registered in the
United States and elsewhere, but may not be
registered in all countries.
William J. Karoly
2. PROBOND® is a registered trademark of
Prochemtechnology, Inc. Bill Karoly is Manager of Industrial Hygiene at Huntsman
3. Method for the Collection, Storage and
Polyurethanes, West Deptford, NJ. He received a B.S.
Determination of Airborne
degree in Chemistry from John Carroll University,
4,4' -Methylenediphenyldiisocyanate
Cleveland, Ohio and a Master of Science in Physical
(MDI, and Oligomers as Vapors and Particulates
Chemistry from Duquesne University, Pittsburgh, PA. He
(Huntsman Method I1024G Rev 1.8).
also received his Master of Public Health in Industrial
4. PYREX® is a registered trademark of Coming
Hygiene from University of
Incorporated. Minnesota in Minneapolis. Bill has
over 25 years of practical
experience in Industrial Hygiene
and is certified by the American
Board of Industrial Hygiene in the
field of Comprehensive Practice.

Parekh, Karoly / 79
 Report on the First ASTM Round Robin of Vacuum Panels

ANDRZEJ BRZEZINSKI THERESE STOVALL

LaserComp Oak Ridge National Laboratory
20 Spring Street 1 Bear Creek Road
Sangus, Massachusetts 01906 Oak Ridge, Tennessee 37831

ABSTRACT INTRODUCTION

A round robin was initiated in February of2000 to compare Vacuum insulation systems have long been used for
different methods of determining the effective thermal cryogenic applications. These systems have historically
resistance of vacuum panels. The outcome of this round consisted of multi-layer evacuated jackets with
robin will provide support for the ASTM material activevacuumsystems. In the early 1990s, sealed evacuated
specification and the development of a future ASTM test panels became commercially available. These panels were
method. Four issues were identified and addressed: (1) filled with either fiberglass or silica and had either metal or
calorimetric vs center-of-panel/barrier conductivity plastic barriers. The continuing design evolution includes
approaches, (2) comparison of available finite open-celled foam and advanced powdered fillers, specialty
difference/element models, (3) appropriate boundary multi-layer films, and the inclusion of new adsorbent
conditions for all measurements/models, and (4) comparison systems. In order to help potential users understand the
of center-of-panel measurements. Six conventional vacuum performance of these panels, a task group was formed in
panels were constructed. All six shared the same 1995 to create an ASTM material specification [1]. Due to
dimensional configuration, the same core material, the same the complexity of this non-homogenous insulation form,
getter insert, and the same manufacturing techniques and several evaluation methods were developed by researchers
equipment. Two different barrier materials (three panels and panel manufacturers. The task group initiated efforts to
from each) were used because barrier thermal conductivity systematically compare the results of these differing
is recognized as a key factor in the determination ofeffective approaches.
thermal resistance for vacuum panels, and because the
different methods used in this round robin comparison The resulting round robin was initiated in February of
should be sensitive to the barrier thermal properties. The 2000, with the goal of comparing different methods of
getters were included in these panels to help them remain determining the effective thermal resistance of vacuum
stable throughout the duration of the round-robin. panels. The outcome of this round robin will provide
Each of the eight participating laboratories measured the support for the ASTM material specification and the
center-of-panel resistance of each of the six panels as development of a future ASTM test method. Four issues
described in the ASTM standard C1484-00 and reported were identified and addressed: (1) calorimetric vs center-of-
those results along with pertinent information about the panel/barrier conductivity approaches, (2) comparison of
transducers) size and location. Several laboratories also available finite difference/element models, (3) appropriate
determined the whole-panel effective thermal resistance, boundary conditions for all measurements/models, and (4)
using two assumed sets of boundary conditions. comparison of center-of-panel measurements.
Considering the exploratory nature ofthis round-robin, and
the complexity of the measurements, the results showed ROUND ROBIN DESIGN
surprisingly good agreement.
Six conventional vacuum panels were constructed by Dow
in January, 2000. All six shared the same dimensional

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 81

configuration, the same core material, the same getter insert, temperature plate. For others, an arrangement where the
and the same manufacturing techniques and equipment. fiberglass batt also covers the bottom and top of the
The specimens were each 12" x 12" by 1", and each was vacuum panel was used. When the fiberglass blanket is
clearly and uniquely marked and evacuated to the same thus inserted between the constant temperature plates and
pressure. Two different barrier materials (three panels from the vacuum panel, thermocouples are attached directly to
each) were used because barrier thermal conductivity is the center of the vacuum panel to record the temperature at
recognized as a key factor in the determination of effective that location. This last arrangement is typically used with
thermal resistance for vacuum panels, and because the an array ofheat flux transducers and is directed more toward
different methods used in this round robin comparison measurement of whole panel performance, because it allows
should be sensitive to the barrier thermal properties. a temperature gradient to develop along the face of the
Aggressive getters were included in these panels to help barrier material. Despite this limitation, center-of-panel
them remain stable throughout the duration of the round- resistivity measurements were also made using this
robin. arrangement.
The seven participating laboratories were Advantek, Dow,
Dupont, Holometrix, LaserComp, Oak Ridge National RESULTS
Laboratory, and the Product Design Center. Each laboratory
measured the center-of-panel resistance of each of the six The measured thermal resistivity values are summarized in
panels as described in ASTM C l484-00 and reported those Fig. 1. For the panels with a more conductive barrier, the
results. Several of the laboratories made multiple resistivities are all between 26.4 and 31.4 h#ft2-°F/Btu*in. or
measurements using different types of apparatus. Pertinent between 92.3 to 109.7% of the average value of 28.5
information about the test equipment used to date in this h«ft2*°F/Btu»in. The standard deviation for these ten
round robin is shown in Table 1. Several laboratories also measurements is 5.1%. For the panels with the less
determined the whole-panel effective thermal resistance, conductive barrier, the resistivities are all between 27.6 and
using two assumed sets of boundary conditions. 32.6, for a range of 92.9 to 109.8% of the average value of
29.5 h*ft2»°F/Btu#in. The standard deviation for the ten
Tablet. Heat flux meter parameters measurements on the less conductive barrier panels is 5.5%.
Previous modeling work on vacuum panels has shown
Plate size Central transducer size that the center-of-panel measurement will be more accurate
for smaller transducer sizes [1]. This is most important if the
24" x 24" 3" x 3"
barrier is more conductive, and the results for this round
24" x 24" 4" x 4" robin show that effect, as seen in Fig. 2. In this figure, the
effect of heat transfer through the panel edges becomes
24" x 24" 4" x 8" more important as the transducer size approaches the panel
size. Indeed, the value measured by a 10 x 10 in. transducer
24" x 24" 10" x 10"
is almost the same as the whole panel effective thermal
12" x 12" 4"x4" resistance calculated with a finite difference model of the
panel (that model used typical heat flux meter apparatus
12" x 12" 3" x 3" boundary conditions).
The center-of-panel thermal resistivity measurements are
12" x 12" 3" diameter
also highly dependent on the measured thickness of the
panel. For the test configurations, the thickness is
An examination of Table 1 shows that some of the test automatically measured by the test apparatus. For other test
devices were the same size as the vacuum panels, 12" x 12". configurations, especially those that employ a fiberglass
In these devices, the entire surface ofthe vacuum panel was blanket above and below the panel, independent
in direct contact with the controlled temperature plate. This measurements are requires. When such measurements are
configuration therefore represents a constant temperature made, they are typically the average of eight locations over
boundary condition, where the temperature gradient from the surface of each panel. A summary of the measured
the center of the panel to the edge ofthe panel is minimized panel thicknesses is shown in Fig. 3. The data show a
and the edge heat transfer through the boundary material is variation from ­6 to +11% relative to the nominal value of 1
reduced. Other test devices were twice as wide as the in.Because vacuum insulation panels are non-homogenous,
vacuum insulation panels. For these tests, a high-density various approaches have been developed to determine their
fiberglass blanket was sculpted to fit tightly around the overall thermal effectiveness. One method employs an
vacuum panels and to match the area of the test apparatus overall hot box technique where mathematical models are
plate size. For some of these large plate tests, the vacuum used to correct for the effects of materials used to surround
panel was still in direct contact with the constant the test panel. That method has not yet been tested with

82 / Brzezinski, Stovall
 Figure 1. Measured center-of-pane! thermal resistivity

the round robin specimens. analysis. That latter method was used by three of the
The other method in common use employs a finite participating laboratories and the results are shown in Table
difference model of the panel, and requires a priori 2.
knowledge of the barrier thermal conductivity, the thermal Four boundary conditions were considered. The first
conductivity of the evacuated region within the panel, and represents the typical wall or door of a refrigerator. For this
some definition of the thermal boundary conditions for the configuration, one side of the panel would face a thin sheet
of steel (0.0006 m thick, 69.23 W/m-K) which is in turn
exposed to indoor convective transfer to an environment at

Figure 2. Measured center-of-panel thermal resistivity

varies with transducer size fo r more conductive barrier
materialf comparison to calculated effective whole panel Figure 3. Comparison o f measured panel thickness,
therm al resistance. norm alized relative to nom inal 1 in. thickness .

Brzezinski, Stovall / 83
 Table 2. Finite Difference/Element Model Results for Whole Panel Thermal Resistance

Model used Boundary conditions Whole panel thermal resistance (h*ft2*°F/Btu)

barrier: more conductive barrier: less conductive
LabB Refrigerator door 10 26
Lab D Refrigerator door 10 23
LabB Wall 12 32
Lab D Wall 10 27
Lab D Test apparatus 11 27
Lab C Undefined 13 27

21C. The other side of the panel would be surrounded by the program and measurements will be completed using their
2.54 cm. of foam (0.024 W/m-K) and a thin sheet of ABS apparatus and methodologies as soon as possible.
(0.003048 m thick, 0.2596 W/m-K) exposed to an air One of the most important future efforts will be the
temperature of 4C. The second boundary condition addition ofcalorimetric measurements at a facility dedicated
represents a wall section of a building. In that wall, one side to such work. This will provide an invaluable benchmark for
of the panel would face 1.27 cm. gypsum board (0.16 W/m- the finite element modeling efforts.
K) exposed to indoor convective conditions (21C) . The Another important issue to be addressed is that of the low
other side would face 1.27 cm of foam (0.03 W/m-K)), heat flux calibration. Many of these test devices are
followed by a thin cladding (wood, 2.5 cm. thick, 0.19 W/m- typically used to measure insulation with lower resistivity,
K) exposed to external convection at -1C. The third so the heat flux is usually much greater than that measured
boundary condition is a bit simpler, because it represents a during these tests. Special calibration procedures have
heat flux meter apparatus with standard high-density been developed and reported by some participants, and
fiberglass surrounding the panel which is in turn encased efforts will be made to determine procedures used by others.
within two constant temperature plates. The fourth The ultimate goal ofthe round robin is to determine which
boundary condition was not known by the laboratory, test methods give the most useful results, and to provide a
because they employed a canned program provided to them definition of the expected accuracy. The information
by others. gathered during this effort will lead us to improved industry
Considering the different mathematical models, the consensus standard test methods for vacuum insulation
different values used for the element conductivities, and the panels.
different boundary condition implementations, there is a
surprising degree of agreement. The whole panel thermal REFERENCES
resistance for the more conductive barrier was 11.2
h*ft2*°F/Btu with a standard deviation of 11%. The whole 1. ASTM C l484-00, Standard Specification for Vacuum
panel thermal resistance for the less conductive barrierwas Insulation Panels, Annual book ofASTM Standards, Vol.
26.9 h*ft2*°F/Btu with a standard deviation of 12%. 04.06, West Conshohocken, PA.

FUTURE PLANS BIOGRAPHIES

Andrzej Brzezinski
Considering the exploratory nature of this round-robin,
and the complexity ofthe measurements, the results showed
Andrzej Brzezinski, LaserComp’s President and Product
surprisingly good agreement. However, there is still much
Development Director, has been in the thermal conductivity
to learn. Future plans to expand the round robin include the
business in a variety of capacities for more than 20 years.
addition of more measurements and more laboratories.
After graduating from the University of Illinois in 1977 with
Specifically, a standard foam board has been added to the
a degree in Material Science, he worked at Foxborough
round robin to allow us to compare the baseline performance
Corporation before taking a position at Dynatech as
ofall the included test apparatus devices. This board will be
Manager in their Testing Laboratory. He then ventured into
circulated to all the participating laboratories. There are also
Sales, Research and Development (he developed a
several other laboratories that have asked to be included in

84 / Brzezinski, Stovall
LaserFlash System still being sold) and Management at the insulation programs. She updated the DOE Insulation Fact
highest levels. When Mr. Brzezinski created LaserComp he Sheet, the second most requested document published by
utilized his experience in running a testing lab and then sales DOE, and has contributed to laboratory evaluations of a
to pin point the needs of the marketplace. Years in design variety of insulation materials, focusing her efforts on
and development provided the technical background to evacuated insulation panels. Previously, Therese managed
determine the feasibility ofwish lists for instruments. All of the Ice Storage Test Facility, a commercial ice storage
this resulted in the creation ofthe FOX Instrument Line, the testing program sponsored by the Electric Power Research
finest Thermal Conductivity Measurement Instruments in Institute. She received a patent for a load management
the world method using wallboard containing a phase change material.
Therese Stovall received her B.S.M.E. with distinction
Therese K. Stovall from Purdue University in 1977, her M.S.M.E. from the
University of Tennessee in 1988 and is a Licensed
Therese K. Stovall has worked at Oak Ridge National Professional Engineer in the state of Tennessee. She is also
Laboratory since 1978. She is currently the manager of the a member ofthe American Society ofMechanical Engineers,
Engineering Science Section in the Engineering Technology the American Society for Testing and Materials, and the
Division. This section performs work in thermal hydraulics, American Society of Heating Refrigerating and Air
structural mechanics, and irradiation engineering. Conditioning Engineers.
Previously, Therese worked on distributed generation and

Brzezinski, Stovall / 85
 SESSION D
CONSTRUCTION

TECHNICAL SESSION ORGANIZERS

George Gula, Chair

CANNON USA

William Nicola
BAYER CORPORATION

A1 DeLeon
OXID, INC.

TECHNICAL SESSION MODERATOR

M ichel Drouin
JOHNS MANVILLE

TECHNICAL SESSION VICE-MODERATOR

Paul N eill
STEPAN COMPANY
The Development of All Water-Blown Polyisocyanurate Foam
System for Metal-Faced Continuous Sandwich Panels

AKIRA NARUSE, HISASHI NANNO, MITSUHIRO KURITA, HIDEKI INOHARA

and TAKAO FUKAMI
Nippon Polyurethane Industry Co., Ltd.
Central Research Laboratory
440 Akiba-cho, Tostuka-ku, Yokohama-shi
Kanagawa 2450052
Japan

ABSTRACT INTRODUCTION

Metal-faced sandwich panels have been extensively used Rigid polyisocyanurate (PIR) foam as well as rigid
as exteriorwalls in thebuilding industry. Experience andtesting polyurethane (PUR) foam has been widely used as an
have demonstrated that HCFC-141b blown polyisocyanurate insulation m aterial for construction industry. The
(PIR) foams provide acceptable fire performance when demand for PIR foam has increased, due in part, to its
incorporated into these types of panel systems. inherently better fire performance and thermal stability
Nowadays, the requirement for the development of all characteristics when compared to PUR foams. Whereas
water-blown PIR foam system is getting higher, because HCFC-141b used as a blowing agent in PIR foams must
HCFC-141b is to be phased out and water-blown system is be phased out by the end of 2003 in Japan. Therefore
environmental-friendly, safe and economical. alternative blowing agents such as HFC-245fa,
However, in general, all water-blown PIR foam system HFC-365mfc, pentanes and water have been proposed [1],
has some problems such as poor flowability, foam [2], [3].
friability and poor adhesion strength to metal and paper. However these alternative blowing agents have some
Moreover, the poor storage stability of polyol premix of problems for construction applications. HFC-245fa and
PIR system due to hydrolysis is a serious issue, since HFC-365mfc are zero ozone depletion potential (ODP),
polyester polyol modified PIR foams are generally used to but HFC-245fa has higher vapor pressure than
enhance incombustion properties. HCFC-141b and HFC-365mfc is flammable. Furthermore
As for storage stability, the effect of various polyester they have relatively higher global warming potential
components and catalysts for PIR foam preparation were (GWP) and higher price is expected. Pentanes are mainly
examined as the factors of the increase of acid value and used in the refrigerator market in Japan, but for
the change of reaction profile. Through these tests, it was construction, especially for spray application, it is
found that the specific combination of polyester considered to be difficult to use pentanes because of their
components and selected catalysts led to good storage flammability. So, in Japan, as an environmentally-friendly
stability of polyol premix. Furthermore, to improve foam system, all water-blown foams have been mainly studied
properties, new polyisocyanate was developed and finally [4]-
we developed new all water-blown PIR foam system Recently, all water-blown PIR foam systems having
which showed similar storage stability to current excellent fire resistance property have been reported [5],
HCFC-141b system, high flowability, less-void [6]. However, all water-blown PIR foams have not been
appearance, good dimensional stability, good adhesion actually accepted yet. Because the use of water as blowing
strength and excellent incombustibility. agent for PIR foam systems has several problems. They
In this paper, the relationship between each component are such as poor dimensional stability, poor surface
of PIR foam and various properties will be discussed. friability, and poor processability. Poor surface friability
will lead to poor adhesion strength to facer materials and
poor processability will cause voids.
Besides the above problems, the poor storage stability

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 89

 Table 1. Typical formulation
Raw m aterial Parts by weight
Phthalic Acid Based Arom atic Polyester Polyol 80
Mannich Condensate -Based Polyether Polyol 20
TCPP (Flame Retardant) 5% /Total
Silicone Surfactant 0.6% /Total
M ixture o f Potassium Acetate,
Q uaternary Ammonium Salts and Tertiary Amine GT =55sec
W ater 5
Polymeric MDI (Isocyanate) lndex =275

of polyol premix of PIR foam system is a serious issue followed by was cut to the size according to the standard
since polyester polyol used in PIR formulation is easily described below.
hydrolyzed.
To solve these problems in all water blown PIR foam Test Method
system, we have investigated the storage stability of polyol
premix and the relationship between system components RISE PROFILE
and foam properties.
Through these studies, we have developed new all water The profile of free rise foaming was measured by
blown PIR foam system, which is especially suitable for using Time Tech DAPS Model QA System.
metal-faced continuous sandwich panels.
FOAM PHYSICAL PROPERTIES
EXPERIMENTAL
Various physical properties were measured in
Formulation and Raw Materials accordance with the following standards.

The typical formulation used in this study is shown in Foam Density JIS K-7222
Table 1. Phthalic acid based aromatic polyester polyol Compressive Strength JISA -9511
and Mannich condensate-based polyether polyol were Dimensional Stability ASTM D-2126
used to enhanced thermal properties. As additives, tris Incombustibility JIS A-1321 Grade 2 Test
j3 -chloropropyl phosphate (TCPP) as a flame retardant Cone Calorie Meter Test ISO -5660
and silicone surfactant were used. Tertiary amine, Thermal Conductivity JIS A -1412
quaternary ammonium salts, and organic potassium salts (Anacon model TCA8)
were used as catalysts. An amount of catalyst was
adjusted to obtain the gel time of 55 seconds. As an ADHESION STRENGTH
isocyanate, Millionate MR-200 which is polymeric MDI
and supplied by Nippon Polyurethane Industry Co., Ltd. The foam for adhesion test was prepared by using the
was used. aluminum open top mold with 400 (length) X 400 (width)
X35mm (thickness). The mold temperature was kept at
Foam Preparation 60 °C . The mixture (ca.300g) of polyol premix and
isocyanate was poured into the center of the mold where
All foams were prepared by using standard hand metal facer (thickness of 0.35mm) was preset. Then the
mixing method. Each temperature of polyol premix and mold was closed with upper aluminum plate. The foam
isocyanate was kept at 20°C. with metal face was demolded after 5 minutes. The sample
For the measurement of foaming reactivity and free was conditioned for 24 hours at room temperature,
rise density, the mixture (ca.550g) of polyol premix and followed by was cut to the size of 100 X 100 X 35mm.
isocyanate was poured into the aluminum open box mold Tensile tester (IMADA model DPS-50R) was used as
of 250 X 250 X 250mm which was adjusted to 40°C and spring balance. Metal facer was holed with diameter of
then its reactivity was measured. 7mm and the hook of tensile tester was hooked to the hole
For the preparation of foam sample for physical and pulled. Then maximum adhesion strength was
properties, the mixture (ca.750g) of polyol premix and measured.
isocyanate was poured into the aluminum open top mold
of 500 (width) X 60 (breadth) X 500mm (height) which FLOWABILITY
was adjusted to 60°C. After demolding, the obtained
foam was conditioned for 72 hours at room temperature, The foam for flowability test was prepared by using

90 / Naruse, Nanno, Kurita, Inohara, Fukami

 EFFECT OF CATALYST

Under the combination of catalysts, polyester polyol

and water in polyol premix, the polyester polyol is easily
hydrolyzed. It appears as an increase of acid value in
polyol premix and as reaction delay.
The test was carried out by using typical PIR
formulation. Sample were stored for 30 days at room
temperature. As shown in Figure 1, there was a
significant difference in the delay of gel time. Organic
potassium salts led to the best storage stability of polyol
premix in this study, followed by quaternary ammonium
Figure 1. The effect o f catalyst type on storage stability
o fpolyol prem ixture
and tertiary amines. As tertiary amine is generally
necessary to control foaming profile, we have to pay
careful attention for the choice of amine catalysts.

the aluminum open top mold of 500 (length) X 250 EFFECT OF AROMATIC POLYESTER POLYOL
(width) X 15mm (height). The mold temperature was kept
at 60°C. The mixture (ca.80g) of polyol premix and The effect of the components of phthalic based aromatic
isocyanate was poured into the edge of the mold. Then polyester polyol on hydrolysis was examined. Sample
the mold was closed with upper aluminum plate. After were stored for 30 days at room temperature then the
demolding the flow length was measured. change of acid value was measured accordance with JIS
K-1557. These results were shown in Figure 2, 3, 4, 5, and
RESULTS AND DISCUSSION 6. Figure 2 and 3 show the effect of acid component. As
shown in the both figures, tere-phthalic acid based
Storage Stability of Polyol Premix polyester polyol showed poor storage stability. On the

F ig u r e 2 . T h e e ffe c t o f a c id c o m p o n e n t in p o ly e s te r Figure 3. The effect o f acid component in polyester

p o ly o l o n g e l tim e polyol on acid value

Figure 5. The effect o f glycol component in polyester

Figure 4. The effect o f glycol component in polyester polyol on acid value
polyol on gel time

Naruse, Nanno, Kurita, Inohara* Fukami / 91

 of ester group content and steric hindrance should be
considered. DEG based polyester polyol has more ester
groups than TEG or DPG based polyester polyol and does
not have bulky substituent like DPG. On the other hand, in
the case of DPG based polyester polyol, it can be said the
hydrolysis is hard to occur because the ester group is
relatively protected with bulky methyl substituent.
All data were plotted in Figure 6. As shown in Figure 6,
in the initial stage of hydrolysis, the delay of gel time was
in proportion to the increase of acid value. But from over
around 1.5 mgKOH/g acid value, a considerable delay of
gel time was observed with a small increase of acid value.
Figure 6. A Geling Time v.s. A Acid Value
This phenomenon will relate to the reduced amount of free
tertiary amine.

The Effect of Foam Components on Physical

other hand, ortho-phthalic acid based polyester polyol Properties
showed relatively good stability. As a remarkable point,
the storage stability of ortho/tere = 50/50 based polyester EFFECT OF AROMATIC POLYESTER POLYOL
polyol was similar to that of ortho-phthalic acid based
polyester polyol. Table 2 shows foam physical properties and
Figure 4 and 5 show the effect of glycol component of flammability as a function of tere -phthalic acid content in
aromatic polyester polyol. As a glycol, diethylene glycol aromatic polyester polyol. Significant effect of acid
(DEG), triethylene glycol (TEG), and dipropylene glycol component was not seen in physical properties. But there
(DPG) were tested. DEG based polyester polyol was much was a tendency in flammability. With an increase of
sensitive to hydrolysis. As the reason why DEG based tere-phthalic acid component, smoke density and weight
polyester polyol showed poor storage stability, the effects retention were getting better.

Table 2. Effect o f tere-phthalic acid content in polyester potyol on foam properties

Yere -phthalic acid content h polyester polyol — ITS—
------------------ 33------------------------- 523---------- --------- 78-------------
Dimensional Stability (vol. %)
150 “C x48h / 7days - 3.6/ - S.3 - 4 .1 / - 4 .7 - 3 .6 / - 4 A - 3 .3 / 4 .7
9 0 °Cx48h / 7days - 1.4/ - 1.5 - 2.0 / - 1 .7 - 1 .4 / - 1 .7 - 1.5 / 1 .2
70 ° C95% RHx48h / 7days - 0 .1/ - 0.2 - 0 .6 / - 0 .9 0.0 / - 0 .9 0.2/ 0.1
- 20 °Cx48h / 7days 0 .4/0.0 0.4/ - 0.4 0.3 / - 0 .5 0.1 / 0 .2
Compression Strength (MPa)
Parallel 0.394 0.368 0.400 0.404
Perpendicular 0.204 0.200 0.201 0.211
Thickness 0 .2 09 0.198 0.190 0.2 09
Flammability (JIS A - 1321 Grade 2 Test)
Smoke Density (CA) 57 62 45 44
Temperature Time A rea ( °C -min) 14 3 71 14
Flaming Time (sec) 37 43 37 34
Weight Retention (%) 64.0 64.0 69.0 69.0

Table 3. Effect o f polyether polyol on foam properties

Polyether Polyol A Polyether Polyol B Polyether Polyol C
None
O H V -2 0 0 O H V -3 5 0 O H V = 470
Dimensional Stability (vol.%)
150 ° Cx48h - 32.1 - 22.1 - 20.3 17.3
9 0 °Cx48h -14.4 -4.8 -4 .6 -4.8
70 ° C95%RHx48h - 16.4 - 5 .7 - 5 .2 5.4
-20 °Cx48h -3 .7 -0.4 -0.1 -0 .7
Compressbn Strength (MPa)
Parallel 0.333 0.339 0.3 25 0.332
Perpendicular 0.138 0.128 0.1 19 0.121
Thickness 0.116 0.1 17 0.115 0.122
Adhesion Strength (kgfflOcm)
To metal facer after 24h 3.1 6 .2 6 .0 4 .2
Fbwability (mm/80g) 394 37 9 372 35 5

92 / Naruse, Nanno, Kurita, Inohara, Fukami

 Table 4. Effect o f polymeric isocyanate on foam physical properties *
Isocyanate MR-200 C-1350 C-1400
Viscosity (mPa s 25 °C) 200 300 450
NCO contents (%) 31.1 30.9 30.4
Functbnality 2.7 2.8 2.9
Dimensional Stability (vol.%)
150 K x48h -5.5 -3.3 -1.9
90 V x48h -3.9 1.9 -1.0
70 V9S%RHx4Bh -1.6 -1.2 -0.6
-20 V x48h 0.3 0.2 0.2
Fbw ability (mm/80g) 376 369 351
Adhesion strength (kgf/10cm) 6.0 4.2 2.1
Compression Strength (MPa)
Parallel 0.341 0.359 0.371
Perpendicular 0.229 0.254 0.266
Thickness 0.255 0.261 0.270
*: These isocyanate are polymeric MDI and sup/ied b y Nippon Polyurethane Industry Co., Ltd.

Table 5. Results o f cone calorie meter test

Isocyanate Index
200 225 250 275 300
Peak Heat Release Rate (kW/m2) 109.4 99.9 94.1 86.8 87.3
Average Heat Release Rate (kW/m2) 31.4 31.1 25.5 22.7 30.0
Total Heat Release (MJ/m2) 18.8 18.6 15.3 13.6 17.9
Average Effective HOC (MJ/kg) 23.6 23.3 20.9 19.0 22.6
Average Specific Ext Area (m2/kg) 470.4 389.9 262.2 235.9 351.5
Average Mass Loss Rate (g/s -m2) 1.66 1.55 1.34 1.30 1.44
Weight Retention (%) 14.3 13.9 21.7 23.0 14.9

Table 6. Results o f physical properties

Isocyanate Index
175 200 225 250 275 300
F o am D ensity (kg/mS) 3 9 .8 3 8 .7 3 8 .6 3 9 .2 4 0 .4 4 0 .0
Dimensional Stability 15 0 V x 4 8 h -45.1 -2 9 .4 -1 2 .4 -4 .0 -1 .0 6 .9
(vol.% ) 9 0 'C x 4 8 h -1 3 .5 6 .2 6 .1 -1 .2 6 .8 6 .7
70 ° C 95% R H x48h - 1 7 .5 6 .0 2 .3 6 .5 1.7 3 .9
- 2 0 ° C x48h 0 .2 0 .3 0 .9 0 .6 0 .4 0 .2
Compression Strength Parallel 0 .3 7 0 0 .3 6 9 0 .3 4 5 0.351 0 .3 5 4 0 .3 6 5
(M P a) Perpendicular 0 .1 3 7 0 .1 3 6 0 .1 4 5 0 .1 4 5 0 .1 4 3 0 .1 4 9
Thickness 0 .1 2 8 0 .1 3 3 0 .1 3 6 0 .1 4 0.151 0 .1 5 5

EFFECT OF MANNICH CONDENSATE-BASED EFFECT OF ISOCYANATE

POLYETHER POLYOL
The effect of various isocyanates on physical
The effect of the type of Mannich condensate-based properties was investigated. Three types of polymeric
polyether polyol (Mannich polyol) on foam physical MDI were examined as shown in Table 4. MR -200 is a
properties is summarized in Table 3. Three types of the standard polymeric MDI. C -1350 and C-1400 have
Mannich polyol having different OH values were higher functionality compared with MR-200. These
examined. By adding the Mannich polyol, dimensional polymeric MDI are supplied by Nippon Polyurethane
stability and adhesion strength were improved and the Industry Co., Ltd.
Mannich polyol having low OH value of three polyols Polymeric MDI having higher functionality showed
led to better adhesion strength and flowability. higher compression strength, better dimensional stability
and poorer adhesion strength.

Naruse, Nanno, Kurita, Inohara, Fukami / 93

 Table 7. Cone calorie meter test conditions Novel All W ater Blown PIR Foam System
Foam Sample Core foam : Density 37kg/m3
Test Time 10min NEW PIR SYSTEM
Heat Flux 50kw/m2
Based on the fundamental studies, new all water blown
PIR foam system DRC-552 / DRN-2343 has been
developed. DRC-552 having NCO content 30.4%,
viscosity of 230 mPa • s 25°C is a special polymeric MDI
EFFECT OF ISOCYANATE INDEX developed for new PIR foam system.

The effect of isocyanate index on physical properties STRAGE STABILITY OF NEW PIR SYSTEM
was investigated. The results of cone calorie meter test
and physical properties as a function of isocyanate index The storage stability of the new all water blown polyol
are shown in Table 5 and 6, respectively. The test premix, DRN-2343, is shown in Figure 7. The polyol
conditions of cone calorie meter test were shown in Table premix was aged under two conditions, 20°Cand 40°C.
7. Among evaluation items listed in Table 5, total heat At 40°C, the delays of gel time and rise time were seen in
release (THR) and peak heat release rate (PHRR) are the
the case of the storage at 40°C. However, the change of
most important factors. From these data, THR and PHRR
reaction time after 30 days was within 10% against that
decreased as isocyanate index increased till at the point
at 20°C. This storage level is similar to that of current
of isocyanate index of 275. But isocyanate index of 300
showed poorer result compared with isocyanate index of HCFC -141b blown system.
275. It will be influenced by less reaction ratio in the
range of high isocyanate index.

Table 8. Physical properties o f novel all w ater blown P IR foam system

N ovel all w ater blown system Current H C F C - 1 4 1 b blown system
F re e Foaming Reactivity
(sec) C ream Time 7 7
G e l Time 38 35
T ack F re e Time 47 47
Rise Tim e 80 68
F re e Rise Density (kg/m3) 3 2 .8 3 1 .3
Test P an el F oam
F oam Density (kg/m3) 4 2 .6 42.1
Dimensional Stability 150 ° Cx48h - 2.1 1 .6
(vol.%) 9 0 ° Cx48h - 1 .5 1 .0
70 ' C 9 5 % R H x 4 S h 0 .5 1.2
-2 0 V x 4 8 h 0 .2 -0 .2
Compression Strength Parallel 0 .3 6 8 0 .3 0 3
(M P a) Perpendicular 0 .2 5 5 0 .1 8 2
Thickness 0 .2 7 2 0 .1 9 6
Adhesion Strength Initial good good
(kgf/10cm ) A g e d for 2 4 hours 7.8 7.1
Therm al Conductivity Initial 2 0 .9 17.6
(m W /K) A g e d for 2 7 0 days 21.1 2 0 .5
Flowability (mm /80g) 376 377
Closed Cell C ontent (%) 93 92
Cone Calorie M e te r Test
P e a k H e a t R elease R ate (kW /m 2) 8 7 .5 121 .2
Total H e a t R elease (M J/m 2) 14.2 18.6
A verag e Effective H O C (M J/kg) 19.3 2 1 .3
A verag e Specific E x t A rea (m2/kg) 2 3 4 .5 2 1 5 .3
A verag e M ass Loss R a te (g/s - m 2) 1.28 1.69
W eight R etention(% ) 2 3 .4 16.2
JIS A - 1 32 1 G rad e 2 Test
Sm oke Density (CA) 42 38
Tem perature Time A re a ( ° C 'm in) 0 10
Flaming Time (sec) 22 20
W eight Retention (% ) 6 8 .7 6 3 .6

94 / Naruse, Nanno, Kurita, Inohara, Fukami

 FOAM PROPERTIES OF NEW PIR SYSTEM

Foam preparation was carried out by using

MARUKA -KAKOUKI high pressure foaming machine
MEG-HK270-B type with MQ type mixing head.
The properties of new all water blown PIR foam
system are summarized in Table 8. New all water blown
PIR foam system showed sufficient physical properties,
adhesion strength, and flowability compared with current
HCFC-141b blown PIR system. Especially the new
system showed excellent flammability and appearance
F ig u re 7. S to ra g e s ta b ility o f n o v e l a ll w a te r b lo w n P IR (less void) compared with HCFC-141b blown system.
s ys te m Figure 8 shows the foaming profile of the new system
and HCFC-141b blown system. The new system showed
almost the same smooth rising profile as well as
HCFC-141b blown system. Figure 9 shows the
flowability of the new system and HCFC-141b blown
system. The new system showed the same flowability as
HCFC-141b blown system. It is obvious that the novel
all water blown PIR foam system showed almost the
same processability as current HCFC-141b blown
system.
Figure 10 shows changes of thermal conductivity of
the new system and HCFC -141b blown system with and
without metal facer. Foam samples were kept for 9
months at 25°C. Without metal facer, the new system
F ig u re 8. T h e d iffere n c e o f ris e p ro file showed a remarkable increase of thermal conductivity.
But with metal facer, a change of thermal conductivity of
the new system was little as well as HCFC-141b blown
system.

CONCLUSION

The fundamental study of all water blown PIR foam

system was carried out. The relationship between
components of raw materials and foam physical
properties was examined. Through the studies, new all
water blown PIR foam system suitable for metal-faced
continuous sandwich panels has been developed. This
new all water blown PIR foam system has good storage
stability of polyol premix, good dimensional stability,
Figure 9. Flowability o f novel water blown system
good adhesion strength and good processability
comparable to current HCFC-141b blown foam system,
which is especially suitable for metal-faced continuous
sandwich panels.
Therefore the new system will be applied to actual use
in the near future.

REFERENCES

1. J.Goto, K.Sasaki, S.Mashiko, Y.Kataoka, Y.Kanbara

and I.Ohki, 1997. Proceeding o f the Polyurethane World
Congress 1997, p.17.

2. R.E.Berrier, S.N.Singh, J.S.Costa and V.Bonapersona,

F ig u re 10. T h e r m a l cond uctivity 1998. “Hydrocarbon Blown Rigid Polyurethane Foam for
the Boardstock Industry - A Novel Approach”,

Naruse, Nanno, Kurita, Inohara, Fukami / 95

P roceeding o f the POLYURETHANE EXPO >98, M itsuhiro Kurita
Mitsuhiro Kurita joined Nippon
3. Pierre Doumel, Lothar Zipfel and Pierre Barthelemy, Polyurethane Industry Co. in
1999. “Optimization of Formulations Based on 1990 after graduating from
HFC-365mfc and Blends Thereof”, Proceeding of the Tsuruoka National College of
POLYURETHANE EXPO >99, p.227 Technology. He has been engaged
in the R&D of rigid polyurethane
4. H.Sato, CSuzuki, H.Tanaka, N.Tsukida and foam for eleven years in Central
H.Takeyasu, 1999. “Novel Polyols for All Water-Blown Research Laboratory.
and HCFC-141b Blown Rigid Polyurethane Foams”,
Proceeding of the POLYURETHANE EXPO >99, p.209

5. K.Moriya, A.Naruse, M.Kurita and KSasaki, 1996.

“Fire Resistant All Water Blown Polyisocyanurate Foam” Hideki Inohara
Proceeding of the POLYURETHANE EXPO >96, Hideki Inohara joined Nippon
Polyurethane Industry Co. in
6. K.Moriya, K.Tani and K.Tanaka, 1990. “Novel Fire 1997 after receiving B.S. and
Resistant High Index Polyisocyanurate Foam for Exterior M.S. degrees in Chemistry from
Finish Insulation Application”, 33rd Annual Hiroshima University. He has
Polyurethane Technical / Marketing Conference, p266. been engaged in the R&D of
rigid polyurethane foam for four
years in Central Research
BIO G R A PH IES Laboratory.

A k ira N aruse
Akira Naruse joined Nippon Takao Fukami
Polyurethane Industry Co. in Takao Fukami joined Nippon
1990 after receiving B.S. and M.S. Polyurethane Industry Co. in
degrees in Applied Chemistry 1972 after receiving a B.S.
from Tokai University in 1990. degree in Industrial Chemistry
He has been engaged in the R&D from Kyoto Institute of
of rigid polyurethane foam for Technology in 1972. He has been
eleven years in Central Research engaged in the R&D of TPU,
Laboratory. He is currently team TSU, coatings, adhesives and PU
leader of PIR foam team. foams. He is currently Assistant
to Research Director and rigid
foam R&D manager.
H isash i N anno
Hisashi Nanno joined Nippon
Polyurethane Industry Co. in 1991
after receiving a B.S. degree in
Industrial Chemistry from Kyoto
Institute of Technology in 1991.
After development work on
polyurethane coating and
polyurethane flexible foam, he has
been engaged in the R&D of rigid
polyurethane foam for two years in
Central Research Laboratory.

96 / Naruse, Nanno, Kurita, Inohara, Fukami

Alternative Blowing Agent Solutions for Polyisocyanurate
Continuous Laminate Metal Panels

MICHAEL J. SKOWRONSKI and CAMILLA RIGHI

MICHAEL R. HUSPEK Dow Italia S.p.A.
The Dow Chemical Company Via Carpi 29
1881 West Oak Parkway 42015 Correggio (RE)
Marietta, Georgia 30062 Italy

ABSTRACT
INTRODUCTION
Isocyanurate foams are commonly used as the core foam
in the manufacture of continuous laminate metal-faced It is known that polyisocyanurate (PIR) foams show
panels. With the impending phase out of HCFC 141b, panel outstanding thermal stability. Unlike polyurethane (PUR)
manufacturers are facing the difficult decision of which foams, they do not readily decompose even at very high
alternative blowing agent to adopt. Formulators have the temperatures, and this results in superior reaction to fire
daunting task of providing tailored foam systems based on performance with low smoke development, and also in an
the alternative blowing agents which meet the processing excellent overall resistance to fire [1].
characteristics and performance properties for this All foam systems producers and insulated panels
application. Since metal-faced panels are often used in manufacturers know how delicate it is to process PIR foams.
construction governed by stringent combustibility standards, Due to the high index of PIR foams, special “trimer” -
the flammability performance of the foam core is a major promoting catalysts are used. The activity of these catalysts
issue to be considered. is not efficient at low temperature. Therefore, the initiation
Dow Polyurethane Systems Business has been working to reactions are extremely delayed, followed by a sudden and
develop polyisocyanurate foam systems based on alternative very fast gelation. This drawback that makes processing
blowing agents capable of meeting the different processing difficult has been partly overcome by Dow using a proper
needs and performance requirements of a large customer co-catalysts. The promotion of the trimerization reaction in
base. Polyisocyanurate foam systems for continuous PIR foams gives a high crosslink density resulting in
laminate metal-faced panels have been developed using the improved mechanical properties, but also in the formation of
following blowing agents: HCFC 22 or cyclo/iso pentane for polymers more friable than PUR. This problem is
the U.S. market, and normal pentane for the European aggravated at the surface of the foam where the rapid loss of
market. Furthermore, development is also under way for the heat produced by the exotherm reaction keeps the
systems blown with C 0 2 (from the reaction water/isocyanate reacting polymer relatively cool. It is evident from
or gas) HFC 134a and HFC 245fa. Water, pentane, and C 0 2 morphology studies that a polyisocyanurate foam needs heat
(gas or liquid) are the most attractive alternative blowing to cure; if the conversion of isocyanate is incomplete due to
agents from an environmental standpoint since they offer a poor cure, the foam exhibits high friability, poor adhesion
zero ODP (Ozone Depletion Potential) and zero or very to metal, poor dimensional stability, and poorer fire
marginal GWP (Global Warming Potential). behavior. This problem has been solved by Dow using
In Dow’s developmental work, foam flammability is proper polyols and additives, and by optimizing laminator
examined using the Butler Chimney, NBS Smoke Chamber, process parameters.
UL 723/ASTM E-84 Tunnel, the Radiant Panel (Italian CSE For many years PIR foams have been present in the North
/ RF3), and the German DIN 4102 B-2 tests. Some metal American market in continuous line applications varying
sandwich panels are also tested according to the French M l from garage doors, to building panels, to walk-in cooler
and the European standard draft SBI. In this paper, physical panels. The ability of these foams to meet building code
property and insulation value data are included and flammability standards such as the UL 723 / ASTM E -84
processing characteristics are thoroughly discussed. Tunnel Test and the Factory Mutual 4880 Full Scale Comer

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 97

Test in a cost effective way has made this technology 141b blowing agent. System formulators and panel
attractive for many manufacturers. manufacturers are faced with a multitude of options as we
In Europe, PIR foams are less popular, but interest is move toward this date.
growing rapidly mainly due to the increasing fire-resistance For the United States, the HCFC 22 based technology is
consciousness exhibited by regulations, insurance of interest due to performance, economics, and longevity of
companies, and local authorities. Up to now, each country use as per current EPA regulations. The use o f HCFC 22 in
has had its own regulations and a big effort is underway for foams is allowed until January 1, 2010. However, this date
the European code harmonization. However, it is still not is subject to change given dynamics in the environmental
clear when a European norm will finally enter into force. So and regulatory communities. A new developmental effort is
far, only a portion of the foams used in Europe in metal­ a thorough examination of foam systems based on
faced sandwich construction are rated DIN 4102 B2 and hydrocarbons, long used in Europe but until recently not
PIR foams are produced for niche applications or for very seriously considered in the U.S.. This work based on
severe fire requirements. However, manufacturers are hydrocarbons for rigid faced continuous laminate panels has
showing more and more interest in PIR foams due to the primarily focused on isomers of pentane, especially cyclo
forthcoming Euroclasses of reaction-to-fire and to and cyclo/iso blends.
compensate for the increased flammability of foams blown In Europe, hydrocarbons, mainly normal-pentane or the
with hydrocarbons. commercial blends of normal-pentane and iso-pentane are
Because of the contribution of hydrocarbons to foam gradually replacing HCFC 141b. Due to the high conversion
combustibility behavior, PUR formulations with pentane costs, flammability concerns, and the VOC (Volatile
require an additional amount o f flame retardants to achieve a Organic Compound) nature of hydrocarbons, other routes
given level of foam reaction to fire, with impact on both such as C 0 2 (from the reaction of water with isocyanate)
foam mechanical properties and cost. The use of PIR foams and HFC134a are also being developed and introduced in
is beneficial in terms of better economic and environmental the market.
considerations, allowing a relevant reduction or even the The safest and most environmentally friendly choice
elimination of halogenated flame retardants. Preliminary among possible alternative blowing agents is the use o f C 0 2.
trials conducted by Dow showed that an Euroclass B can be It is an inert gas with a Global Wanning Potential (GWP)
achieved also with a panel produced using a PIR pentane several orders of magnitude lower than HCFC’s and HFC’s,
system [3]. and zero Ozone Depletion Potential (ODP). Moreover C 0 2
Furthermore, Europe is evaluating the introduction of a has no photochemical smog creation potential because,
smoke classification, which would be an additional element unlike hydrocarbons, is not a Volatile Organic Compound
in favor of PIR foams. Moreover, PIR systems are an option (VOC).
for those countries like the United Kingdom where When selecting an alternative blowing agent attention
resistance to fire (British Standard 476 part 22) is an must be paid to costs, thermal insulation properties,
important requirement for building panels. processability, foam overall properties, flammability
January 1, 2003 will signal a very important date for requirements and regulatory compliance.
North American and European insulated panel producers
with the phase out of foam panels produced with the HCFC

Table 1. Therm al co n d u ctivity (gas ph ase) a n d b o ilin g p o in t o f d iffe re n t blo w in g agen ts [3 ]

Blowing agent K-factor gas Boiling point (°F)
(BTU inffi2hr°F at 77°F)
HCFC 141 b 0.066 + 90
HCFC 22 0.073 - 42
HFC 134a 0.095 - 15
HFC 245 fa 0.085 + 59
HFC 365 mfc 0.074 +104
Normal-pentane 0.104 + 97
Iso -pentane 0.097 + 82
Cyclo-pentane 0.083 +120
CO2 (Carbondioxide) 0.113

EXPERIM ENTAL European market to be able to meet the DIN 4102 B2

classification or other national regulations such as Italian
In Table 2 is reported the list of the different PIR systems Class 2 or 1, and the French M l for metal sandwich panels.
presented in this paper. Some of these systems are The new zero ODP experimental systems have been
especially designed for the U.S. market to be able to meet designed to be able to meet both the U.S. and the European
the UL 723 / ASTM E- 84. While others are specific for the Fire Standards.

98 / Skowronski, Huspek, Righi

 Table 2. PIR Systems for the U.S. and European m arket
PIR Systems for the U.S. Market Blowing agent Fire Classification
A HCFC 141b UL 723 E 84
B HCFC 22 UL 723 E 84
C Blends of cyclo/iso pentane UL 723 E 84
PIR Systems for the European Market
D HCFC 141b DIN 4102 B2 / NF P 92 -501 M 1/
ULC S 127-CA N /U LC -S 102
E n -pentane DIN 4102 B2 / NF P 92 -501 M 1(*)
PIR Experimental Systems
F CO 2 from water/iso reaction DIN 4102 B2(*) / NF P 92 -501 M 1(*)
G CO 2 from water/iso reaction DIN 4102 B 2(*) / NF P 92 -501 M1 (*)
H C02from water/iso reaction and HFC DIN 4102 B2(*)
134a
1 CC^from water/iso reaction and HFC DIN 4102 B2(*)
245fa
L CO 2 from water/iso reaction and CO 2 DIN 4102 B2(*)
gas
(*) Foams tested internally at Dow. NF P 92 -501 carried on metal sandwich panels.

Panels produced with System D met the Underwriters’ - NBS Smoke Chamber ASTM E-662
Laboratories of Canada tests ULC S I27 and CAN/ULC - UL 723/ ASTM E84 Steiner Tunnel (core foam)
S102. - German DIN 4102 B2 Foams are classified B2 for a
Methods utilized in generating data include both lab scale flame height lower than 6” (15 cm), otherwise they are
as well as production line investigation. Larger scale classified B3 or not rated.
evaluations were executed at several manufacturers’ - UNI 9174 (Italian CSE /RF3 Radiant panel). Foams
facilities. HCFC 22, cyclo/iso pentane, and n-pentane data have been tested in wall position. In this paper we
were generated using double band laminators approximately report only the value of the damaged foam length. For
100 feet in length, running at 25 - 30 feet/minute. lengths equal or inferior to 8” (200 cm) the Italian Class
Additionally, the C 0 2 blown formulations were processed in 1 can be reached.
a 39 feet long pilot laminator running at approximately 13 - French NF P 92-501 carried on steel sandwich panels.
feet/minute. Concerning the C 0 2 (from water or gas) and - EN 13823 (SBI) according to the European Standard
the HFC’s, the evaluations were carried out in the laboratory Draft of 8th February 2000 implementing the Council
with a high pressure Cannon A-40 dispensing machine. Directive 89/106/EEC carried on steel sandwich panels.
Foams resulting from these investigations were tested
subject to the following criteria:
RESULTS AND DISCUSSION
The discussion of the results achieved with this study for
• Physical Properties PIR systems is reported by the different blowing agents used
- Density ASTMD-1622 in the formulations.
- Compressive Strength ASTM D-1621
- Dimensional Stability ASTM D-2126 HCFC 22
- k-factor ASTMC-518
- Green Compressive Strength - This is a method While the thermal conductivity of HCFC 22 is higher than
developed by Dow to measure the foam curing at a very that of HCFC 141b, at lower operating temperatures foams
early stage: the free rise foam is compressed at 10% in expanded with HCFC 22 show better thermal performance
the perpendicular-to-rise direction by the use of an than foams expanded with HCFC 141b due to the much
Instron 4466 at a fixed time. lower condensation temperature of HCFC 22 (2].
- The foam skin formation has been determined at a In comparison to HCFC 141b, the processing o f HCFC 22
certain temperature opening the mold, previously was characterized by an instant froth consistancy upon
treated with a release agent, at a fixed time and laydown on the bottom skin. This is the result o f attempting
evaluating the percentage of skin formed on the panel. to optimize the k-factor by utilizing a high load o f HCFC 22
in the polyol component. A homogeneous laydown was
• Flammability and Smoke Development obtained with the use of a fixed pouring bar. Table 3
- Butler Chimney ASTM D 3014

Skowronski, Huspek, Righi / 99

provides a comparison between a standard HCFC 141b laminator is void-free. Table 3 provides also a comparison
formulation versus the experimental HCFC 22 system. of physical and flammability properties between the two
During our developmental work it was noted that with foam systems.
HCFC 22 several distinct and separate trails of froth foam As expected, a higher k -factor was measured for HCFC
were evident prior to the movement of the foam/metal into 22 compared to HCFC 141b. The compressive strength and
the laminator, whereas the liquid nature of HCFC 141b dimensional stability results show that the two systems are
chemistry allows for the reacting mixture to create a thin comparable. A Smoke Developed rating of 450 at 3”
uniform rising foam layer on the bottom skin prior to thickness was borderline to meet Class I criteria, the
entering the laminator. For this reason it is critical to formulation can be optimized to lower this value.
balance the rheology of the reacting polymer with catalytic
activity to ensure that the finished panel exiting the

Table 3 . C om parison o f p h ys ic al and flam m ab ility pro p erties o f in d u stria l p an els p ro d u ced w ith
H C FC 22 a n d cyclo /iso pentane blow n system s versus H CFC 141b fo r th e U.S. m arket
System A System B System C System C
HCFC141b HCFC 22 cyclo/iso cyclo/iso
80/20 60/40
Polyol Blend A 75
Polyol Blend B 89
Polyol Blend C 89 89
HCFC 141b 25
HCFC 22 11
cyclo/iso pentane 80/20 11
cyclo/iso pentane 60/40 11
Isocyanate index 240 240 275 275
Cream time, sec 17 froth 12 11
Gel time, sec 35 35 33 31
Tack Free Time, sec 45 43 46 44
Free rise density, pcf 2.30 2.30 2.26 2.25
Physical mechanical properties
Panel Core Density, pcf 2.55 2.70 2.35 2.35
K- factor, BTU in/ft^ hr°F, 75°F mean temperature 0.132 0.149 0.132 0.132
Adhesion to metal face very good very good very good very good
Compressive Strength, psi
X - Axis (across machine) 19 19 19 20
Y - Axis (parallel to rise) 19 18 19 17
Z - Axis (down machine) 23 23 36 37
Dimensional stability % volume
158°F, 95% RH., 28 days 0.4 -0.2 4.1 4.9
158°F, 28 days 0.2 -0.5 0.7 0.0
-20°F, 28 days 0.2 -0.3 0.0 0.0
ASTM D 3014 Butler Chimney % retained weight 93 88.9 93.9 94.6
NBS Smoke Chamber Corrected Max DS 92 171 57 55
UL 723 / E 84 testing (core foam)
Flame spread (*) 25 20 25 20
Smoke Developed (*) 400 450 75 75
Thickness 4" 3" 4“ 4"

(*) These numerical ratings are not intended to reflect hazards presented by this or any other material under actual fire
conditions.

1 0 0 / Skowronski, Huspek, Righi

Cyclo/Iso Pentane and possibly a temperature controlled storage area for the
freshly produced panels.
Isomers of pentane are being evaluated in the United
States for a variety of foam applications. Pure cyclopentane Water
offers the best solubility in typical polyol blends, as well as
the best thermal performance. As a drawback, its relatively One well-known method to produce a C 0 2 blown foams is
high boiling point (120°F) can cause cold age dimensional through the reaction of isocyanate and the water contained
stability problems when the cyclopentane begins to in the polyol blend. The water blown polyurethane systems
condense at cooler temperatures. Meanwhile, iso and developed by Dow have been successfully supplied in the
normal pentane have lower polyol solubility and thermal insulated panels market since 1994. This has been a real
performance, but lead to better foam dimensional stability challenge due to the intrinsic physical obstacles of these
due to their lower boiling points (82°F and 97°F, foams: in fact, the C 0 2 has the highest vapor thermal
respectively), n the attempt to extract only the positive and conductivity amongst different blowing agents. Moreover,
beneficial characteristics of these isomers, this part of our the C 0 2 defuses fast out of the cells with a concurrent (but
work focused on the development of systems based on two several times slower) inward diffusion of air. This means
commercially available pentane blends: Exxsol 2000 (80/20 that with time in a foam which is not covered by
cyclo/iso blend), and Exxsol 1600 (60/40 cyclo/iso blend). impermeable facers, the air replaces the C 0 2 in the cell gas
Table 3 provides specifics on the formulations tested. composition resulting in unacceptable thermal conductivity
The test results show very little difference between the values.
two isomer blends. Both systems demonstrated excellent In practice, for metal sandwich panels in which metal acts
thermal performance and we attributed this to exceptionally as a diffusion tight barrier, the diffusion phenomena can
fine cell structure compared to conventional HCFC 141b only take place through the open sides of the panels. [4]
systems. Both systems also showed similar y/z compressive Extensive studies showed that the initial thermal
ratios, which typically should be as close as possible to a 1:1 conductivity value remains for a long time essentially
ratio. Deviations may be attributable to the laminator line unchanged for the largest part of the panel due to the long
speed and the system reactivity profile not being perfectly path of diffusion. The initial higher thermal conductivity
optimized. Still, the dimensional stabilities did not suffer value of foams blown with C 0 2 (about 15 % higher than
because of these deviations, producing very acceptable HFC 141b blown foams) resulting in slightly worse or equal
results. Perhaps the most significant achievement of this insulation properties of the final panel has been largely
study was the UL 723 test results, which gave significantly accepted by the European industry.
lower Smoke Development values compared to HCFC 141b. Dimensional stability is another issue to be carefully
considered. The foam dimensional stability is the balance
Normal Pentane between external and internal forces. As mentioned before,
in an unfaced foam, the C 0 2 moves out of the cells faster
Normal-pentane is now largely used as blowing agent in than the inward diffusion of air. This results in a minimum
Europe in the continuous laminate metal-faced panels cell pressure that in certain circumstances, such as high
production with polyurethane core. Because of the external load, elongated foam cells or too low applied
contribution of hydrocarbons to the foam combustibility densities, may be responsible for foam shrinkage. Dow’s
behavior, formulations with pentane require an additional experience over years indicates that in case o f metal
amount of flame retardants to achieve a given level of foam sandwich panels foamed with a proper technology a good
reaction to fire, with impact on both foam mechanical panel dimensional stability can be reached when utilizing
properties and costs. A solution, adopted now by some the water blown technology for foam applied densities not
manufacturers, is the use of polyisocyanurate foams. lower than 2.7 pcf.
Table 4 shows a comparison of foam properties for panels Challenging is the use of isocyanate reacting with water as
produced in Europe. The use of n-pentane, compared to sole blowing agent in PIR foams. In fact, the urea formation
HFC 141b results in a worse thermal conductivity, as from the reaction water-isocyanate increases the friability of
deductible from the worse lambda gas of n-pentane, but in the foam, mainly on the surface. Therefore during our
improved mechanical properties and dimensional stability. developmental work we redesigned the chemical
In spite of the pentane flammability, foams are DIN 4102 formulation as needed.
B2, can easily reach the Italian Class 2 and the sandwich The developmental work described in this paper lead to
panels meet both the French M l classification and also the the definition of fully water blown polyisocyanurate foams
Euroclass B for a specific joint design [3]. with isocyanate index 240-260 that showed good final
Compared to polyurethane systems, the industrialization physical-mechanical properties, reduced friability and
of a n-pentane blown PIR system requires an accurate superior flame behavior.
temperature control through the whole panel manufacturing The results obtained at our internal pilot plant laminator
process, with preheating of facers, a good conveyor heating are shown in table 4; two different polyol approaches have
been compared to the above mentioned HCFC 141b and n-

Skowronski, Huspek, Righi / 101

pentane PIR formulations. The initial thermal conductivity is HCFC 141b was evident. Unfortunately the data from the
higher but dimensional stability and compressive strength NBS smoke chamber predicts a possible failure in terms of
values are good. Moreover, the foams showed an excellent Smoke Development in the UL 723 E -84 test for isocyanate
cure at the conveyor exit, as confirmed by the green index 240-260, but the positive effect that the isocyanate
compressive strength values reported in Table 5. index increase had on smoke reduction was also evident.
During the pilot plant laminator trials, the temperature System G at index 300 showed excellent flammability and
was carefully controlled. Both systems F and G showed smoke values (see Figure 1).
good adhesion with metal faces preheated at 104°F and with All the above trials confirmed that polyisocyanurate water
the conveyor at 122°F. Also the adhesion tests carried on in based foams can be processed with little to minor process
the laboratory showed acceptable adhesion when the mold and equipment modifications. These foams showed a good
was heated at 122°F and the panel remained at that adhesion with preheated facers and a conveyor temperature
temperature for 5 min. Encouraging results have been of 122°F and a good dimensional stability at a density of
obtained also at 113°F but for a retention time of at least 7 2.82 pcf at isocyanate index of 240 -260. A higher index to
minutes. In any event, the poor skin formation of the water reduce smoke development is also possible, but higher metal
based systems underlined how critical and temperature- and conveyor temperature may be necessary.
dependant this technology is.
As far as flammability is concerned, the superior behavior
of the two water systems compared to n-pentane and even to

Table 4. Comparison o f physical and flammability properties of industrial panels produced with normal-pentane and developmental
_________________________ water systems versus HCFC 141b for the European market
System D System E System F System F System G System G System G
HCFC141b n -pentane w a te r(1) w a te r« w a te r<1> w ater <1> w a te r<1)
Polyol Blend D 83
Polyol Blend E 92
Polyol Blend F 100 100
Polyol Blend G 100 100 100
HCFC 141b 17
normal - pentane 8.2
Isocyanate index 300 290 240 260 240 260 300
Cream time, sec — .... 6 6 5 5 6
Gel time, sec — — 26 28 25 29 33
Physical mechanical pro pertiesw
Overall foam density, pcf 2.79 2.85 2.82 2.69 2.81
Core foam density, pcf — 2.58 2.45 2.55
K- factor, BTU in/ft2 hr°F, 50°F mean 0.134 0.152 0.163 0.164 0.164
temperature
Compressive Strength parallel to rise, psi 17.4 29.7 39 35 30
Dimensional stability %volume
158°F, 95% RH., 7 days 2.8 1.8 -4.5 -3.6 -3.1
+ 230°F, 7 days 2.7 -0.6 -3.9 -2.5 -3.2
-22°F, 7 days -2.1 -0.4 -0.5 -0.3 -0.4
Reaction to fire
German DIN 4102 B2 flame height (cm) 7 12 7.4 6.1 6.1 5.8 5
Italian CSE /RF3 wall position damaged length 275 300 275 250 220 — 195
(cm)
ASTM D 3014 Butler Chimney %retained weight 92.6 89 91.3 90.8 — 91.9 94.6
NBS Smoke Chamber Corrected Max DS 100 135 220 182 ... 148 118
French NF P 92-501 (steel sandwich panel) M1 M1 M1 ... M1 M1 —
EN 13823 (SBI) fire/smoke classification B S3 C S3 or B
(steel sandwich, standard joints) S 3 (3»

* These numerical ratings are not intended to reflect hazards presented by this or any other material under actual fire conditions.
(1) The water systems F and G have been run on a laboratory pilot laminator.
(2) All panels have a thickness of 3".
(3) Two different classifications have been obtained for two different joint design.

1 0 2 / Skowrvnski, Huspek, Right

 Table 5. Physical-m echanical properties and flam m ability o f the new developm ental zero ODP systems.
Foams produced in the laboratory
System F System H System 1 System L System L
water HFC 134a HFC 245 fa CQ2 gas C 02gas
Polyol F (1) 100 100 100 100 100
HFC 134a 2.8
HFC 245fa 11
C02 gas (solubilized on the isocyanate) 1% 1%
Isocyanate Index 240 240 307 277 360
Cream Time, sec 7 froth 5 froth froth
Gel Time, sec 36 32 35 31 33
Free Rise Density, pcf 2.28 2.3 2.17 2.45 2.76
Green Comp. Strength, psi 3 min 11.3 10.6 12.3 18.8 —
Green Comp. Strength, psi 6 min 16.8 14.9 17.1 22.9 —
Physical m echanical properties

overall density, pcf 2.8 2.67 2.7 2.83 2.95

Initial Kfactor,BTU in/ft2 hr°F, 74°F mean 0.164 0.166 0.153 0.159 —
temperature
Compressive Strength parallel to rise, psi 42.8 29.1 38.1 47.6 —

Compressive Strength perpend, to rise, psi 17.4 17.1 16.5 25.5 —

Dimensional Stability (2) + 158°F, 95% RH., 7 days -1.7 -1.1 -0.3 + 1.0 ...

Skin formation mold at 113°F 5 min (%) 0 0 20 0 0

Skin formation mold at 122°F 5 min (%) 20 20 100 100 90
Adhesion to metal at 113 °F 7 min good / very good good good
Adhesion to metal at122°F 5 min good / / / good

Sm all scale fire tests

German DIN 4102 B2 flame height (cm) 6.6 6.8 5 5 5
ASTM D 3014 Butler Chimney %retained weight 91.3 — 91.4 92.3 95.7
NBS Smoke Chamber Corrected Max DS 226 — 191 161 129
(1) for each formulation a catalysts, surfactant and a water level adjustement was needed to obtain similar reactivities and
densities.
(2) % volume change of samples 25" x 25" x 2" without facings

pressure. At this level, or lower, the polyol containing HFC

W ater / H FC ’s 134a is liquid and doesn’t need particular precautions during
handling. HFC 134a may also be added on the high pressure
The use of HFCs as co-blowing agents in combination polyol stream by the use of a static mixer. Within the
with the carbon dioxide produced by the reaction solubility limit the liquid laydown is maintained, whereas at
isocyanate-water offers some advantages in terms of higher levels a froth effect is noticed.
dimensional stability or thermal conductivity. The Properties of foams produced in the laboratory are shown
possibility to replace part of the water with a physical in table 5. In System H, 2.8 % of HFC 134a in the polyol
blowing agent results in a lower urea formation with replaces a very small amount of water with no major
beneficial effects on the adhesion to facings at a reasonably differences in performance compared to a fully water blown
high conveyor temperature; moreover, higher isocyanate system. The experience achieved on polyurethane systems,
indexes are possible at still reasonable isocyanate/polyol anyhow, indicates that an initial cell gas composition with a
ratio with improved flammability characteristics. molar fraction 0.1 / 0.9 of HFC 134a / C 0 2 as in System H,
HFC 134a is a zero ODP low boiling blowing agent with should be enough to raise the minimum cell pressure in the
a limited solubility in the polyol blend. It may be slowly foam thanks to the contribution of the vapor pressure o f the
added to the polyol blend via a static mixer. Studies showed HFC 134a with a consequent positive impact on dimensional
a maximum solubility in a typical polyol for PER foams of stability. This means that compared to a PIR fully water
2.S % in weight at room temperature and atmospheric blown system, the applied density may be reduced. In fact,

Skowronski, Huspek, Righi / 103

the dimensional stability after aging at 70°C and 95% RH of lower lambda value of this gas compared to C 0 2, an
foams produced with System H is good at lower density. A improvement in terms of thermal conductivity of the foam
further density reduction is also possible, but should be has been obtained. Compressive strength values and
carefully checked in the laboratory. Significant applied dimensional stability at a lower density were good when
density reductions can be achieved with higher amounts of compared to the fully water blown system. The values for
HFC 134a but the strong froth effect may be critical in terms skin formation and the adhesion to metal obtained in the
of subsurface voids. laboratory show an improvement compared to Formulation
HFC 245 fa has a zero ODP and a low GWP and it is not F and a conveyor temperature of 113°F should provide a
considered a VOC in the US. Furthermore its boiling point good adhesion to the facings.
of 59°F allows the handling as a liquid under moderate A drop- in replacement of HCFC 141b with HFC 245fa in
pressure. Its solubility in the polyol blend is very good and System A for the U.S. market has been evaluated in the
there is no limitation in the quantity to be used. At high laboratory. As shown in table 6, a part from a lower but still
levels, HFC 245 fa gives a moderate froth while at medium- acceptable green compressive strength of HFC 245fa, results
low levels the laydown is still liquid - as reported in table 5 were very promising mainly in terms of flammability and
for System I. A rather limited quantity of HFC 245 fa has smoke development.
been used in order to balance a significant improvement in
processing with foam performance and c o s t. Thanks to the

Table 6. Comparison HFC 2451a versus HCFC 141b. Foams produced In the laboratory
System A System A
HCFC 141b HFC 245 fa
Polyol Blend A 100 100
HCFC141b 28
HFC 245 fa 28
isocyanate index 240 240
Cream time (sec) 12 froth
Gel time (sec) 27 25
Free rise density , pcf 2.45 2.46
Initial K-factor, BTU in/ft2 hr°F, 74°F mean temp. 0.131 0.130
Green comp, strength, psi 3 min 9.7 7
Green comp, strength, psi 4 min 11.2 8.1
Green comp, strength, psi 6 min 14.1 10.7
Comp, strength parallel to rise, psi - 54.2
Comp, strength perpend, to rise, psi - 24.8
Skin formation mold 104°F 5min 20 95
Skin formation mold 113°F 7min 80 100
Adhesion to metal at 104°F , psi very good Very good
% volume change + 158°F, 95% RH.,7 days - -0.2
Small scale fire tests
DIN 4102 B2 flame height (cm) 9.0 7.4
Butler %retained 93 87.8
NBS smoke 92 66

ratio, more isocyanate is made available for the trimer

W ater / Carbondioxide gas formation with improvement in fire performance. In our
developmental work we have decided to add the C 0 2 gas on
The addition of C 0 2 gas as blowing agent is still the best the isocyanate side in spite of the fact that its solubility is
environmental choice without the drawback effect of urea lower on the iso compared to the polyol. The advantages of
formation during the chemical isocyanate-water reaction doing this derive from the good chemical stability o f C 0 2 in
with the consequent reduction in the surface brittleness of the isocyanate and from the fact that PIR foams have a very
the foam due to the partial replacement of the water with the high level of isocyanate in the system that allows anyhow a
C 0 2 gas. Another significant advantage is that, reducing the good introduction of the C 0 2 in the final foam. During the
water level and maintaining the same isocyanate/polyol laboratory trials the C 0 2 has been slowly added to the iso

1 0 4 / Skowronski, Huspek, Righi

back pipe of an HP machine through a static mixer . The Figure 1 shows an evident improvement in smoke release
maximum quantity of CO 2 we could add to the isocyanate with the isocyanate index increase.
was 1% (as reported in Table 6 - System L), this created a With a proper modification of the foaming head and of
very strong froth with an evident reduction in flow, and a the distribution laydown of the material it is also possible to
very fine cell structure. Fine and less isotropic cells are most use higher amounts of C 02 with further reductions in the
probably the reason of the improved thermal conductivity water level.
compared to the fully water blown systems. Also skin The effect of the strong froth of the C 02 on panels’
formation and adhesion were improved as expected. aesthetics is something to carefully evaluate.

Figure 1. Smoke developm ent in the NBS Chamber vs. isocyanate index

Normal-pentane is now largely used as blowing agent in

CONCLUSIONS Europe in the continuous laminate metal-faced panels
production with polyurethane core. PIR foams can meet the
With the phase out HCFC’s approaching fast, Dow offers DIN 4102 B2 classification. Furthermore on steel sandwich
a wide variety of polyisocyanurate (PIR) systems based on panels, the NF P 92-501 M l and the Euroclass B with
different alternative blowing agents options especially proper joint design can also be met.
designed to meet the most important flammability and The Carbondioxide solutions for PIR foams are
smoke development regulations. The choice of the blowing challenging but feasible. Trials showed that panels can be
agent strictly depends on thermal conductivity requirements processed provided facers preheating, conveyor temperature
and overall economics that for certain solutions may include higher than 120°F, and possibly a temperature controlled
the significant cost of machinery modification. storage area for the freshly produced panels. The reaction to
For the U.S. market, where the differentiate phase out of fire of these foams is excellent and the smoke development
the blowing agents allows HCFC 22 as a transitional can be reduced by increasing the index. The replacement of
solution, a HCFC 22 froth system has been designed to meet part of the water with C 0 2 is in favor of this index increase.
the UL 723 /ASTM E 84 Class I requirements. The resulting froth may require a fixed pouring bar for an
An extensive developmental work carried out by Dow led homogeneous laydown. The Flammability results obtained
to the design o f a system for the use of cyclo/iso pentane during this developmental work confirm that with these
blends capable of meeting the UL 723 /ASTM E 84 and foams the most severe fire standards will be met.
characterized by low smoke development, thermal The addition of soluble quantities o f HFC 134a in the
conductivity similar to the traditional HCFC 141b foams high water containing polyol blend may favor foam applied
and easy processing. density reductions. Nevertheless evaluations are still in
progress for higher quantities o f HFC 134a that could

Skowronski, Huspek, Righi / 105

improve flammability and allow farther density reductions Appliance and Construction Applications. Since 1994 she
due to a strong froth effect that may be critical in terms of has project responsibilities in Research and Development
subsurface voids. for rigid polyurethane foam applications. She is also flame
During this developmental work we noticed that HFC retardancy focal point for Dow Polyurethane Systems
245fa may be a choice to successfully replace HCFC 141b. Business.
In fact, laboratory data show similar performance, initial
thermal conductivity and improved flame behavior.
Furthermore when considering economics, HFC 245fa can Michael R. Huspek
also be used as a co-blowing agent in high water containing
PIR systems with some advantages in terms of processing, Mike Huspek received his B.S. in
flammability and smoke development. Chemistry and M.B.A. from Loyola
University of Chicago, in 1985 and
REFERENCES 1987, respectively. Mike began to
work in the urethane field in 1984 as
1. W.F. Gum, W. Riese, H. Ulrich “Reaction polymers” a technician for Stepan Company,
Hanser Publishers, 1992. where he was involved in the
2. S. E. Moore, D. M. Ference “Polyurethane system preparation of polyester polyols for
optimization for energy efficiency” POLYURETHANE rigid foam applications. In 1987,
EXPO, 1998. Mike joined Flexible Products Company - Marietta,
3. R. Van Den Bosch, A.Ottens, L. Bertucelli “Challenges Georgia, where he held a number of rigid foam applications
of metal-faced sandwich panels with polyurethane core” development positions, culminating in 1999 as Product
UTECH 2000. Development Manager. In 2000, Flexible Products
4. L.Bertucelli, M. Vreys, F. Pignagnoli, A. Ottens, P. Company was acquired by The Dow Chemical Company
Keller “Phase-out of ozone depleting substances in the and since that time, Mike has been the Rigid Foam Research
manufacture of metal-faced sandwich panels with Leader for Dow’s North American Systems House business
polyurethane foam core” UTECH Asia 1999. unit.
BIOGRAPHIES

Michael J. Skowronski

Michael Skowronski is a Senior

Research Specialist in the North
American Systems House of The Dow
Chemical Company located in
Marietta, Georgia. He holds a B.S. in
Chemistry from Loyola University,
Chicago and a M.S. in Management of
Technology from the University of
Miami. He has authored five
publications and been awarded 17
U.S. patents. His responsibilities
include the development of alternative blowing agent
technologies for fire rated rigid urethane and isocyanurate
foams.
Camilla Righi

Camilla Righi received her

Organic Chemistry degree from the
University of Modena (Italy). She
joined Dow in Italy in 1985. She
was involved in Technical Service
and Development for Domestic*

* Trademark o f The Dow Chemical Company.

1 0 6 / Skowronski, Huspek, Righi

Processing Gaseous Blown Spray Applied Polyurethane Foam

IAN WHEELER, RICHARD CROOKER and DEAN WICKWIRE

ATOFINA Chemicals, Inc.
900 First Avenue
King o f Prussia, Pennsylvania 19406

Abstract HCFCs are phased out or are no longer

available, HFC Forane® 134a the zero
HFC-134a is a substitute-blowing agent ODP can be readily utilized as the most
for HCFC-141b. It has been successfully cost-effective substitute to produce spray
used in many applications, such as applied foams with excellent properties.
appliance and pour-in-place. However,
due to its boiling point and vapor Introduction:
pressure many have discounted HFC-
134a acceptability as a blowing agent for At ATOFINA Chemicals Inc. we have
spray applied polyurethane foam. evaluated gaseous blowing agents
(GBAs) for various applications. The
The addition of high levels of HFC-134a demonstrated advantages include good
to B-side blends results in blends with low temperature thermal conductivity,
vapor pressures significantly above improved low temperature dimensional
atmospheric pressure. This will result in stability, increased compressive
these blends having to be handled in strengths, reduction of pressure on jigs
pressurized containers. However, low and improved part
level of HFC-134a can be added to B- appearance{ 1,2,3,4,5,6,7,8 }.
side blends and the vapor pressure can
be low enough to handle them in Spray applied polyurethane foam,
standard containers. however, has some unique differences
from other applications. Spray foam is
This paper focuses on further typically done at a construction site
developments and improvements rather than in house application. The
achieved in the spray foam systems common equipment is a fixed-ratio
using gaseous blowing agent (GBAs), dispensing unit supplied by drums of A-
which are also relevant to other side and B-side. The need to formulate
Polyurethane Rigid (PUR) foam to meet these restrictions has made
applications. In particular, we replacing 141b difficult.
demonstrate that Forane® 22/142b is the
best choice to produce spray-applied At ATOFINA Chemicals Inc. we have
PUR foam products when HCFC-141b is examined gaseous blowing agents for
phased out by 2003. However, if this use in spray applications. We have
option is unavailable, the paper will also developed systems for spray foams using
demonstrate that when and where 141b, 22/142b and 134a. Although there

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 107

 are difficulties using gaseous blowing
agents for spray foam, all of these are The North American spray foam
lower ODP and their cost/performance industry is, therefore, required to deal
ratios are very good. with a blowing agent that is either
gaseous at room temperature or
In this paper, we present the results we flammable. Although these are difficult
have obtained with these blowing agents. issues, ATOFINA believes they can be
However, our results are not fully dealt with in a cost-effective manner.
optimized. Systems formulators with
more raw material options to work with Experimental
should obtain better results.
There are many spray applied foam
formulations from numerous suppliers.
Each supplier has a unique formulation
Comparison of Blowing Agents for each application depending on the
properties required. In this paper, we
focus on “typical” non-fire rated 2.0 pcf
Table 1 shows the popular blowing agent density cavity fill foam and non-fire
choices currently in use or proposed as rated 2.5 pcf density roofing foam.
substitutes forHCFC-141b. Other than
HFC-365mfc and cyclopentane, the At ATOFINA, we developed the GBA
other options are gaseous at room formulations using published data on
temperature and atmospheric pressure. formulating spray systems we used both
Both cyclopentane and HFC-365mfc are commercially available polyether and
flammable. Additionally, HFC-365mfc polyester based polyols, surfactant and
is not planned to be available in North catalyst packages. The formulations are
America as a foam-blowing agent for shown in Table 3 (Formulations for
some time owing to patent restrictions.
Table 1 Blowing Agent Properties

HCFC HCFC
HCFC HFC HFC HFC- Cyclo
Properties -142b -22/142b
-141b -134a -245fa 365mfc pentane
(40/60)
Gaseous Phase
Thermal 0.069 0.080 0.078 0.094 0.085 0.074 0.083
Conductivity
(a). 77°F & 1
Boiling Point
89.6 15.4 -18.4(1) -15.7 59.5 104 120.7
(°F)
Molecular 70
117 101 95 102 134 148
Weight
Ozone 0.065 0.06 0 0 0 0
0.11
Depletion
Vapor Pressure 6.8 4.9
10 41.8 60.2 68.5 17.8
@ 68 F (psia)

(1)Bubble Point

1 0 8 / Wheeler, Crooker, Wickwire

 We used low viscosity high amine
pcf Cavity Fill Spray Foam) and Table 4 polyols that gave low vapor pressures
(Formulation for 2.5 pcf Roofing Spray (i.e. better compatibility with 134a)and
Foam). The same polyols, surfactants, reduced viscosity of the B-side blends.
catalysts and isocyanates were used for The low viscosity helps to improve the
each formulation (Shown in Table 2) . mixing and compensates for the
tendency of B-sides containing relatively
The low level GBA formulations were low levels of blowing agents to have
formulated to enable them to be shipped high viscosity. Additionally, the catalyst
in standard drums. The pressure rating package was modified to include Polycat
for Standard drums (UN 1Al) is 24 psig. 5 a catalyst which speeds up the
Therefore, we formulated to be 50% of beginning of the reaction. This helped to
that pressure rating at 110°F. offset the cooling effect of the GBA
Additionally, We formulated based on frothing (i.e. coming out of solution
maintaining the 1:1 volumetric ratio of when exposed to atmospheric pressure).
A-side to B-side that is standard in the This cooling of the initiation of the
industry. reaction can impact peak exothermic
temperature, and delay the time to reach
For the lower density 2.0 pcf cavity fill this peak.
foam we were able to formulate spray
foam using gaseous blowing agents. We sprayed the materials using
However, for HFC-134a we could not Polyfoam Products, Inc. equipment with
maintain the 1:1 ratio, achieve the chemical temperatures of 110°F. We
blowing level required for 2.0 pcf foam sprayed on to cardboard at room
and keep the blend vapor pressure below temperature.
12 psig @ 110°F, because the high CO2
blowing level required increased the The formulations are shown in Tables 2
amount of A-side need to react with and 3. The Cost Relative to HCFC-141b
water to create the CO2 . We evaluated is based on the formulations developed.
2 levels of GBA to produce system We used market prices for ingredients
suitable for use using the current used.
standard drums or for use with
containers rated for higher pressure. The foam properties were measured
using Standard ASTM methods used for
foam testing.

Table 2 Poyol Blend Formulation

COMPONENTS % OF TOTAL
SILTECH 425LV 35.7
Voranol 391 29.8
Jeffol SA-395 11.9
Terate 2541 14.3
Antiblaze 80 8.3

Wheeler, Crooker, Wickwire / 109

 Table 3 Formulation for 2.0 pcf Cavity Fill Spray Foam
HCFC - HCFC-22/142b HCFC-22/142b HFC-
COMPONENTS 141b (40/60) (40/60) 134a

POLYETHER POLYOL
65 64.94 62.22 62.39
(Weight % of B-side)
POLYESTER POLYOL
9.7 9.64 11.5 10.9
(Weight % of B -side)
%FLUOROCARBON BLOWING
AGENT 12.5 12.4 4.5 12.16
fWeieht %of B-side!
% C02 BLOWING
32 37 70 32
(Mole %of Total Blowing)
B-SIDE BLEND VAPOR PRESSURE
<0 28 5 45
@ 75°F (psig)
B-SIDE BLEND VAPOR PRESSURE <0 NA 10 NA
(Si 110°F (psig)
COST RELATIVE TO HCFC-141B
0 0.00 -0.045 0.05
($/lb of foam)
NA -not Available

Table 4 Formulation for 2.5 pcf Roofing Foam

COMPONENTS HCFC-141b HCFC- HCFC- HFC-134a HFC-134a

22/142b 22/142b (high level) (low level)
(high level) (low level)
% PO LYETHER
PO LYO L 70.9 74.32 72.2 74.4 77.6
(Weight % of B-side)
% POLYESTER
PO LYO L 7 11.5 11.5 11.5 10.0
(Weight % of B -side)
% PH YISIC A L
B LO W IN G A G EN T 10.5 9.9 3.6 9.6 3
(Weight % of B-side)
% C 02 B LO W IN G
(Mole % of Total 35 35 70 32 80
Blowing)
B -SID E B LEN D VAPOR
PRESSURE @ 75°F <0 17 3 40 5
(psig)
B -SID E B LEN D VAPOR
PRESSURE @ 110°F <0 30 7 NA 12
(Psig)
COST R E L A T IV E TO
HCFC -141B 0 -0.003 -0.04 0.036 -0.03
($/lb of foam)
NA -not Available

1 1 0 / Wheeler, Crooker, Wickwire

Blending & Handling The resulting B-side was then placed in
a separate container. The blends having
We blended each system in a 2 gallon vapor pressures above 15 psig were put
pressure vessel rated to 250 psig. The in a 2 gallon pressure pot. A over
polyols, surfactant, catalyst and water pressure of 10 psig nitrogen was applied
were added to the vessel. The resulting to be able to deliver liquid from the
premix was then set to circulate through container. The blends with low levels of
a static mixer. The blowing agent was GBAs and low vapor pressures we put in
then added upstream of the mixer to the a standard container.
blend, which was then circulated until it
reached a constant pressure.
Foam Properties

Table 5 Properties: 2.0 pcf Cavity Wall Foam

Properties HCFC-141b HCFC- HCFC- HFC-134a

22/142b 22/142b (high level)
(high level) (low level)
Density (pcf) 2.1 2.1 2.1 2.1
R-Value (Initial) 8.7 7.5 6.7 6.25
Initial Thermal 0.115 0.133 0.149 0.160
Conductivity @75°F (Btu-
in /tfh r T )
InitialThermal 0.110 0.120 0.137 0.145
Conductivity @50°F (Btu-
in/f^hrTF)
Parallel Compressive 33 41 45 47
Strength (psi)
Perpendicular 16 20 22 21
Compressive Strength
(psi)
%Volume Change 28
days @
-20 °F -1.1 0.13 0.15 0.12
158 °F 2.8 0.75 0.95 1.2
158 °F& 100% Relative 5.2 2.8 3.9 2.7
Humidity

Wheeler, Crooker, Wicbvire / 111

 Table 6 Properties: 2.5 pcf Roofing Foam

Properties HCFC-141b HCFC- HCFC- HFC-134a HFC-134a

22/142b 22/142b (high level) (low level)
(high level) (low level)
Density (pcf) 2.5 2.5 2.5 2.5 2.5
R -Value (Initial) 8.7 7.5 6.6 6.3 5.9
Thermal Conductivity 0.115 0.133 0.151 0.159 0.169
(®/75°F (Btu-in/ft2hr °F )
Thermal Conductivity 0.110 0.120 0.139 0.146 0.156
@50°F (Btu-in/ft2hr °F)
Parallel Compressive 45 53 55 54 58
Strength (psi)
Perpendicular 22 27 28 26 28
Compressive Strength
(psi)
% Volume Change 28
days (a}
-20 °F -0.71 0.16 0.15 0.13 -0.12
158 °F 2.2 0.63 0.52 1.1 -0.95
158 °F& 100% Relative 4.2 2.3 2.1 2.5 -3.3
Humidity

All foams produced gave good adhesion difference in R-value, especially with
to the cardboard substrate. The surface HCFC-22/142b . These levels, however,
of the foams blown with higher levels of require that the B-side material be
GBAs was slightly rougher than the maintained as a liquid under pressure.
HCFC-141b and low levels of GBA. This would require some modifications
in the handling of the B-side for the
The results in Tables 5 and 6 show that formulators and sprayers. Foams made
foams blown with the high levels of with HCFC-22/142b are similar in cost
HCFC-22/142b have R-Values within to those made with HCFC-141b, while,
15% of those produced with HCFC- the high HFC-134a system would cost
141b. The initial R-value of HFC-134a approximately $0.03 to $0.05 per pound
blown foams were within 30% of those of foam more than the HCFC-141b
produced with HCFC-141b. In other system.
experiments, we found that GBA blown
foams age at the same or better rates At lower levels of GBAs with partial
than HCFC-141b blown foams. pressures less than 12 psig at 110°F,
permitting the use of standard drums,
Compressive strength and Dimensional we reduce the cost per pound of foam
Stability are better with GBAs than with $0.03 to $0.05 compared to HCFC-141b.
HCFC-141b. This indicates that the Additionally, based on the compressive
densities could be reduced to obtain the strengths and dimensional stability
same values. results lower density foam could be
used. We plan to formulate to achieve
Cost Performance equivalent compressive strength. This
will allow us to better calculate cost per
GBAs can be formulated at levels R-value.
similar to HCFC-141b. This reduces the

1 1 2 / Wheeler, Crooker, Wickwire

Conclusions: Density, Flowability, Cabinet
Energy and Product Appearance,”
Proceedings of the Polyurethane
GBAs can be used in spray applied Expo 1999, September 12-15,
polyurethane foam. 1999.
3. Walker, G., D. Tomlinson, G.
Incorporating GBAs into a spray English and W.Bazzo. “An
formulation at relatively high or low Effective Strategy to Eliminate
CFCs in Refrigerator Insulation,”
levels each has its own advantages. At
Proceedings of the Polyurethane
levels similar to HCFC-141b, 9 to 13 World Congress 1993,
percent of B-side, they require October 10-13,1993.
pressurized containers. The formulations 4. Gurecki, J. R. and I. A Wheeler.
are processed through standard spray “The Introduction of HCFC-
22/141b Appliance Foam
equipment with no modifications. The
Technology in North America,”
cost/pound of foam made with HCFC- Proceedings of the SPI
22/142b is similar to HCFC-141b at Polyurethane Division Conference
these levels, while the cost/pound of 1995, September 26-29, 1995.
foam is $0.03 to $0.05 higher for HFC- 5. Kuhn, E. and P. Schindler.
“Advances in the Understanding of
134a foams. Additional reductions in
the Effects of Various Blowing
foam cost should be possible by lower Agents on Rigid Polyurethane
the density, to reduce compressive Appliance Foam Properties,”
strength to that which is required. Proceedings of the Polyurethane
World Congress 1993, October 10-
13, 1993.
At lower levels of GBAs, the resulting
6. Gurecki, J, I. Wheeler and C. M
blends have < 12 psig at 110°F, can be Sungail. “Comparison of
shipped using standard drums, the same Appliance Insulation Systems,”
B-side handling as HCFC-141b. These Proceedings of the Polyurethane
lower levels result in reduced cost/pound Expo 1998, September 17-20,
1998.
of foam relative to HCFC-141b. The
7. Krueger, D. C. and C. J. Reichel.
foams produced with low levels of “Chlorodifluorometane as the
GBAs also show improved compressive Primary Blowing Agent for Rigid
strength and dimensional stability, Urethane Foams Using
indicating the possibility or reducing the Conventional Low and High
cost by lower density. Pressure Foam Mixing
Equipment,” Proceedings of the
Polyurethane World
REFERENCES Congress 1991, September 24-26,
1991.
1. Wu J., Albouy A , Crooker R., 8. Creazzo, J. A and H. S. Hammel.
Wheeler I., “Smooth Transition for “Replacing CFC-11 with HCFC-22
the Appliance Industry Using Safe in Polyurethane Foam,”
and Cost-effective Gaseous Proceedings of the Polyurethane
Blowing Agents” Proceedings of World Congress 1991,
Polyurethane World Congress September 24-26,1991.
2000, Oct. 8 -11,2000 9. Logsdon, P. B. and D. J. Williams
2. Haworth, J, J. Sinaga, P. Irwin, D. “Third Generation Blowing Agent
Latham and R. Morgan. “New Update for The Appliance
Developments with Appliance Industry,” Proceedings of the
Rigid Foam Systems Utilizing Polyurethane Expo 1998,
Liquid and Gaseous Blowing September 17-20,1998.
Agents to Achieve Optimum

Wheeler, Crooker, Wickwire / 113

 10. Logsdon, P. B. and D. J. Williams. entry door, spray, panels and appliance
“The Next Step: markets; and subsequently, in 1994, he
Commercialization Timeline for
HFC-245fa,” Proceedings of the was a Global Development Specialist for
Polyurethane Expo 1999, the Appliance and Construction Systems
September 12-15, 1999. markets.
11. Boze, C. D., and K. J. Elsken.
“Effects of Processing Variables on Richard Crooker
a Polyurethane Foam System
Blown with HFC 245fa,” Richard
Proceedings of the Polyurethane Crooker is the
Expo 1998, September 17-20, Director of
1998. Fluorochemicals
R&D for ATOFINA
BIOGRAPHIES Chemicals, Inc., with
responsibility for
Ian Wheeler process development
Currently, Ian and new applications
Wheeler is a project for fluorochemical blowing agents and
leader for Rigid refrigerants. He received his BA from
Polyurethane Foam Knox College, Galesburg, IL, and Ph.D.
Products at in physical organic chemistry
ATOFINA’s
Chemicals, Inc. at DISCLAIMER
the Technical Center Nothing contained herein should be taken as
an inducement to infringe any patent and the user
in King of Prussia, is advised to take appropriate steps to be assured
PA. He received his diploma in that any proposed use of these mowing agents
will not result in patent infringement. In this
Chemical Engineering from Mohawk regard, among others, the reader should be aware
College in 1982, and has since received of such basic patents as Bayer’s U.S. 5,496,866
on the use o f HFC -245fa as a foam blowing
his Business Degree from Northwood agent, DuPont’s U.S. 5,439,947 & U.S.
University in Michigan. Previously, he 5,516,811 on the use of HFC-134a, HCFC-124
and HCFC-22 as foam blowing agents, and non-
worked for ICI Polyurethanes, Canada as U.S. counterparts of the forgoing patents.
a Technical Service Representative for

1 1 4 / Wheeler, Crooker, Wickwire

Meeting the Needs of the Future with HFC-245fa and HFC-245fa
Blend Technology in the Construction Industry

MARY BOGDAN and DAVID WILLIAMS

Honeywell
20 Peabody St.
Buffalo, New York 14210

A B S TR A C T IN T R O D U C T IO N

Energy conservation has become a key concern of Honeywell, previously AlliedSignal, has been dedicated
today’s consumer. Globally, insulation of buildings, homes to supplying blowing agents to the polyurethane foam
and offices to conserve energy has become critical. market for over 40 years. Per the Montreal Protocol
Fluorocarbon blown insulation has historically provided blowing agent technology will transition from the use of
the highest R-value per inch. HCFCs to zero ODP alternatives in the next several years.
Each nation has an established phaseout date for HCFCs.
Due to environmental concerns HCFC technology will The specific date of this transition varies from country to
soon be replaced with HFC technology. Honeywell is country.
dedicated to providing an HFC blowing agent to the Honeywell continues to be dedicated to providing
construction industry. It has dedicated many years blowing agents to the polyurethane industry. Honeywell
conducting research on HFC-245fa (1,1,1,3,3- has identified HFC-245fa (1,1,1,3,3-pentafluoropropane)
pentafluoropropane) as a replacement for HCFC-141b. as the third generation blowing agent. We feel that a HFC-
Honeywell is currently supplying semi-commercial 245fa technology provides the optimum foam performance
quantities of HFC-245fa to the construction industry for to the construction market.
industrial scale evaluations and field trials. Construction of However, today’s consumer is not only concerned about
the commercial facility is underway and is targeted to be energy savings but also cost when specifying insulation.
operational mid - 2002. The use of HFC ­245fa blend technology allows an
insulation manufacturer to design the insulation sold to
There are a variety of foam applications in the meet the needs of the insulation market. This technology
construction industry. Each has specific product allows the manufacturer to meet a variety of market
requirements. In some applications the R-value of the demands with a quality nonflammable alternative.
insulation is critical. In other applications the insulation
value is not as critical as the cost. Honeywell intends to H FC -245fa O V E R V IE W
supply HFC-245fa technology to meet all of these needs.
HFC-245fa technology provides insulation with the highest HFC-245fa is chemically identified as 1,1,1,3,3-
R-value. HFC-245fa blend technology makes it possible pentafluoropropane. Over the last 5 years, we have
produce foams with the optimal balance between blowing published a series of papers, which highlighed our work
agent performance and cost. with HFC -245fa in the construction industry. (1) To give
you a quick summary, HFC -245fa is a non-ozone depleting
This paper will discuss the progress made developing compound with an acceptable global warming potential
HFC-245fa and HFC-245fa blends in a variety of (GWP). It is a not considered a volatile organic compound
construction applications. It will include a discussion of (VOC). HFC -245fa has no long-lived atmospheric
laboratory and field evaluations vs. current HCFC -14lb breakdown products. It has low toxicity and is a non­
technology. flammable material with a boiling point of 59.5°F
(15.3°C). A summary of this data is contained in Table 1.

,
POLYURETHANES EXPO 2001 SEPTEMBER 30-OCTOBER 3, 2001 115
 Table 1: Physical Properties of HFC-245fa

Molecular Formula............CF3CH2CHF2
Molecular W eight........................... 134.0
Boiling Point (°F )............................59.5
(°C).............................15.3 Table 2: Regulatory and Environmental
Liquid Density (g/cc@ 20°C )..........1.32 Information on HFC -245fa
Freezing Point (°C )....................... <-160
Vapor Pressure: CAS Number.................................... 460-73-1
(PSIA @ 68°F )........................... 17.8 Ozone Depletion Potential...........................0
Global Warming Potential...................... 990
(kPa @ 20°C )............................... 123
US VOC status.................................. Exempt
Vapor Thermal Conductivity @ 40°C
Exposure guideline
(BTU in / ft2 hr°F)...................... 0.097
ACGIH TLV..................................... None
(m W /m ° K )............................... 14.0 OSHA P E L ...................................... None
Water Solubility (in HFC-245fa) ..1600 ppm WEEL (AIHA) TWA 8 hrs..... 300 ppm
Flash Point * ...................................... None US DOT Hazard C lass............. Not regulated
Vapor Flame Limits * * ....................None R C R A ................................. Not RCRA Waste
TSCA Inventory Status......................... Listed
*Flashpoint by ASTM D 3828-87; ASTM SNAP Approval....... All Foam Applications
D1310-86
**Flame Limits measured at ambient
temperature and pressure using ASTM
E681-85 with electrically heated match
ignition, spark ignition andfused wire
ignition; ambient air.

T O X IC IT Y M IS C IB IL IT Y

HFC -245fa is currently listed on the US EPA TSCA The miscibility of a blowing agent in the polyol is
Inventory, the European EINECS Inventory, and the critical in any foam application. HFC -245fa has exhibited
Japanese MITI Inventory. Extensive toxicity testing acceptable miscibility in a wide range of polyols and
indicates that HFC-245fa is of low toxicity. Overall isocyanates. We continue to add information to our list as
results from a series of genetic studies indicate that HFC- customers request it. Table 3 contains the most recent
245fa is non-mutagenic. It was also not a teratogen. The information on HFC -245fa miscibility in polyols and
American Industrial Hygiene Association has established isocyanates.
a Workplace Environmental Exposure Level (WEEL) of Note: To determine miscibility a mixture containing 40
300 ppm. Anyone who uses or handles HFC-245fa wt. % HFC-245fa and 60 wt.% polyol is prepared in a
should carefully review the MSDS and product label prior calibrated miscibility tube. The mixture is thoroughly
to use. A summary of this data is in Table 2. mixed at an elevated temperature. The tube is then placed
in a constant temperature bath for 24 hours. If separation
occurred, the height of the blowing agent was measured
and the amount dissolved in the polyol was calculated.

1 1 6 / Bogdan, Williams
 Table 3 Miscibility of H F C - 2 4 5 fa in P olyols@ 70 ° F (21 ° C)
P olyol % P o lyo l %
M iscible M iscible
POLYETHERS
(S u c ro s e )
Voranol 360 a >40.0
(S u c ro s e - A m in e )
Rubinol R 170° |I____ >40 I| Rubinol P 180° |I >40
(A ro m a tic - A m in e )
Voranol R 350Xa 24.3 Rubinol R 144° >40
Voranol R-470X3 36.8 Rubinol R159° 21
(T D A )

Pluracol - 8 2 4 0 35.4 | Voranol 4 9 0 a | >40

PO LYESTER
Terate 2541° 23.3 Terate 5521 0 23.0
Terate 2541 L° 27.9 Terate 254° 23.4
Terate 2031 0 18.8 Stepanol 2352® 32.3
Terate 2542 d 21.5 PHT 4 Diol' 6.2
MDI
Luprinate M20Sc |I 15.0 Luprinate M70R c |I 10.0
a Dow Chemical ° Huntsmanc BASF a KOSA “Stepan 'Great Lakes

VAPOR PRESSURE From Figure 1 it is obvious that HFC -245fa does not
have the same processing pressure requirements needed for
The boiling point and the vapor pressure of a blowing HFC-134a, HCFC-22 or CFC-12. It does however have a
agent can have a significant impact on the storage and higher vapor pressure than CFC-11 or HCFC-141b.
handling of the blowing agent, as well as preparation, Therefore process and shipping HFC -245fa systems may
storage and transportation of the polyol blend and require minor modifications to equipment and packaging.
manufacture of the polyurethane foam. Although HFC- In order to accurately address any processing, equipment,
245fa is an ideal HCFC-141b alternative meeting the needs or packaging issues one must know the vapor pressure of
of the rigid polyurethane foam industry, it does have a the polyol blend.
higher vapor pressure and lower boiling point than HCFC- In previous papers we have discussed the impact polyol
141b. It is not however a gaseous material. This is evident selection and blowing agent concentration has on blend
when one compares the vapor pressure differences in vapor pressure (1). A summary of the most recent
Figure 1 between HFC-245fa and those of HCFC-22 and information on vapor pressure vs. HFC-245fa
HFC-134a. concentration in polyols is contained in Table 4.

Figure 1 Vapor Pressure (PSIA) vs Temperature

Bogdan, Williams / 117

 T a b le 4 : V a p o r P r e s s u r e o f P o ly o ls
Polyol V orano l T e ra te T ero l V o ra n o l I V orano l V o ran o l V o ran o l
800a 203b 352b 270* I 350X* 370“ 470Xa
V a p o r P res s u re (psi<a) ® so ° c I
W e ig h t
%
H F C - 2 4 5 fa
0 5 6 5 5 6 5 9
5 10.5 16 - 8 9.5 6 11
10 19 30 19 13 16 13 20
15 - - - 17 23 - 23
20 29.5 - 33 - - 19 -

a Dow Chemical °KOSA

temperature. Insulation applications occur at many

As previously reported, this information is particularly different mean application temperatures. For instance the
significant for optimization of polyol systems. When one mean application temperature refrigerator foam may be
is formulating a polyol blend, minimization of vapor exposed to is 50 ° F (75 °F home vs. 35 °F refrigerator
pressure minimizes blowing agent lost through vapor interior). In a construction application in the northeast US
liquid equilibrium during processing and the foaming in the winter the mean temperature may be as low as 40° F.
reaction. This is of particular concern for spray foam (70° F home interior vs. 10° F outdoor air temperature)
polyol blends since these materials are shipped and stored Fluorocarbons particularly HFC -245fa perform ideally at
in 55-gallon drums. low temperatures for many reasons. One of the easiest to
Through optimization of polyol blend vapor pressure it understand is boiling point /condensation point. As foam
is possible to use current drums and equipment with HFC- is exposed to heating and cooling conditions in the outdoor
245fa /C 0 2(water) spray foam systems. Not only can the environment the gaseous blowing agent in the cell is
vapor pressure of the system be lowered, it can be done condensed and evaporated based upon the boiling
with minimal effect to foam quality and k-factors. point/condensation point of the blowing agent present.
This is illustrated Figure 2 which shows the relative
K-FACTOR VS. TEMPERATURE differences between polyiso insulation blown with HFC-
245fa, HCFC-141b, cyclopentane/isopentane blend and
One of the insulation purchasing factors for many isopentane. If one compares the boiling point of the
consumer applications is insulation value. Optimum blowing agents contained in Table 7 vs. data in Figure 2
insulation performance and energy efficiency is achieved the effect of condensation on the k-factor performance of a
through selection of blowing agent by mean application foam system is apparent.

Figure 2 Relative k-factor vs HCFC-141b

vs Mean Foam Tem perature Polyiso Insulation

118 / Bogdan, Williams

 Table 7 Boiling Point o f Blowing Agents
B oiling P oin t 8F °c
H C F C -1 4 1 b 8 9 .7 32
C y c lo p e n ta n e 120.7 49
Is o p e n ta n e 9 6 .8 36.0
H F C - 2 4 5 fa 59.5 15.3

There is however a fine distinction between the boiling environment. The focus of this section is to discuss the
point and the processability of a blowing agent. Although recent test results with HFC -245fa technology in
gaseous blowing agents should also have good low commercial environments using existing equipment. This
temperature k-factors, the vapor pressure and solubility information is primarily a summary of field results
considerations in processing the foam with a gaseous reported by systems houses based upon trials with their
blowing agent weaken its low temperature performance as customers.
seen in Figure 3. Although there is a linear relationship
between temperature and k-factor the miscibility of HFC- N O R T H A M E R IC A N SPRAY F O A M R O O F IN G
134a makes its overall performance poorer. The boiling T R IA L S
point of HFC -245fa makes it an ideal blowing agent. It has
the lower boiling point and higher vapor pressure than There have been a number of roofing trials conducted in
other liquid alternative, which reduces condensation upon the US using HFC -245fa/C02(water) technology. All
cooling in a foam cell. It also has a relatively high boiling systems were shipped in standard shipping containers.
point vs. gaseous blowing agents, which reduces the They were sprayed with existing equipment with the
fugitive emissions vs. a gaseous material and has better contractors on site. There were no special precautions
miscibility in polyols. necessary with the spray of HFC -245fa systems that were
not used when spraying the HCFC-141b systems. In fact
F IE L D T R IA L IN F O R M A T IO N in one of the trials the roof surface temperature at time of
spraying was 105 F. Table 6 summarizes the results
Historically there have always been products, which obtained by two US system houses using HFC -245fa/C02
perform well in a laboratory environment but for one (water) technology.
reason or another do not work in a commercial production

Figure 3 Comparison o f Temperature vs. k - factor vs. Biowing A gent in a Polyurethane Formulation

Bogdan, Williams / 119

 Table 6 Summary of Spray Foam Field Trial Evaluations

Parameter/ System House Resin Technologies (6) BASF

Technology Used HFC-245fa/C02(water) technology HFC-245fa/C02(water) technology
System Shipping Container Standard 55 gal drum Standard 55 gal drum
Trial Location CA CA, AZ (2), GA,
Equipment Standard used with HCFC-141b Standard used with HCFC-141b
systems systems
Foam Quality
• Core Density 2.99 pcf 3.00 pcf
• 6 Month Aged k-factor Equivalent to HCFC-141b Equivalent to HCFC­141b
• Compressive strength 48.6 psi > 50 psi
• Foam surface Smooth/ orange peal, no fisheyes Smooth/ orange peal, no fisheyes
Age of roofs/ Aged quality >1 year/ excellent < 1 year

The results of the spray foam field evaluations are very US POUR IN PLACE TRIAL
positive. The HFC -245fa/C02 (water) technology has
allowed the industry to spray HFC-245fa systems without This trial was also conducted by Resin Technologies using
any significant modification to equipment, or shipping an HFC -245fa system identified as RT-5075. (9) It is
containers. Just as important the foam sprayed has been commercial cavity fill pour foam. Again this system was
equivalent or superior in quality to the HCFC-141b foam shipped in a standard 55 gal drum. The drum showed no
currently sprayed in the market. Work will continue in this signs of bulging and was not under excessive pressure
market to optimize the formulations and evaluate the long­ when opened. The trial was conducted under the
term field performance. manufacturing conditions outlined in Table 7.
The foam processed well. There were no voids noted in
the part molded. A cross section of the finished article is
displayed in Photo 1.
The physical properties of the foam are summarized in
Table 8.

Table 7 Trial M anufacturing Conditions

Foam Machine Gusmer H6
High Pressure
60 lb/ min
Gun Gusmer AR “C”
Impingers 66 C-l-C
Shot Size 2.4 lb
Substrate Treated Polyethylene
Foam Profile Standard Pour Rise and Cure
Fluid Pressure 1400-1600 psi
Temperature MDI: 85° F
Resin: 85° F
Hose: 85° F
Demold Time 16 Minutes

1 2 0 / Bogdan, Williams
 P h oto 1: C om m ercial c a v ity fill
p o u r foam
T able 8: F oam P h y sic a l P ro p erties T able 9 G en eric J a p a n e s e
Property US Metric S p ra y F oam S y s te m
Core Density 2.2-2.6 35-42 COMPONENT PBW
pcf kg
TEROL* 80.00
Compressive 41.8 288 k POLYETHER 20.00
(parallel) psi Pa/m2 ADDITIVE 20.00
SURFACTANT 1.50
Compressive 22.8 157k CAT 1 2.00
(perpendicular) psi Pa/m2 CAT 2 1.50
CAT 3 1.50
“FT value 6.3
WATER 1.65
k-Factor 0.158
BTU in/hr ft2°F HFC-245fa 32.50
@ 77°F
Iso-
Dimensional Stability( Vol %) Luprinate M 489**
Cold Temp -0.3 -0.3 Index 2.01
(-20°C/28 Day)
DryHeat >1.0 >1.0 *Made by Oxid Chemicals
(70°C/28 Days) ** Made by BASF

In summary this was a successful trial. The foam research program with Oxid, we have conducted some
systems were shipped in standard containers. The material initial evaluations of a generic Japanese spray foam
was processed through existing equipment. The foam formulation. It is important to note that this is not an
produced has superior physical properties in comparison to optimized system. It represents very preliminary work on
the current HCFC-141b system and had comparable k- this application.
factors to the existing HCFC-141b system The generic formulation used in this study is contained
in Table 9.
JAPANESE SPRAY FOAM As with any spray foam system vapor pressure of the
polyol blend is important when one considers packaging
The Japanese spray foam market is significantly and shipping of the polyol blend. Figure 4 illustrates the
different from the North American spray foam market. vapor pressure of the polyol blend at various temperatures
The market is primarily wall insulation. The construction is in comparison to HFC -245fa neat and the vapor pressure
primarily concrete. The system is traditionally higher index limit of a 16-gauge drum.
than that used in the US and a lower density. In a joint

Figure 4: Vapor P ressure vs Tem perature Generic Jap an ese Foam

Formulation

Bogdan, Williams / 121

 Table 10 J a p a n e s e S p ra y F oam P ro c e ssin g
C onditions
Equipment Gusmer H2000
Resin:MDI pump ratio 1:1
Pressure MDI: 1300 psi
Table 11: P h ysica l P ro p erties
Pressure Resin: 1000 psi
J a p a n e s e S p ra y F oam
Hose MDI: 100 °F/Resin: 120 °F
Temperatures
Physical Properties Data
Preheater 120°F Overall density, pcf 1.79
temperature with skin on
Gun Gusmer GX7 Core density, pcf 1.79
Module MDI.028/Resin.023
2 ” lifts sprayed on cardboard

consistency. Photo 3 shows the pass line between passes

The foam was prepared using the equipment listed in of the foam. There is excellent interlayer adhesion on
Table 10. double passes.
There was no problem processing the Japanese spray Table 11 Summarizes initial test data on the foam
foam system. Photo 2 shows the skin and process quality sprayed. Figure 5 displays the initial k-factor vs.
of the spray foam. The skin is a rough orange peel temperature information on the foam.

Figure 5 HFC -245fa/C02 (water)

1 2 2 / Bogdan, Williams
 Although the system tested is not optimal this initial 4. D.J. Williams, M.C. Bogdan, R.C. Parker and G.M.
work is very encouraging. The Japanese system did Knopeck . “Update on the Development of HFC -245fa as
process well. It did not have an extreme vapor pressure a Liquid HFC Blowing Agent” presented at Polyurethane
indicating, it may not require any special high-pressure 1995, September 26-29, 1995
packaging. The foam quality was good and the system can 5. M.C. Bogdan, D.J. Williams, P.B. Logsdon, and R.C.
be processed in existing foam equipment. Parker. “Status Report on the Development of HFC -245fa
as a Blowing Agent” presented at Polyurethanes Expo’96,
CONCLUSION October 20-23, 1996
6. H.P. Doerge. “ New Developments in HFC-245fa
HFC-245fa is a nonflammable, non-VOC alternative. It Appliance Foam” presented at Polyurethanes World
produces foams with k-factors equivalent to current Congress’97, September 29-October 1,1997
industry standards. The foam produced has superior k- 7. M.C. Bogdan, and D.J. Williams HFC -245fa Systems
factor aging to current HCFC-141b systems. They also Co-blown with Water (C 02):Quality, Cost Effective HFC
have superior k-factors to other alternatives at a wide range Alternative for Construction Foam Applications presented
over a wide range of temperatures. at Polyurethanes Conference 2000, October8 -l 1,2000
HFC-245fa and HFC-245fa blends technology has been 8. Daniels, Farrington, Alberty, Robert A. 1975. Physical
proven to work well in both field trials and lab evaluations. Chemistry: John Wiley & Sons, Inc., pp.l 19
Therefore, HFC-245fa and HFC-245fa blends 9. Website www.Resintechnologies.com
technology provides the foam construction industry with a
solution for all markets. HFC -245fa blown foams meet the ACKNOWLEDGEMENTS:
critical requirement for insulation sensitive applications
while the HFC-245fa blend technology provides the The authors would like to thank the following
answer for the more cost driven markets. individuals and firms who without their assistance this
paper would not have been possible
DISCLAIMER • Jim Doose and Resin Technology
• BASF
All statements, information and data given herein are • Oxid LTD Chemicals
believed to be accurate and reliable but are presented
• Les Bement, Bin Zhong & Ron Riegal and all the
without guaranty, warranty or responsibility of any kind,
support groups at the Honeywell Research Laboratory in
expressed or implied. Statements or suggestions
Buffalo, NY
concerning possible use of our products are made without
representation or warranty that any such use is free of
patent infringement and are not recommendations to
infringe any patent. The user should not assume that all
safety measures are indicated, or that other measures may BIOGRAPHIES
not be required.
M.C. Bogdan
REFERENCES Mary Bogdan holds a BS degree in
Chemistry and an MBA from
1. M.C. Bogdan, D.J. Williams, and P. Verbiest. Canisius College. Since joining
Utilization of HFC -245fa in Spray Foam” presented at Honeywell in 1989, she has been
Polyurethanes Expo’98, September 17-20, 1998 involved in a variety of fluorocarbon
2. M.C. Bogdan, R.C. Parker and D.J. Williams. and fluorine specialty research
“Utilization of HFC-245fa in the Construction Industry” programs. She has been a Group
presented at Polyurethanes World Congress 1997, Leader in the Analytical Services Group. Currently, Mrs.
September 29-October 1,1997 Bogdan is a Principal Scientist in the Genetron ®Products
3. G.M. Knopeck, R.C. Parker, R.G. Richard and Technical Service and Applications group at Honeywell’s
I.R.Shankland. “Status Report on the Development of a Buffalo Research Laboratory. She has global
Liquid HFC Blowing Agent” presented at Polyurethanes responsibility for construction and systems applications.
94, October 9-12, 1994

Bogdan, Williams / 123

D. J. Williams
Dave holds a B.S. degree in
Chemistry from The University of
New Haven, New Haven, CT, USA.
From 1975 to 1994, he worked as a
Research and Development Chemist
and Technical Service Representative
for the Upjohn Company and later
The Dow Chemical Company in a wide variety of rigid
polyurethane and polystyrene foam application areas.
Dave joined AlliedSignal (now Honeywell) in 1994 and is
currently the manager of the Blowing Agent Technical
Service and Development Group. In this capacity, he is
responsible for technical service and product development
of Honeywell’s line of CFC, HCFC, and HFC foam
blowing agents. Dave is a member of the UNEP Flexible
and Rigid Foam Technical Options Committee.

124 / Bogdan, Williams

Pour-in-place Discontinuous Panels Using Pentanes
Al D eL eon David S h ieh
OxidLP . Oxid L.P.
101 Concrete Street 101 Concrete Street
Houston, TX 77012 Houston, TX 77012

E. F. Feske
Albemarle Corporation
P. O. Box 14799
Baton Rouge, LA 70898

ABSTRACT The use of R-134a in pour-in-place systems have been

demonstrated and several commercial pour-in-place, class
The conversion from R -14IB to suitable alternate I rated and non-rated systems are available commercially.
blowing agents in the rigid foam markets has accelerated Usually this blowing agent is used in conjunction with
as the deadline for its disappearance nears. Pentanes have water.
been selected by the flexible faced laminate industry to The use of hydrocarbons by system houses to replace R-
replace R -141B. Plant conversions are underway. Other 14IB has been lagging because of flammability concerns.
foam applications are considering pentanes. One such Because of this, the use of hydrocarbons by system houses
large application is the metal faced panels using pour-in- could be limited. This paper will discuss using pentane as
place systems. a blowing agent in class I rated pour-in-place systems. We
The work reported here will show that if properly will show foam formulations and E -84 tunnel test. We
formulated, systems containing pentanes can be used in will also discuss new polyester polyols developed and their
this application. The paper will show several fire retardant application.
combinations needed to reach the desired results in the E-
84 Tunnel Test. BACKGROUND

INTRODUCTION Unmodified aromatic polyester polyols whether based on

PET, DMT bottoms, pure terephthalic acid or phthalic
As the end o f the year 2002 nears, the urethane foam anhydride have poor pentane solubility. Cyclopentane is
industry is yet faced with another challenge. What would only slightly soluble in these polyols while isopentane or
be the replacement for R -14IB as the blowing agent? n-pentane have little or none. Table I shows the pentane
The answer is becoming clearer as this day approaches. solubility of commonly found aromatic polyester polyols.
In the US, the flexible faced laminators are converting to Pentane solubility is expressed as parts per hundred parts
pentane. It appears that a blend of cyclopentane and of polyol (PPHP). In order to use these materials one must
isopentane is favored at this time. However, n-pentane is depend on the addition of emulsifiers or solubilizers.
being evaluated because it offers the best economics of all The selection of an emulsifier can be difficult, as they
the pentanes. are system specific. Emulsion stability is dictated by the
The majority o f the US appliance manufacturers will application. In the production of flexible faced laminate
switch to 245FA, while R-134a will be used by one board where pentane is introduced via in-line blending,
producer. emulsion stability needs to be only a few minutes. If a
The US System Houses are faced with a more polyol/pentane blend is to be stored or shipped, the
complicated conversion. The choices are R -245fa, R- emulsion stability needs to be several weeks. The use of
134a, water, hydrocarbons and a combination of these. emulsifiers limits the use of pentane in pour-in-place
Work with R -245fa has progressed well with Honeywell. systems because of the high probability of losing pentane.
Pour-in-place systems with this blowing agent have been This can be significant as the polyol/pentane blend is
developed. The challenge with R -245fa in these systems is exposed to elevated temperatures.
balancing the ratio o f water/R - 245fa to get optimum Table 2 shows new polyester polyols with pentane
properties. solubility. This was accomplished by modifying the
polyester polyol backbone with pentane soluble materials.

,
POLYURETHANES EXPO 2001 SEPTEMBER 30-OCTOBER 3, 2001 125
Table 1. Pentane Solubility of Standard Polyols Table 4. U.S. Lamination Formulation
Pentanei Solubility, pphpi Ingredient Name EQ WT PBW
Polyol n -Pentane Isopentane Cyclopentane
Terol 256 2 4 2 Terol Terol 588 235 100
Terol 352 2 4 2 Fire Retardant TCPP 10
Terol 198 3 4 5 Surfactant B-8443 2
Terate 203 0 1 2 CAT 1 Koctoate 3.6
S-3152 0 1 2 CAT 2 Kacetate 0.85
CAT 3 PC5 0.25
Water 9 0.5
Table 2. Pentane Solubility in N e w Polyols
Exxol (60/40)
Pentane Solubility, pphp Cyco/lso 18.9
Polyol n -Pentane Isopentane Cyclopentane Total 136.10
Terol 595 10 12 15 Index 3.000
Iso Total E-489 136 196.00
Terol 588 15 14 15
Core Density 1.78
Terol 681 13 13 15 Overall Density 1.9

Comp. St. 20.4

These polyols still maintain high aromatic content and
yield good foam physical properties. E-84 Tunnel Test
Terol 588 was introduced early in the year 2000 for use Flame spread index 27
in continuous flexible faced lamination. Table 3 and 4 Smoke 64
show typical formulations used commercially in the US
K-factor
and Europe. Table 3 show typical European formulations 0.149
Initial
where the pentane used is n-pentane. Table 4 shows a 1 month 0.158
typical American formulation using Exxsol 1200 as the 2 month 0.160
blowing agent. Both formulations yield foam boards that 3 month 0.158
meet or exceed all specifications.
Terol 588 has sufficient pentane solubility for use in PUR
Table 3. European Lamination Formulation or PIR systems, but its functionality is too low to achieve
Name EQWT PBW the required foam physical properties. Terol 681 has the
same pentane solubility as Terol 588, but has a much
Polyester Terol 588 245.50 85.75
higher functionality.
FR Additive 1 TCPP 12.50
Surfactant 2 B8443 1.75 The data presented in this paper was generated using
CAT 1 K Acetate 50.00 2.00 Terol 681 in pour-in-place systems aimed at the
CAT 2 Koctoate 196.15 2.00 discontinuous metal-faced panel application. Specifically,
CAT 3 PMDTA 0.25 the applications where E-84 class I requirements are
Water 9.00 0.70 needed.
N -Pentane 18.00
Total 125.95 E X P E R IM E N T A L W O R K
Index Mondur 489 134.00 3.300
Iso Total 212.00 The initial part of this study was conducted in the
B/A 0.580
10 laboratory using the hand-batch method. The screening
Cream
Gel 31 test used to predict the E-84 tunnel test performance was
T.F. 45 the Cone Calorimeter test developed by Albemarle
Rise 65 Corporation and presented at the 1994 SPI Conference1.
Core Density 1.88 Large foam samples were generated and tested in the E -84
Comp. St. tunnel test at Omega Point Laboratories in San Antonio,
perpendicular, psi 27 Texas.
parallel, psi 18

1 2 6 / DeLeon
RESULTS AND DISCUSSION Table 6. Screening Formulation for E-84 test
Name EQ WT E-6 E-7 E-8
Table 5 shows a formulation combining Terol 681, a Polyester Terol 681 233 36 36 36
sucrose based polyether polyol, and a combination of RB- Polyether Sucrose 151 39 39 39
79 a tetrabromo phthalic anhydride based polyether and Fire Retardant 1 RB-79 249 35 35 35
Fire Retardant 2 Fyrol CEF 25 30 35
Fyrol CEF, a tris, chloroethyl phosphate, a liquid non- Surfactant B8443 2 2 2
reactive fire retardant. The objective was to obtain the CAT 1 PC5 0.2 0.2 0.2
lowest possible flame spread and smoke numbers as CAT 2 TMR-3 2.76 2.8 2.85
predicted by the Cone Calorimeter. At a fixed level of CAT 3 PC-8 0.1 0.1 0.1
fyrol CEF, increasing the RB-79 from 25 to 35 PPHP Water 9 0.7 0.7 0.70
accomplished this. This is the equivalent to increasing the Exxsol 1200*** 22.4 22.8 23.2
Bromine content from 3.6% to 4.8% at essentially a fixed Total 163.2 168.6 174.0
chlorine and phosphorous content
Index 1.80 1.80 1.80
Isocyanate E-489 134 175 175. 175
Table 5. Screening Formulation for E-84 test
Core Density,
Name EQ WT E-2 E-5 E-6 LB/CUFT 1.9 1.9 1.9
Cone Calorimeter
Polyester Terol 681 233 36 36 36 Projections
Polyether Sucrose/PO 151 39 39 39 Flame Spread 23.3 24.2 25.8
Fire Retardant 1 RB-79 249 25 30 35 Smoke Density 735 763 819
Fire Retardant 2 Fyrol CEF 25 25 25 %Chlorine 2.75 3.20 3.68
Surfactant B-8443 2 2 2 %Bromine 4.83 4.69 4.61
CAT 1 PC5 0.2 0.2 0.2 %Phosphorous 0.81 0.94 1.08
CAT 2 TMR-3 2.6 2.7 2.76
CAT 3 PC-8 0.1 0.1 0.1
Water 9 0.7 0.7 0.7 It is well known that PIR foams are more thermally
Exxsol 1200*** 20.77 21.6 22.4 stable than PUR foams. Typically, PIR foams display low
Total 151.4 157.0 163.2 flame spread and smoke in the E-84 tunnel test with little
or no fire retardants. Table 7-1 shows the E-84 tunnel test
Index 1.80 1.80 1.80 results for two levels of fire retardants. These foams were
Isocyanate E-489 134 165 170 170 made using Terol 588.
In Table 7 we explore the effect of increasing the index.
Core Density, LB/CUFT 1.9 1.9 1.9
Here the comparison is between a 2.50 index foam and a
Cone Calorimeter 1.85 index foam. The lower index foam contains a higher
Projections level of fire retardant and yet the effect of a higher index is
Flame Spread 23.3 26.5 23.3 overwhelming in reducing flame spread and smoke.
Smoke Density 854 798 735 Increasing the loading of polyester, reducing the level of
polyether produced foams with the lowest smoke ratings.
%Chlorine 2.90 2.80 2.75 Table 8 shows these results. At the same index and same
%Bromine 3.63 4.22 4.83 fire retardant loading, reducing the polyether content
%Phosphorous 0.85 0.82 0.81 reduces smoke significantly with little change in flame
spread.
Table 6 shows the effect of increasing Fyrol CEF at a
fixed level of RB-79. Here, there is a noticeable negative In table 9, we combine the highest level of bromine,
effect on smoke generation with a slight increase in flame constant level of phosphorous from three different fire
spread. At fixed % bromine in the foam, increasing the % retardant sources and high index. These foams were
phosphorous from 0.8 to 1.1 increase smoke and flame collected from large buns and tested in the E-84 tunnel test
spread at 4 inches in thickness. The sample with the highest
index yielded the lowest smoke. All three samples fell in
Table7-1. E-84 Tunnel Test Results, PIR Foams the 25-30 flame spread range.
Polyester Polyol Terol 588 Terol 588
Fire Retardant TCPP TCPP CONCLUSION
Fire Retardant pbw 5 10
Index 3.0 2.5
E-84 Tunnel Test We believe that it is possible to develop a system for
Flame Spread 35 40 discontinuous panels using pentane to meet Class I E-84
Smoke 50 110 tunnel test rating.
Board Thickness, in 4.0 4.0

DeLeon / 127
 Table 7. Screening Formulation for E-84 test Table 9. Screening Formulation f o r E - 84 test
Name EQ WT E-13 E-15 Name E-18 E-19 E-20
Polyester Terol 681 233 40 40 Polyester Terol 681 60 60 60
Polyether Sucrose/PO 151 16 16 Fire Retardant 1 RB-79 40 40 40
Fire Retardant 1 RB-79 249 40 40 Fire Retardant 2 Fyrol CEF 0 26 0
Fire Retardant 3 TEP 15 15 Fire Retardant 3 TEP 17 0 0
Surfactant B8443 2.5 2.5 Fire Retardant 4 DEEP 0 0 16
CAT 1 PCS 0.2 0.2 Surfactant B8443 2.5 2.5 2.5
CAT 2 TMR-3 3.1 2.9 CAT 1 PC5 0.2 0.2 0.2
CAT 3 PC-8 0.1 0.1 CAT 2 TMR-3 2.9 2.9 2.9
Water 9.00 0.7 0.7 CAT 3 PC-8 0.1 0.1 0.1
Exxsol 1200*** 20.5 20.85 Water 0.7 0.7 0.7
Total 138.1 138.3 Exxsol 1200 16 18.3 15.9
Total 139.4 150.7 138.3
Index 2.50 1.85
Isocyanate E-489 134.00 185 151 Index 2.15 2.5 2.15
Isocyanate E-489 165 192 165
A&B@ 28 °C
Core Density,
Core Density, LB/CUFT 1.9 1.9 LB/CUFT 1.9 1.9 1.9
Cone Calorimeter
Cone Calorimeter Projections Projections
Flame Spread 25.8 22.1 Flame Spread 23.8 26 26.2
Smoke Density 676 868 Smoke Density 314 182 328

%Chlorine 0.00 0.00 E-84 Tunnel Test

%Bromine 5.69 6.36 Flame Spread 25.8 28.5 27.3
%Phosphorous 0.79 0.89 Smoke 450 250 400

Table 8. Screening Formulation for E-84 test To accomplish this, one must use a combination of
bromine from RB-79, phosphorous from three different
Name EQWT E-16 E-17
fire retardants, the lowest possible amount of polyether,
Polyester Terol 681 233 40 60 and the highest possible index.
Polyether Sucrose/PO 151 16 0 Would this combination result in a commercially viable
Fire Retardant 1 RB-79 249 40 40 system? We will be addressing this in future development
Fire Retardant 4 DEEP 15 15 work. This will include pouring panels in actual molds to
Surfactant B8443 2.5 2.5 determine processing characteristics, optimization of foam
CAT 1 PCS 0.2 .2 physicals, and optimization of the fire retardant package to
CAT 2 TMR-3 2.9 2.9 get the lowest possible smoke.
CAT 3 PC-8 0.1 0.1
The Cone Calorimeter is a reasonable tool to predict
Water 9.00 0.7 0.7
Exxsol 1200*** 15.5 16 flame spread in the E-84 tunnel test. However, smoke
Total 132.9 137.4 prediction from this test is less reliable. We will try to
improve this in future experiments.
Index 8.85 I.85
Isocyanate E-489 134.00 158 161 ACKNOWLEDGMENTS
Core Density,
LB/CUFT 1.9 1.9 The authors would like to acknowledge James Buis of
Cone Calorimeter Oxid’s R&D group for sample preparation and testing
Projections
and to Albemarle Corp for the Cone Calorimeter tests.
Flame Spread 24.0 25.6
Smoke Density 801 659
%Chlorine 0.00 0.00 REFERENCES
% Bromine 6.33 6.17
%Phosphorous 0.96 0.94 1. Dowling and Feske Proceedings of the Polyurethanes
Conference. Pages 357-363.

128 / DeLeon
BIOGRAPHIES as PUR/PIR foam development, using these polyols. Prior
to joining Oxid in 1990, David was a Research Chemist at
A1 DeLeon Chardonol, where he developed numerous polyester
polyols for use in foams. He holds several patents in the
A1 DeLeon is the VP of R & D. He began his career in field of Polyester polyol, and he is recognized as an
1970 working at Jim Walter Research Corp., where he was authority in this field. He holds a BS degree in Chemistry
involved in the development of isocyanurate chemistry. from the Catholic University of Taiwan, and graduate
After fourteen years there, he became Technical Director degrees from the University of Houston.
of Flexible Products Company where he stayed for four
years, directing their R & D as well as TS & D efforts, In
1988, A1 joined Oxid in his current position. He holds E. F. Feske
numerous patents in urethanes, and is the author of
numerous publications. A1 holds BS and MS degrees in Elbert (Bert) Feske received a BA degree in Science from
Chemistry from the University of Miami (FL). He is a Graceland College in Lamoni, Iowa. Bert has worked in
member of Sigma XI and ACS. all aspects of Polyurethane product development in the last
30 years. He joined Ethyl Corp. in 1988 and is currently
David Shieh an R & D Specialist at Albemerle Corp. Bert has
numerous technical papers in the area of Fire retardancy of
As Manager of R & D for Terol polyols, David Shieh is rigid urethane foams.
responsible for both polyester polyol development as well

DeLeon / 129
 SESSION E
INNOVATIVE POLYURETHANE SOLUTIONS FOR
AUTOMOTIVE APPLICATIONS

TECHNICAL SESSION ORGANIZERS

Jim Tobias, Chair

AIR PRODUCTS & CHEMICALS, INC

Jim Lambach
BAYER CORPORATION

Allan James
DOW CHEMICAL CANADA INC.

TECHNICAL SESSION MODERATOR

Tom Jensen
GENERAL MOTORS

TECHNICAL SESSION VICE - MODERATOR

Allan James
DOW CHEMICAL CANADA INC.
Innovative Polyurethane Seat Trim Cover Solution

ERIC KURTYCZ JOANNE BLADON

The Dow Chemical Company Dow Chemical Canada
1600 Harmon Road, 1086 Modeland Road
Auburn Hills, Michigan 48326 Sarnia, Ontario N7T 7K7
Canada
STEVEN ENGLISH
The Dow Chemical Company ANDREW HOGG
Midland, Michigan 48674 Dow Europe S.A.
Am Kronberger Hang 4
JOSE GODOY Schwalbach
Dow Europe S.A. Germany
13, rue de Veyrot
CH-1217 Meyrin 2/Geneva ALFREDO LARRE
Switzerland The Dow Chemical Company
2301 North Brazosport Blvd.
Freeport, Texas 77541

ABSTRACT passenger cars, niche and cross-over vehicles or any

vehicle platform interested in achieving a boldly
Innovative seating solutions that combine flexibility in differentiated seat system, and still realize cost savings.
first surface appearance, increased functionality, and cost The ability to recycle automotive seating is improved in
reduction opportunities have been developed by Dow for this one-constituent solution. Furniture applications can
automotive and non-automotive applications that will meet also greatly benefit from this technology. Explore the
your needs. For instance, current automotive seats are possibilities of this seating development with Dow.
typically manufactured either with polyurethane foam pads,
which are subsequently topped with a fabric, vinyl, or BACKGROUND AND OBJECTIVE
leather seat covering, or by molding seat foam with
separate covering materials in a pour-in-place operation. Flexible molded polyurethane foams continue to be the
Dow’s technology integrates the first surface material into material of choice for most forms of transportation seating.
the foam pad, eliminating costly cut-and-sew operations or Successful seating designs maintain their support and
coverstock pre-forms for pour-in-place operations. comfort performance over time and it is not uncommon for
This polyurethane application development enables polyurethane foam pads to still meet physical property
automotive seats with a supple, durable outer layer, which requirements at the end of a vehicle life [1 - 2].
exhibit excellent touch and grain definition, mold-in color Current automotive seats are typically manufactured
capability, and the ability to tune seat comfort, while either with polyurethane foam pads that are subsequently
supporting washable interior initiatives. Coupling this new topped with a fabric, vinyl, or leather seat covering, or by
material development with the seating design element, molding seat foam with separate covering materials in a
Dow’s technology unleashes the potential for lower profile pour-in-place operation. Dow recognized the need for
seating, and unique seating styles and aesthetics. This alternative, innovative seat trim cover solutions. Dow’s
combination is expected to be ideally suited for a wide technology integrates the first surface material into the
range of vehicles including trucks, sport utilities, mini vans, foam pad, eliminating costly cut-and-sew operations or

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 133

coverstock pre-forms for pour-in-place operations. This all-MDI foam was used, however, any typical seating foam
development enables automotive seats with a supple, could be used behind the elastomer layer including all-
durable outer layer, which exhibit excellent touch and grain MDI, TDI/MDI or all-TDI foam. The hardness was
definition, mold-in color capability, -and the ability to tune measured as CFD 40% according to DIN 53577 [3]
comfort, while supporting washable interior initiatives. (equivalent to ISO 3386 [4]) before aging as well as after
This technology gives enhanced design and styling heat aging a sample at 140 °C for 22 hours and humid
flexibility. This development enables easier recycling due aging another sample at 120 °C for 5 hours. The density of
to the all-polyurethane composition. The surface is also the composite in-mold coating/elastomer/ MDI flexible
easy to clean and repair. In essence, it is a perfect solution foam was measured according to DIN 53420 [5]
for today’s seating requirements since it is both innovative (equivalent to ASTM D 3574 [6].
and cost effective. Since this development involves an integral trim cover,
the adhesion of the two layers is very important. The Tear
EXPERIMENTAL CONDITIONS strength of the elastomer/foam interface was measured
according to Renault RP 1048-73 [7] where the elastomer
Foam Preparation was pulled away from the foam.
All interior components including seating foams must
We produced both prototype automotive seat foams and meet flammability requirements and thus, the flammability
foams for physical property testing. The foams were performance was measured according to FMVSS 302° [8].
manufactured on a Krauss-Maffei foam machine using two Durability or fatigue resistance in seating foams relates to
different molds. For the foam/elastomer composite how well the seating foam maintains its loadbearing
property testing, a square, 80 cm x 80 cm x 5 cm mold was capacity and dimensions following extended use. It is
used. Specially designed prototype automotive seat molds desirable to have minimum effects after extended use.
for both cushion and back pads with side bolsters were Foam is said to have undergone fatigue if it loses
used to manufacture the prototype seats. Two grains were loadbearing capacity or height after repeated use. This can
selected for the tool surfaces: a leather-like grain for the lead to a decrease in overall seat comfort and a shift in the
bolster areas and a snakeskin-like grain for the central driver position. A variety of laboratory and field tests are
seating areas. used to measure fatigue performance [9]. The Peugeot
The inside surfaces of the molds were kept at 65°C and dynamic fatigue test, Peugeot D 42 1047 [10], was selected
were sprayed with Kluber* 41-2013 mold release agent for our evaluation since it is a relatively stringent test that
prior to application of the in-mold coating system. The can be conducted on relatively small specimens. Pieces of
prototype evaluations and physical property evaluations the foam/elastomer composite were evaluated. This
were carried out using UV -stable in-mold coatings from Peugeot D 42 1047 [10] method was cycled for 200000
two different suppliers. During production of the prototype cycles of 25 - 75% compression at 3 Hz and both the
seats, the elastomeric formulation was sprayed on the side height loss and hardness loss were measured.
rolls and seating portion of the open mold using a low Compression set tests have also been used in the industry
pressure Unipre M 12 spray machine. The MDI-based for many years as a predictor of foam durability [11]. In
foam formulation was poured across the entire mold using this study, dry, wet and humid aged compression sets were
all measured The 50% dry compression set was measured
a mixing head mounted on a 5-axes ASEA^ 90 robot. The
according to Ford-ISO 1856-80 [12]. The 50% humid
mold was closed and the polyurethane article was allowed
aged compression set was tested according to DIN 53578-
to cure. After the appropriate curing time, the article was
demolded and crushed to break cells and avoid shrinkage 88 [13] with humid aging at 120 °C for 5 hours. The
of the part. During use, the in-mold coating layer would be Renault RP 1637 [14] method was used for 70% wet
the top surface on which an occupant would sit. compression set at 40 °C, 95% R.H. for 22 hours.
Fogging is critical for all automotive interiors since it is
Foam/Elastomer Composite Testing undesirable to have any build-up on the windshield interior.
The fogging test was conducted on the foam/elastomer/in-
We evaluated the properties of the foam/elastomer mold coating composite according to SAE J1756 [15] with
composite using typical requirements for automotive a 100 °C bath temperature and 21 °C cooling temperature
seating foam. Hardness is very important for the overall for 3 hours. These test conditions are similar to the Ford
comfort of automotive seating. Various geographical areas specified conditions.
show differing preferences for foam hardness. In general,
foams in Europe tend to be much harder that those in North
American automotive seating. For these evaluations,

° Results of small-scale flammability tests are not intended

* Trademark of Kluber-Chemie KG to reflect the hazards presented by this material or any
p Trademark of The ABB Group other material under actual fire conditions.

1 3 4 / Kurtycz, Bladon, English, Hogg, Godoy, Larre

Trim Cover Testing blotted and rubbed gently with white cleaning cloths wet
with various sequences of cleaners. The color difference
We evaluated the seat trim cover according to typical was measured with a BYK -Gardner Colorview
OEM seating trim tests. Following an OEM Spectrophotometer. It is desired that the cleaning cloth
recommendation, the elastomer was peeled off the foam should have a rating of at least 4 on the stain rating scale
layer and it was tested mainly against the GM 2737M and there should be no staining, discoloration, surface
supported vinyl specification [16]. The tensile strength and softening nor loss of grain when tested with the cleaners.
elongation were measured according to ASTM D5034-95 The water vapor transmission was measured on samples
Breaking Strength and Elongation of Textile Fabrics (Grab of the elastomer using the Association of the Non Woven
Test) [17]. The samples were taken from the elastomer Fabrics Industry (INDA) Standard test: 1ST 70.2 (95) [25].
peeled from the center snakeskin-grained area and from the The sample material was attached to a disposable
front leather-grained area on cushion samples in either aluminum dish with epoxy. Distilled water was injected
direction. Since the elastomer was sprayed, there should into the side of the dish with a syringe and the hole was
not be directional influences. The specimen thickness was plugged with epoxy. The rate of water vapor transmission
taken as the minimum thickness of the averages across the through the specimen from the water to the controlled
width at the middle and at each end. atmosphere was determined from weighing the dish at
The tear strength was measured by the ASTM D5587-96 regular intervals. The controlled atmosphere was 22 °C
Tearing Strength of Fabrics by Trapezoidal Procedure [18]. with a relative humidity of 50%. The weights were taken at
The specimen thickness was taken as the average of three intervals of 8 to 24 hours after the initial weighing.
measurements along the tear path. The average tearing The colorfastness to light was measured by Xenon arc
force was taken as the average of the 5 peak forces as aging according to SAE J1885 [26]. GM 2737M [16]
specified in the method for materials that exhibit a jagged requires ‘no more than a noticeable on-tone change in color
tearing force curve. or gloss and little or no plasticizer migration, increased
Seats must be able to withstand cold temperatures for pliability, loss of grain, appreciable wrinkling or any other
northern climates and one critical parameter is the undesirable effects’ after 225.6 kJ/m2 exposure and states
resistance to cold cracking. For both unaged and aged that ‘spectrocolorimeter readings are for information only.’
samples, we measured the resistance to cold cracking with GM 2617M Interior Trim Parts Performance [27] 4.3.3
an impact of 4 J at - 30 °C [19]. The aging sequence requires that ‘after exposure the 8E reading for plastic parts
involved two repeats of the following cycle: - 30 °C for 4 shall be no greater than 3.0 units when measured in the
hours, 38 °C, 100% R.H. for 16 hours, - 30 °C for 4 hours, same location on the sample before and after exposure.’
38 °C, 100% R.H. for 16 hours and 93 °C for 72 hours.
In order to ensure the trim cover can withstand the wear
typical of an automotive seat, the abrasion resistance was
measured according to ASTM D3884 [20] with CS10
abrader wheels and a 1000 gram load for 500 cycles.
Another test that predicts the wear resistance is the Scuff
and Mar test. The resistance to scuffing and marring was
evaluated according to GM 9150P [21] with a 750 gram
weight. Both unaged samples and samples aged at 93 °C
for 24 hours were tested. The stretch and set test is more
applicable to a separate trim cover which could stretch and
sag with use, however, it was measured here to demonstrate
the integrity of the integral trim cover. Stretch and set was
conducted according to SAE J855 [22]. It is also very
important that the color integrity of the surface remains
constant. Color crock was evaluated using AATCC Test
Method 8 [23] under both dry and wet conditions as well as
after sample aging at 70 °C for 7 days. Resistance to
blocking of face-to-face samples was measured on both
aged and unaged specimens at 80 °C for 30 minutes with a
908 gram weight on top. The aging regimen for the
blocking test was 72 hours at 38 °C, 95% R.H. The soil and
cleaner resistance test was conducted according to a new
GM Draft method, GMW 3402TM [24]. Samples of
ketchup, soil, soy sauce, coffee, chocolate sauce, sewing
machine oil, petroleum jelly and ball point pen were F ig u re 1. Prototype S e a t A ssem bled on the Fram e
ground into the surface of the samples and then they were

Kurtycz, Bladon, English, Hogg, Godoy, Larre / 135

GM 2617M Interior Trim Parts Performance [27] 4.3.3 TEAR STRENGTH (PEEL-TYPE TEAR)
requires that ‘after exposure the 8E reading for plastic parts
shall be no greater than 3.0 units when measured in the The tear strength of the elastomer/foam interface was
same location on the sample before and after exposure.’ found to be 700 N/m compared to the Renault requirement
The seating requirement in Table 1A of GM 2617M [27] of 150 N/m tear strength for the foam itself.
says ‘per covering material spec’ which means it refers
back to GM 2737M [16] in the case of a vinyl trim cover. DURABILITY AND COMPRESSION SETS
The color was measured before and after aging at
exposures of 225.6 kJ/m2, 452 kJ/m2 and 677 kJ/m2. The dynamic fatigue testing according to Peugeot D 42
1047 [10] method gave a CFD 50% hardness loss of 15.3%
Seat Prototype Evaluation - Jounce & Squirm and and a height loss of 5.05% compared to the Peugeot
Ingress/Egress Testing requirements of 15% maximum hardness loss and 5%
maximum height loss respectively.
Both Jounce & Squirm and Ingress/Egress tests were The 50% dry compression set as measured according to
conducted on the prototype seats. The prototype cushion Ford-ISO 1856-80 [12] was 8.5% which meets the
and back pads were mounted on corresponding frames as requirement of less than 10%. The 50% humid aged
shown in Figure 1. The Jounce & Squirm test, GM CPC compression set was 30%, which meets the GM
1473 [28], was run for 300000 cycles. This test simulates requirement of 30%. The Renault 70% wet compression
the motion of the occupant during driving. The Ingress- set 8.9% Cd which is well within the specification limit of
Egress [29] test was run for 54000 cycles. This simulates 20% maximum.
the sliding entry into the vehicle. These test lengths
represent three times the life expectation of the car. The FLAMMABILITY
Ingress/Egress test method was relatively stringent using
modified truck seat test methodology. During the FMVSS 302 flammability test [8], the
samples had a rating of D.N.I. (did not ignite)0 . Many of
RESULTS AND DISCUSSIONS the OEM’s require a burn rate of less than 100 mm per
minute, so this is well within these requirements.
Evaluation of foam/elastomer composite according to
typical foam specifications FOGGING

Overall, there was excellent compatibility between the in­ Fogging results for the foam/elastomer/in-mold coating
mold coating, elastomer and flexible foam layers. There composite samples were 86.8% and 100% reflectance.
was no cracking, bubbling nor deterioration of the in-mold This compares well with the requirement of greater than
coating layer under any of the heat or humidity conditions 70% reflectance.
nor after the fatigue testing.
Evaluation of elastomer layer
DENSITY AND HARDNESS
TENSILE STRENGTH AND ELONGATION
The overall density of the composite in-mold
coating/elastomer/ MDI flexible foam was 94 kg/m3. The The tensile strength was 670 N and 695 N and the
density of the foam itself was 55 kg/m3. The CFD 40% [3] elongation values were 322 % and 550% for two of the
before aging was 14.7 kPa. This is a relatively high elastomers evaluated. This exceeds the GM construction
hardness, but was not unexpected since the elastomer layer specification WN1J for expanded vinyl coated knit [28],
imparts additional hardness to the foam and the foam used which has a tensile strength requirement of 400 N
for this composite was medium density MDI foam. After minimum in the warp and fill directions. The elongation
requirement in the GM construction specification WN1J
heat aging at 140 °C for 22 hours, the CFD 40% [3] was
[28] is 8 -25% for the warp direction and 50 - 72% for the
12.4 kPa which equates to a 15.3% force loss after heat
fill direction. Our elastomer samples have significantly
aging. After humid aging at 120 °C for 5 hours, the CFD
higher elongation than the vinyl requirement. This is not
40% [3] was 9.8 kPa, which represents a 33% force loss
an issue since the cover is integrated with the foam.
after humid aging. This matches well with the GM foam
requirement of 30 - 35 % maximum force loss after humid
aging.

° Results of small-scale flammability tests are not intended

to reflect the hazards presented by this material or any
other material under actual fire conditions.

1 3 6 / Kurtycz, Bladon, English, Hogg, Godoy, Larre

TEAR STRENGTH thicker elastomer samples met the specifications of GM
WN1J [28] for stretch and set.
The tear strength ranged from 38.7 N to 125.5 N for
elastomer layer thickness ranging from 1.8 - 5.2 mm. The RESISTANCE TO SCUFFING AND MARRING
tear strength was 83.5 N for the elastomer peeled from the
prototype cushions where the layer thickness 2.7 mm. As There was a slight marring on all samples tested
expected, a thicker elastomer layer gives better tear results. including unaged and aged samples. Vinyl specification
The GM WN1J [28] vinyl construction specification gives GM 2737M [16] requires no scuffing and only slight
minimum requirements of 24 N in the warp direction and marring is acceptable.
20 N in the fill direction. Therefore, all samples tested
exceed this requirement by a significant margin. COLOR CROCK (AATCC TM8 [23])

RESISTANCE TO COLD CRACKING The color crock rating was 5 (best) for all samples.
There was no color transfer to the cloths for dry and wet
No coating failure and no cracking was observed after testing of both unaged and aged samples. Vinyl
testing at - 30 C for 2 aged samples and 2 unaged samples. specification GM 2737M [16] requires no color transfer
Vinyl specification GM 2737M [16] requires no coating to the fabric after 10 strokes on the crockmeter.
failure or cracking.
BLOCKING
RESISTANCE TO WEAR (ABRASION RESISTANCE)
There was no blocking (sticking together when placed
The resistance to wear test results showed a weight loss face-to-face) at 80 °C of three pairs of unaged samples and
range from 0.01% to 0.08%. The appearance was quite three pairs of aged samples. The specimens did not stick to
poor for coating #1 after abrasion testing since the in-mold each other and no pulling was required to separate the
coating layer was very thin. The in-mold coating was worn pairs.
off the sample in as many as 15 places. For coating #2,
there was some loss of gloss due to the abrasion, but the SOIL AND CLEANER RESISTANCE
coating was intact. This minor loss of gloss is acceptable in
the OEM vinyl requirements. This shows the importance of The results of the Soil and Cleaner Resistance test are
the thickness of the in-mold coating. given in Figures 2 to 4 where three different in-mold
coatings were evaluated. For the soil resistance test, all
STRETCH AND SET three coatings showed excellent resistance to the soils
with color change ratings of 4-5 to 5 for all soils. In the
The stretch and set results are given in Table 1. The cleaner resistance test of coating #1, the naphtha stained
thicker elastomer samples, Samples #1 and #2, passed the the cleaning cloth to a 3 rating and all the other cleaners
warp stretch and warp and fill set requirements, but the had ratings of 4 to 4-5 (where 5 is best with no staining
allowable fill stretch is larger for the vinyl due to the knit and 1 is worst). For coating #2, all cleaners had ratings
fabrics used in their construction. This should not be an of 4 to 5 on the cloths. For coating #3, the rubbing
issue with the 'attached' elastomer in our design. Sample alcohol stained the cleaning cloth to a 2 - 3 rating. The
#3 was significantly thinner than the other two samples and other cleaners all had good ratings of 4 to 4-5 on the
the stretch was higher than the fill stretch requirement. The cloths.

Table 1. S tretch a n d S e t R e s u lts

SamDle Percent Stretch (% ) Percent Set (% ) Thickness (m m )

S am ple #1 13.2 0 .3 3 .3
Sam ple #2 18.5 0.3 3 .3
Sam ple #3 78.8 4.2 1.5
Vinyl requirem ent w arp® 8 -2 5 3 max.
Vinyl requirem ent fill® 5 0 -7 2 7 max.
® per G M Vinyl Construction specification W N 1 J .

Kurtycz, Bladon, English, Hogg, Godoy, Larre / 137

 4-5. For coating #1, the part color retention was moderate
with gray scale for color change ratings of 3-4 to 4-5.
Optimization of the in-mold coating should improve these
color retention results.

WATER VAPOR TRANSMISSION

The water vapor transmission properties were compared

to that measured for typical automotive seating vinyl and
automotive leather samples. The water vapor transmission
rate was highest for the leather and the water vapor
transmission rate of the elastomer was between that of the
leather and vinyl.
F ig u re 2. Soil Resistance o f Trim Layer

COLORFASTNESS TO LIGHT

The colorfastness to light was measured by Xenon arc

aging according to SAE J1885 [26] for coatings #1 and #3.
The color was measured before and after aging at
exposures of 225.6 kJ/m2, 452 kJ/m2 and 677 kJ/m2. Both
coating #1 and coating #3 passed the requirement of a 8E
of less than 3.0 units after exposure to 225.6 kJ/m2.
Coating #1 continued to show very little color change after
exposure up to 677 kJ/m2with 8E of less than 1.0 units.

Evaluation of Seat Prototypes - Jounce & Squirm and

Ingress/Egress Testing

The Jounce & Squirm testing showed only minor surface

F ig u re 3 . C lean er Resistance o f Trim Layer - Cloth Staining
cracks in the trench area in the in-mold coating layer and
no cracks in the elastomer layer after 300000 cycles. There
is a level of personal judgement in the interpretation of
whether the seat passes the test and this would be
considered excellent performance and a "pass" according
to a former Tier 1 seating engineer.

F ig u re 4. C lean er Resistance o f Trim Layer - Part Color

Retention

In the cleaner resistance test for coating #3, the part color
retention was good with gray scale for color change ratings
of 4 to 5 (where 5 rating is best with no color change and 1
is worst). For coating #2, the part color retention was
F ig u re 5. Xenon Arc W eathering Results
moderate with gray scale for color change ratings of 3-4 to

138 / Kurtycz, Bladon, English, Hogg, Godoy, Larre

 Figure 6. Jounce &Squirm after 300000 cycles

Ingress/Egress testing was conducted for 54000 cycles

according to a modified truck seat test profile. There was
some very minor wear at the rear of the seat cushion where
the hardware rubbed during the test. Some minor cracking
of the coating was visible in the bolster area of the cushion
and back, but no cracking of the elastomer was observed.
As interpreted by a former Tier 1 seating engineer, this is
excellent performance and would be considered a "pass".

Customer Focus Session

Figure 7. Ingress/Egress after 54000 cycles
A customer focus session was held early in the
development to gage consumer reaction to the innovative
trim cover. The benefit of the ability to clean this type of CONCLUSIONS
seat was recognized by consumers. The feel of the surface
was critical with the smoother samples with less drag being This Dow development enables automotive seats with a
preferred by the majority of customers. The prototype supple, durable outer layer, which exhibit excellent touch
seats will be showcased in a future customer focus session. and grain definition, mold-in color capability, and the

Kurtycz, Bladon, English, Hogg, Godoy, Larre / 139

ability to tune comfort, while supporting washable interior 9. Moore, R., G. Wiltz, R. Priester, D. Hunter and D.
initiatives. The elastomeric layer properties compare well Laycock, 1997. Flexible Polyurethane Foams, 2nd
with typical automotive seating trim requirements. Edition, R. Herrington and K. Hock eds., The Dow
Coupling this development with the seating design element, Chemical Company, pp. 8.1 - 8.15.
Dow’s technology unleashes the potential for lower profile 10. Peugeot D 42 1047, “Flexible Cellular Materials,
seating and unique seating designs, styles and aesthetics. Dynamic Fatigue by Compression at Fixed
This combination is expected to be ideally suited for Deformation,” 1997.
trucks, sport utilities, mini vans, passenger cars, niche and 11. Skorpenske, R.G., R. Solis, R.A. Kuklies, A.K.
cross-over vehicles or any vehicle platform interested in Schrock and R.B. Turner, 1992. “Compression Set
achieving a boldly differentiated seat system, and still Mechanisms In Flexible Polyurethane Foam,”
realize cost savings. Explore the possibilities of this seating Proceedings o f the 1992 SPI Polyurethanes
development with Dow. Conference, Lancaster, Pa.: Technomic Publishing
Co., Inc., pp. 650 - 659.
ACKNOWLEDGEMENTS 12. Ford-ISO 1856-80, “Flexible cellular polymeric
materials - Determination of Compression Set,” 1980.
The authors would like to express their sincere 13. DIN 53578-88, ‘Testing the Ageing Behaviour of
appreciation to Patrice Bocquel, Patrick Thevenon, Latex and Polyurethane Foam,” 1988 (superceded by
Antoine Storione and Jean-Luc Roser, Dow Europe S.A. DIN EN ISO 2440, “Flexible and Rigid Cellular
for their assistance in the development of this technology. Polymeric Materials - Accelerated Ageing Tests,”
2000 ).
REFERENCES 14. Renault RP 1637, “Remaining Deformation under
Humid Atmosphere, Flexible Cellular Materials,” May
1. Wilson, A. and G.R. Blair. 1994. “ Polyurethane 1981.
Automotive Cushioning In Car Durability vs. Foam 15. SAE J1756 DEC94, ‘Test Procedure to Determine the
Properties,” Proceedings of the 1994 SPI Fogging Characteristics of Interior Automotive
Polyurethanes Conference. Lancaster, PA: Technomic Materials,” December 1994.
Publishing Co., Inc., pp. 478 - 488. 16. GM 2737M Supported Vinyl specification, General
2. Blair, G.R. and A. Wilson, 1995. “Polyurethane Motors Engineering Standards, Materials and
Automotive Cushioning: Material Properties After In Processes Textiles, June 1999.
Car and Simulated Durability Testing,” Proceedings of 17. ASTM D5034-95 “Standard Test Method for Breaking
the 1995 SPI Polyurethanes Conference, Lancaster, Strength and Elongation of Textile Fabrics (Grab
Pa.: Technomic Publishing Co., Inc., pp. 413 - 417. Test),” July 1995.
3. DIN 53577, “Determination of Compression Stress 18. ASTM D5587-96, “Standard Test Method for Tearing
Value and Compression Stress-Strain Characteristic Strength of Fabrics by Trapezoidal Procedure,” April
for Flexible Cellular Materials Characteristics” 1996.
(superceded by DIN EN ISO 3386/1, “Flexible 19. GM 9140P, “Resistance to Cold Cracking of Interior
Cellular Polymeric Materials - Determination of Vinyl Materials, General Motors Engineering
Stress-Strain Characteristics in Compression - Part 1: Standards, Materials and Processes - Procedures,”
Low-Density Materials,” 1986.) November 1988.
4. ISO 3386/1, “Flexible Cellular Polymeric Materials - 20. ASTM D3884 - 92, “Standard Test for Abrasion
Determination of Stress-Strain Characteristics in Resistance of Textile Fabrics (Rotary Platform,
Compression - Part 1: Low­Density Materials,” 1986. Double-Head Method),” August 1992.
5. DIN 53420, ‘Testing Cellular Materials; 21. GM 9150P, “Resistance to Marring or Scuffing,
Determination of Apparent Density,” 1978 General Motors Engineering Standards, Materials and
(superceded by ISO 845 Cellular Plastics and Rubbers Processes Procedures,” September 1988.
- Determination of Apparent (Bulk) Density, 1988. 22. SAE J855 JAN94, ‘Test Method of Stretch and Set of
6. ASTM D 3574 - 95, “Standard Test Methods for Textiles and Plastics,” January 1994.
Flexible Cellular Materials - Slab, Bonded, and 23. AATCC Test Method 8-1996, “Colorfastness to
molded Urethane Foams,” May 1995. Crocking: AATCC Crockmeter Method,” 1996.
7. Renault RP 1048-73, “Resistance to Tear, Flexible and 24. GMW 3402TM Draft #3, Procedure for Determining
Semi-Rigid Cellular Materials,” 1973. Soil and Cleaner Resistance of Automotive Materials,
8. FMVSS 302, “Laboratory Procedures for GM Worldwide Engineering Standards, February
Flammability Compliance Tests Federal Motor 1998.
Vehicle Safety Standard No. 302, Flammability of 25. Association of the Non Woven Fabrics Industry
Interior Materials, Passenger Cars, Multipurpose (INDA) Standard test: 1ST 70.2 (95)
Passenger Vehicles, Trucks and Buses,” U.S. 26. SAE J1885 MAR92 - Accelerated Exposure of
Department of Transportation, June 1973. Automotive Interior Trim Components Using a

140 / Kurtycz, Bladon, English, Hogg, Godoy, Larre

 Controlled Irradiance Water-Cooled Xenon-arc
Apparatus, March 1992. Andrew Hogg
27. GM 2617M Interior Trim Parts Performance, General
Motors Engineering Standards, Materials and Andrew Hogg is the Dow
Processes ­ Textiles, December 1998. Automotive Market Manager
28. GM CPC 1472, Seat Cushion Fatigue - Jounce and for Polyurethane Flexible
Squirm, 7/7/79. Foams. He joined Dow in 1984
29. GM CPC 1473, Sliding Entry Test, 7/7/79. and spent six years in
30. Vinyl & Other Rolled Goods Construction Polyurethane sales in the United
Specification, General Motors Corporation, North Kingdom focussing on Rigid
American Appearance Engineering, Construction and Flexible polyurethanes plus
Number WN1J, Expanded Vinyl Coated Knit, ACES and distribution. In 1995,
Application/Process: Seating, Sidewall Trim, 10/1/87. he moved to the Automotive
group in sales for the United Kingdom and he moved to
Schwalbach in 1997 to assume his current marketing role.
BIOGRAPHIES

Eric Kurtycz Steven English

Eric Kurtycz is currently the Steven English is currently the

Advanced Project Manager for Global Business Development
Seating Systems in Dow Director for the Flexibles
Automotive. He received his Polyurethane business of The
Bachelor of Science in Dow Chemical Company. He
Information Systems from has a Bachelor of Science in
Oakland University, his Masters Chemical Engineering from the
of Science in Manufacturing University of Newcastle upon
Engineering from Wayne State Tyne. Steven joined Dow in
University and an MBA in 1985 in the UK and has held
Finance from Oakland University. He has 14 years positions in sales in Chemicals and Epoxies, European
experience in Automotive Interior Engineering including Marketing Manager for Automotive and Global Marketing
seven years in Automotive Seating. Eric joined Dow Manager for Polyurethanes.
Automotive in 2000.

Joanne Bladon Jose Godoy

Joanne Bladon received her Jose Godoy is currently a Senior

Bachelor of Engineering Specialist in the Flexible
Science degree from the Molded Group in Meyrin,
University of Western Ontario Switzerland. He graduated from
in 1987. She is currently in the the University of Geneva
Polyurethanes Flexibles Core (Switzerland) and received his
Development group. She began Ph.D. in Organic Chemistry
her career with Dow Chemical from the same University in
Canada in the Plastics Process 1983. After a Post Doc. at
Research department. She Imperial College of Science
then transferred to the Plastics Technical Service and and Technology in London on the synthesis of natural
Development and has been a Senior Specialist in the products, he returned to Geneva to work on a World Health
Polyurethane Technical Service and Development group in Organizationprogram on synthetic molecules against
Sarnia, Canada for eight years. Malaria. From 1987 to 1989, he worked in polyurethane
R&D for BP Chemicals then for Dow Europe SA in
Meyrin (Geneva) since 1989.

Kurtycz, Bladon, English, Hogg, Godoy, Larre / 141

Alfredo Larre

Alfredo Larre received his B.S.

in Chemical Engineering from
the National Technological
University in Buenos Aires,
Argentina. Alfredo Larre is
Project Director in the
Polyurethanes Research &
Development Department of
The Dow Chemical Company.
After working 15 years in the
PU foam industry in Argentina, Alfredo relocated to USA
to join The Upjohn Company, Polymer Chemical Division
TS&D group in La Porte, Texas. After Dow's acquisition
of the Upjohn PCD, Alfredo held several positions in the
PU Technical Service & Development organization
including PU Technical Manager for Latin America, North
American TS&D Leader for Flexible Foams and Global
Development Leader for Molded Foams.

142 / Kurtycz, Bladon, English, Hogg, Godoy, Larre

Automotive, Moulded Visco-Elastic Foams

PAUL FARKAS, ROMEO STANCIU and LIBERATO MENDOZA

Woodbridge Foam Corporation
Chemical Research and Development
8214 Kipling Ave.
Woodbridge, Ontario L4L 2A4
Canada

ABSTRACT As indicated by their name, these foams have a low

resiliency and good energy dissipating properties, and are
An improved type of moulded visco-elastic foam, also known for their sensitivity to changes in temperature
Intellisense™, extends the effective operating range of the and humidity [2].
material from elevated temperatures to below 0 deg. C., They are generally used in comfort cushioning,
while maintaining the energy dissipation values at packaging, bumper pads, sport and medical equipment,
designed levels. mattresses, helmet liners, pilot seats, earplugs, and for
Until now, visco-elastic (VE) foams have not found a various noise and vibration dampening applications [2].
significant success in automotive interior/exterior Flexible PU foams are amorphous materials. Below the
applications due to an evident loss of some essential Glass Transition Temperature (Tg), they behave as glassy
properties with changing conditions (e.g. diminished solids. In the proximity of the Tg, due to an increased
softness/ ,,friendliness,, in contact with human body when mobility o f the chain segments, the polymer starts to
used at low temperatures or reduced VE properties at high soften. This is the region with potential for maximum
temperatures). energy dissipation [3]— when the ratio of Storage Modulus
A visco-elastic behavior means a time delayed and rate E” to Elastic Modulus E’, the trigonometric tangent of
dependent response to applied stress [1] and makes it delta (Tan 8), expressed as a function of temperature,
desirable in parts designed for specific energy management peaks. Above Tg, the PU foam becomes “rubber like”—
(EM). However, to be acceptable for an application, the Tan 8 drops and its energy dissipation is reduced. With
effectiveness of the foam response has to be in the further heating, and reaching the Rubbery Flow
expected usage temperature range of the final part. Temperature (TRF), the Elastic Modulus (E’) drops sharply
Systems formulated with a balanced mixture of polyols, and the foam suffers a significant loss in mechanical
some special additives, and modified isocyanates generated properties [Fig. 1].
VE foams with the glass transition temperature (Tg) below The stress-strain relation for VE polyurethane foams and
- 20 deg. C. The observed loss of visco-elasticity at similar materials is a time dependent function. Under
elevated temperatures, common with prior-art foam of this
type, is significantly reduced as well.
This novel approach to visco-elastic foams will benefit
energy management automotive applications like head
restraint, seat occupant safety, sound or vibration
attenuation, etc. It is expected that this finding may extend
the area of applicability beyond the car industry.

BACKGROUND

In the context of this paper the term Visco-Elastic foam

refers to low resiliency flexible polyurethane (PU) foams
also known as memory, shape-memory, low recovery, slow
recovery, pressure relief or even zero resiliency foams [2]. Figure 1. Modulus vs. Temperature (typical curve)

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 143

dynamic load, a phase difference (time delay) appears INTRODUCTION
between the applied stress and the generated strain Ml.
The response to an applied stress will be rate dependent, The time delayed response of strain recovery in VE PU
increasing with the rate of application (e.g.: an increased foams of a given crosslinking density, is generally
stiffness with increased rate). In the case of the prior-art attributed, to an increase in hydrogen bonds in the polymer
VE PU foams for any given rate of stress, this response backbone. Factors affecting this relatively weak molecular
may be also temperature dependent [5], bonding may cause a change in a specific property (like an
When comparing the Tg values for High Resiliency (HR) increase in temperature or humidity that may weaken the
and prior-art VE polyurethane foams, it is evident that the Visco-elasticity) [2; 8]. The MW of the polyol is the
first has it well below 0 deg. C. (~ -40 deg. C) while the determining factor of the Tg of a PU polymer, while
second has a Tg of - +20 deg. C. [Fig. 2]. At the same time, additives, crosslinking density, and the nature of the
this foam is temperature “sensitive” [6. 14, 11, 5]. isocyanate used in the process are all influencing factors
As a consequence, while the HR foams remain “friendly [2; 9; 10]. Manufacturing of prior-art VE Foam is usually
to touch” at low temperature, the prior-art VE foams may done with systems that contain:
become too hard and “unfriendly”. When subjected to Polyols with an average MW -700-1000, mostly tipped
elevated temperatures prior-art VE foam shows an increase with secondary OH, relatively high levels of gel catalysts,
in elastic response and the energy dissipation characteristic cell openers and other specific additives for improved foam
is reduced. HR foam softens at elevated temperatures, processing stability and good final foam characteristics.
gradually losing its mechanical strength. Consequently, Isocyanates like polymeric and modified MDIs or,
both types have a relatively limited usability window. depending on application and resin blend, TDI.
For energy management applications, the challenge is to Commercial processing of these systems is described in
extend the temperature interval in which the material has the literature as being difficult and challenging. This is
predictable and good energy dissipation, from a narrow attributed mostly to the preponderance in the resin of the
range around Tg, to one covering the entire operating secondary OH that makes the reaction rates too slow.
interval of the final part. As an interior car part, the foam Addition of more catalyst to overcome process problems
has to keep its soft response to light touch in a wide results in products with closed cells, while supplementation
temperature range. Such polyurethane foam of an of cell openers reduced the foaming systems stability. As a
appropriate behavior is claimed in a recent patent [7]. result of these additions, the percentage of volatile
While recognized for their energy dissipation efficiency, components may be too high for the product. Flame
the prior-art VE foams have not found significant success retardant additives make some of them prone to early
in automotive interior applications due to their temperature scorching.
sensitivity. “For automotive applications... the same
characteristics o ffoam should be kept from below freezing FORMULATING INTELLISENSE™VE FOAM
point to far above room temperature to provide the same
dampening or comfort over a broad temperature range” It is an objective of this development work to achieve
[2 ]. Visco-Elastic properties in PU foams formulated with
From a design perspective, to be acceptable for an conventional, high molecular weight molding polyols,
application, the material response to stress has to be both implicitly conferring to the polymer low Tg values. These
effective and predictable over the entire operating polyols have a high percentage of reactive primary OH and
temperature range of the final part. consequential high reactivity within an acceptable
processing window. For foams, water generated C 02
should be the preferred blowing agent.
Processing should be done in a conventional way and
using the existing molding line equipment.
Another objective of the research work was to design a
new type of visco-elastic polyurethane foam with a wider
temperature range of optimized energy dissipation and a
lesser degree of essential property dependence on
temperature, resulting in the foams remaining essentially
soft at below 0 deg. C. temperatures. These foams have a
large automotive application potential.
The main polyols have a MW of -4700-6000, with a
significant amount of primary ­ O H functional groups.
Crosslinkers are used to achieve the desired foam
properties. Usage of large amounts of potentially volatile
catalysts is avoided. Polymer segment mobility is enhanced
with special additives, while aiming to reduce the
Figure 2. Tan S - H R vs. Prior-art VE Foam (typical curves) temperature sensitivity of the final foam.

144 / Farkas, Stanciu, Mendoza

 Cell openers complement the effect of designer Table 2. Typical Properties - IntelliSense™ VE Foam
surfactant blends in providing both product and process Typical Properties A B
performance. IntelliSense™ Foam
Commercially available modified MDIs (E.W. iso ~ 147­ Density [kg/nY3] 6 5 -7 0 6 7 -7 3
Tg [deg.C] -2 5 -4 3
149) are used.
Ball Rebound [%] 23 21
Table 1 shows two typical formulations.Patent
ILD (50% ) [N] 190-300 3 70-540
application procedures are currently in progress [15].
Tensile Strength [kPa] 160 155
Foam properties were measured, using known ASTM Elongation [%] 150 128
procedures. Table 2 shows a selected range of typical T e a r Strength [N/m] 260 29 0
measured properties. Compression Set (5 0 % ) [%] 2 5 -3 0 88
As mentioned, the objective of the development work is Fogging 100 100
to achieve an enhanced product that is able to overlap the Flammability [mm/min] 43 42
benefits of energy dissipation via cellularity and visco ­ (pass) (pass)
elastic based mechanisms, with the maintenance of the
“friendly touch” characteristic of the foam over the
expected usage temperature range. Impact, dynamic testing (FMVSS 201/GM9685P) is
The degree of Visco-elasticity and the implicit amount of used to confirm the energy dissipating properties of a
the expected energy dissipation are estimated in percent finished part and its impact rate dependency.
values of the Ball Rebound (BR) test (ASTM D 3574-91). Interpretation of the results follows the section describing
The lower the BR value, the higher the degree of visco­ the way in which samples were produced.
elasticity presence in the foam.
The low temperature behavior, the guarantor of the EXPERIMENTAL
“friendly to body part touch” and implicitly the design
predictability o f the VE properties is measured as Tg and Samples were prepared using one or both of the
tan 8 and estimated by Dynamic Mechanical Thermal following procedures.
Analysis (DMTA - ASTM D 4065). The lower the Tg, the
softer the foam should be at low temperatures; if tan L is an Procedure for Hand Poured Foam
- constant function of temperature, the foam will have a
predictable response to an imputed mechanical energy. A The resin blend was prepared by adding the right amount
higher absolute value of tan L means a higher loss modulus of each component in the order shown in Table 1. The
and related stored energy; the foam is more visco-elastic. blend was mixed using a high torque stirrer for 30 minutes,
Foam softness and the “friendly touch” is complementary at 1750 rpm and 23°C. The required amount of resin blend,
measured by another factor: measured in a 1.5 paper (Dixie) cup, was pre-mixed at
-The Ratio o f 25% IFD (GM 6293M), recorded at two 1750 rpm for 30 seconds using a Delta 0 2” lab mixer. The
different temperatures (+22 and -20 deg. C.). Any right amount of isocyanate was added under continuous
deviation from one (1) may indicate a change in the foam mixing and the timer was started. For “free rise”
hardness. estimation, the resin + isocyanate blend was mixed for 10
Elastic properties were measured as tensile, elongation, sec. and then poured into a 2 gallon bucket. Cream, string-
tear strength and compression set (CS) (ASTM D3574). gel and rise time were recorded. Moulded foam samples
Other relevant automotive foam properties like 50% IFD, were produced using the same mixing procedure and then
fogging (GM-9503P), and flammability (FMVSS-302) poured into a 10” x 10” x 4” aluminum test mould, and
were added for full foam characterization and a better heated at 65°C. The foam was demoulded after 6 min., left
understanding. to cool and kept for seven days at ~23°C and - 50%
relative humidity before testing for properties.

Table 1. Typical Formulations - IntelliSense VE™Foams Procedure for Equipment Poured Foam
A B
Typical Formulation Ingredients [%] [%] The resin blend was prepared in an open 50 gal. drum by
Polyol (E W 1500 - 2 000) 4 0 -6 0 4 0 -6 0 adding the right amount of each component, respecting the
Copolymer Polyol (E W 280 0 ) 1 0 -2 0 1 0 -2 0 same order of addition as described above. The blend was
Surfactants 0 .7 -1 .2 0 .7 -1 .5
mixed for 60 min. using a Lightnin NAG 100 air driven
Crosslinkers 2 5 -4 0 2 5 -4 0
0 .5 -0 .8
mixer at 300 - 500 rpm and ~23 °C.
Catalysts 0 .5 -0 .8
Cell Opener 0 -1 -
The resin and the isocyanate were loaded into a high-
Additive Type I 0 -1 0 - pressure foam dispenser (model WFC) equipped with a
Additive Type II - 0 -1 0 WFC modified EMB HP 22 mm straight mixing head
W ater 1 -3 1-3 servicing a carousel with a 4 - 8 mould capacity.
Iso Index 7 0 -1 0 0 7 0 -1 0 0 The foam was poured using a 2000 - 3000 psi injection
pressure at a throughput of 150 - 300 g/sec. The mould

Farkas, Stanciu, Mendoza / 145

used was 10” x 10 “x 4” aluminum test blocks or an actual
production headrest mould, heated at 65 °C.
The parts were demoulded after 6 min., left to cool and
kept at -23 °C and - 50% relative humidity for seven days
before testing. Foam properties were measured using the
known ASTM procedures.

UNDERSTANDING THE TEST RESULTS

In the following paragraph, test result details are

explained and evaluated in their interdependence. This
interpretation carefully indicates the pass chosen in
obtaining a complete set of useful characteristics in our
search for an optimized interior car part with enhanced
energy dissipating properties.
Figure 4. lntelliSenseT" Foam - Type B
The degree of achieved visco-elasticity was measured
using a Time Tech, Ball Rebound Tester Model TT-502. DMTA Analysis
The chosen BR level of -21 ­ 23% is considered useful for
automotive moulded foam padding for head restraint
applications. Lower BR numbers, as sometimes reported in The prior-art VE foams are sensitive to temperature
literature for prior-art VE foams, are considered changes. For temperatures lower than their Tg, they are
unnecessary for the part performance. stiff and “unfriendly”, while at elevated temperatures the
The curves indicated in Figures 3, 4, 5, 6 are the results elasticity is the predominant characteristic— the foam
of DMTA tests performed in compression mode, using becomes too soft to maintain its usefulness in energy
Rheometric Scientific Mark IV equipment. management applications. Foam softness changes were
The glass transition temperature was recorded at the measured as a relative increase in Indentation Force
maximum of the Storage Modulus curve (E”), the direct Deflection (IFD) performance.
measurement o f the most relevant property for the The IFD measurements were made with a Zwick model
dissipated energy [16]. 050 universal testing machine, using a 2.5 KN load cell,
The Tg values of -25°C (foam A) and -43°C (foam B) and a standard 323 cm2 flat indentor foot.
are substantially lower than the prior art VE foams average The ratio values for 25% IFD recorded at two different
of ~+20 °C, and comparable with the, as mentioned above, temperatures of +22 and -20 °C are 0.74 (foam A) and 0.69
HR foam Tg of —40°C. The range of temperature (foam B). This is a load variation of only - 26% for foam A
“insensitivity” is considerably larger than that of the and - 31% for foam B, respectively. These changes are
conventional HR foam or prior-art VE foam. Foam B much lower than in some prior-art VE foams, which are as
shows, in particular, a remarkable constancy of its relative high as -60 - 70% under the same condition changes [7].
energy dissipation ability from —30 to —­ ( 105 °C, All IntelliSense™ foams show a good corroboration of
compared with - +7 to +60 °C for the HR foams of a their thermal behavior with the impact results under
similar Tg value [Fig. 6]. dynamic testing done with temperature-conditioned

Figure 5. FIR Foam

Figure 3. IntelliSense " Foam - Type A
DMTA Analysis
DMTA Analysis

1 4 6 / Farkas, Stanciu, Mendoza

 Figure 6. IntelliSense™ F oam vs. H R F oam
Tan L T em peratu re D e p e n d e n c y

samples [Fig. 7]. The impact test was performed using a Figure 7. In te lliS e n se ™- T ype B vs. H R F oam
WFC Pendulum Impact Tester with a flat 7” diameter
impactor.
Markedly, there is a substantial softening by HR foams at
the maximum test temperature of +80 °C (no curve shown
AUTOMOTIVE APPLICATIONS
as the sample was bottoming-out), while the IntelliSense™
foam has little change from its behavior at ambient
Due to their complex energy dissipation mechanism,
temperature.
extended over a larger temperature range, with a
Impact simulation, with a Hybrid III dummy placed 2
simultaneous conservation o f the much appreciated
inches from the head restraint and an impact velocity of 15
Flexible Polyurethane foam “touch” the usage of the
mph, results shown in Table 3, indicate a substantially
IntelliSense ™ VE Foams becomes an attractive challenge
improved energy dissipation with the new VE foam
for automotive parts designers. The implicitly built-in
concept, over a HR PU or prior-art VE foam padded
impact rate sensitivity, combined with a known
headrests. The lower Flexion Bending Momentum value
may provide the improvement needed for whiplash effect
avoidance.
The expected visco-elastic behavior of IntelliSense™
foam is evident in its impact rate sensitivity. Comparing
the impact curves at different velocities, for HR [Fig. 8a]
and IntelliSense™ [Fig. 8b] foams (recorded under similar
experimental conditions), the latter’s increased hardness
with increased impactor velocity is demonstrated.
IntelliSense™ may provide with the comfortable softness
of a light, friendly touch and when needed the stiffness
required against a bottoming-out effect as in the case of a
high velocity impact.

Table 3. Im pact Sim ulation

PU Foam Type Flexion Bending

Momentum [Nm]
HR (Control) 291
IntelliSense™ 87
Prior-art VE Foam 1* 108
Prior-art VE Foam 2** 192
♦A Temperature sensitive product Figure 8 . . E ffect o f Strain R a te on H R H e a d R estrain t Foam
* * Commercial automotive product a n d In telliS en se ™ VE F oam

Farkas, Stanciu, Mendoza / 147

 REFERENCES

1. Birley, A. W. at al. 1991. Physics o f Plastics. New

York, NY: Oxford University Press, pp. 260-261.
2. Herrington, R. July 20, 2000. “Low Resiliency
Flexible Polyurethane Foams”, Tech Note /Report to
WFC, Dow Plastics/Vortex.
3. Saha, P. 1992. “Materials for Interior Acoustics in an
Automobile”, Body Engineering Journal: 44-47
4. 4.Riande, E. at al. 2000. Polymer Viscoelasticity. New
York, Basel: Marcel Decker, Inc. pp. 238-240
5. Davies, O. at al. 1999. “The rate dependence of
Confer Polyurethane Foams”, Cell. Pol. 18(2): 117-
136
6. Davies, O. at al. 2000, “Seating Pressure Distribution
Using Slow Recovery Polyurethane Foams”, Cellular
Polymers 19(1): 4.
Figure 9. S a feG u a rd ™ H e a d R estra in t - In telliS en se ™ VE 7. Nishida, K. at al. 1999. EP.0913414 Al
Foam ; Im pact T e st 8. Saunders, J.H. at al. 1978. “Polyurethanes: Chemistry
and Technology Part I Chemistry. Krieger, New York,
p.331
manufacturing process may extend the potential of 9. Smith, T.L.at al. 1960, “Diisocyanate-Linked
IntelliSense™ Foams far beyond the currently envisaged Polymers. Mechanical and swelling properties of some
headrest application [Fig. 9] to other interior, exterior car polyurethane elastomers”, J. o fAppl. Sc., 62,: 391
10. Whitman, R.D. at al. 1963. “Urethane Foam Properties
parts. IntelliSense™ VE Foams may benefit energy
related to The Fundamental Polymer Molecular
management applications for seat occupant “comfort of
Structure”, Proceedings o f the 1963 SPI Polyurethane
safety”, be it in head restraints, arm rests, seat backs,
cushions, sound and vibration attenuation programs or
Conference, Technomic, Lancaster, PA. 1963, p. 1
11. Raucholtz, D. “Using Specialty Engineered Foams in
other energy management parts.
Seating Design”, E.A.R. Specialty Composites,
In particular in head restraint application the
Indianapolis, IN.
IntelliSense™ VE foams helps to meet the new energy
12. Stoffer, H., Jan.8, 2001, “NHTSA seeks tougher head
absorption requirement [12; 13].
restraint rules”, Automotive News,
13. NHTSA, Jan. 2001. 49 CFR Part 571 [Docket No.
CONCLUSION NHTSA-2000-8570] “Notice of Proposed
Rulemaking”; MVSS; Head Restraint.
New VE Polyurethane Foam, IntelliSense™, was 14. Temper Foam, 1982. “A Unique Polyurethane Foam
designed and successfully tested for potential automotive Material... ”, Edmont-Wilson, Coshocton, OH.
usage. The new foam provides: 15. Farkas, P. at al. 2000. US Pat. Appl.
• Comfort and Feel of a conventional foam; 16. Rieger, J., 2001. ’’The Glass Transition temperature Tg
• Energy-Dissipation/Absorption over an extended ­ of polymers-Comparison of the values from different
temperature range in a recoverable foam form; thermal analysis (DTA, DSC) and dynamic
• Temperature Stability and general predictable mechanical measurements (torsion pendulum).,
behavior within a wide temperature range; ELSEVIER, Polymer Testing, 20:199-204
• Strain-Rate Synergy - Increased energy­absorption
with increase impact force;
• Product Performance ­ Reducing rebounding loads IntelliSense™ & Safeguard ™ are trademarks of
to help meet new Regulatory and Due-Care Woodbridge Foam Corporation
requirements..

ACKNOWLEDGEMENTS

The authors would like to thank Woodbridge Foam

Corporation for the funding of this research; the Chemical
Research and Development Department and the Troy
Product Development Group for their continuous
encouragement and the help with sample making and
testing.

1 4 8 / Farkas, Stanciu, Mendoza

BIOGRAPHIES Liberato Mendoza

Liberato V. Mendoza has

Paul Farkas been in the polyurethane
industry for 23 years. He is
Paul Farkas is a Corporate Senior currently a Senior Chemical
Scientist in charge of the Technologist in the Chemical
Emerging Technologies Group Research and Development
within the Chemical Research Department with Woodbridge
and Development Department Foam Corporation,. Liberato’s
Woodbridge Foam Corporation. expertise is in the
Paul’s education is in Chemical development of urethane
Engineering, Chemistry and systems for automotive and
Technology of Macromolecular non-automotive applications.
Compounds, and Organic
Chemistry. Paul is active in the
urethane industry for over 30 years and is the author/co-
author of technical presentations, papers, a book chapter
and numerous U.S. and foreign patents in the field of
isocyanate chemistry based products.

Romeo Stanciu

Romeo Stanciu is a Corporate

Senior Chemist in the
Emerging T echnologies
Group of the Chemical
Research and Development
Department with Woodbridge
Foam Corporation. Romeo’s
education is in Chemical
Engineering, Chemistry and
Technology of Macromolecular
Compounds. He joined
Woodbridge Foam
Corporation in 1997 as R&D chemist working in the area
of flexible PU foams and is currently involved in the
development o f new energy absorbing flexible PU foams.

Farkas, Stanciu, Mendoza / 149

HFC-245fa: An Ideal Blowing Agent for Integral Skin Foam

BIN ZHONG, JIM M. BOWMAN and DAVID WILLIAMS

Honeywell
20 Peabody Street
Buffalo, New York 14210

DISCLAIMER foams prepared using CFC-11 or HFC-245fa, especially

skin thickness. A blowing efficiency study was conducted.
Nothing herein should be construed as an inducement to For each blowing agent, two 16pcf and 32pcf foams were
infringe any patent including, but not limited to, U.S. molded with free rise densities of 8pcf and 16pcf and pack
Patent No. 5,906,999, U.S. Patent No. 6,218,443, and U.S. factor of 2. The relationship of integral skin foam
Patent No. 6,010,649. properties with regard to variables such as mold
temperature and packing level was also explored. The
ABSTRACT results showed that at the same weight percent loading of
blowing agent in the formulation, the free rise density of
Hydrochlorofluorocarbons have been banned in the US HFC-245fa blown foams was approximately one-half that
from use in integral skin foam applications for over five of the CFC-11 blown foams. This indicates that HFC-
years. During this period, C 0 2(water) and HFC- 134a have 245fa has much higher blowing efficiency than CFC-11
been used with varying degrees of success in these even though the two have very similar molecular weight.
formulations. However, compared to CFC-11, foams This result is consistent with our previous work. The
blown with water and HFC-134a are both characterized by results also showed that HFC-245fa gave similar skin
poor skin formation. thickness compared to CFC-11 and showed very similar
Since the phase-out of HCFCs in integral skin foam in change in skin thickness as mold temperatures or packing
January 1, 1994, Honeywell has done considerable work to levels were varied.
develop a next generation blowing agent and is now As a more cost-effective albeit lower performance
commercializing HFC-245fa (1,1,1,3,3- option, HFC -245fa/C02(water) blends were also studied in
pentafluoropropane) to replace HCFCs. A world-scale this work. Two water levels, 10mol% and 30mol%, were
HFC -245fa plant is now being built by Honeywell in formulated to make foam with same free rise density as
Geismar, Louisiana. The plant is on schedule to be neat HFC-245fa. The results showed that addition of water
commercially operational by mid-2002. HFC -245fa is well lowered the vapor pressure of the polyol premix. The
known as a third generation blowing agent with zero ozone results also indicated that water caused higher in-mold
depleting potential (ODP), non-flammability, and pressure. Even at 30mol% water level, a good skin quality
acceptable global warming potential (GWP). With the was obtained.
commercialization of HFC-245fa, applications like integral
skin foam that need a better performing blowing agent may INTRODUCTION
find a solution to problem areas that have not been met by
water and HFC-134a. Integral skin foam has a combined structure of a cellular
In this paper, a CFC-11 based integral skin foam core and an elastomeric skin. It is widely used in a variety
formulation was used to compare the physical properties of of applications, such as automotive interior trim, shoe

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 151

 T able 1. Blowing a g e n t p h y sic a l a n d en viron m en tal p ro p e rtie s
Carbon Cyclo­
CFC-11 HFC-134a HFC-245fa n-Pentane
Dioxide pentane
Formula CCI3 F CH2 FCF3 C 02 CF3 CH2 CF2 H (CH2) 5 CH3 (CH2 )3 CH3
Molecular Weight 137.4 1 0 2 .0 44 134.0 70.0 72
Boiling Point (°F) 74.8 -15.1 -218.2 59.5 120.7 161.6
Pvao (PSIA @ 68°F) 1 2 .8 82.9 17.8 4.9
Vapor Thermal Conductivity
0.051 0.086 0 .1 0 1 0.0932 0.076 0.097
(BTU in/ft2 hr °F) @50°F1
Flash Point (°F) None None None None -35 -56.2
Vapor Flame Limits (Vol. %) None None None None 1.4-9.4 1 .3-8.0
VOC Status (US) Exempt Exempt Exempt Exempt Yes Yes
Ozone Depletion Potential 1 .0 0 0 0 0 0
Global Warming Potentialb
4600 1600 1 9906 117 117
(C0 2 = 1 , 1 0 0 yr)
1998 UNEP Rigid and Flexible Foams Technical Options report, Table 1, unless otherwise noted
2Geller, Bivens, Yokozeki, “Transport Properties and Surface Tension of Hydrofluorocarbons HFC 236fa and HFC-245fa”20th
International Congress of Refrigeration, IIR/IIF, Sydney 1999
3 Zipfel et al, “HFC-365: A Versatile Blowing Agent for Rigid Polyurethane Foams”, Polyurethanes Expo ‘98,1998. Note, data at 25°C
4Albouy et al, “A Status Report on the Development of HFC Blowing Agents for Rigid Polyurethane Foams”, Polyurethanes World Congress
‘97, 1997
5 “Scientific Assessment of Ozone Depletion: 1998”WMO Global Ozone Research and Monitoring Project Report No. 44, unless otherwise
noted
6Magid et al, “Atmospheric Lifetime and Global Warming Potential of HFC-245fa”, J. Geophysical Research, Vol. 104, No. D7,
pp. 8173-8181, 1999
7From 1998 UNEP Rigid and Flexible Foams Technical Options report, Table 1

soles, and furniture. The mechanism of integral skin potential (ODP), non-flammability, and acceptable global
formation is quite complex, but it is widely accepted that warming potential (GWP). With the commercialization of
the skin is formed through the condensation of the blowing HFC -245fa, applications like integral skin foam that need a
agent. Two factors contribute to the condensation process. better performing blowing agent may find a solution to
One is the high temperature gradient across the reacting problem areas that have not been met by C 0 2 (water) and
materials adjacent to the relatively cool mold surface. The HFC-134a.
other is the pressure in the mold. From technical perspective, important properties of
CFC-11, due to its moderate boiling point and good blowing agents include boiling point, blowing efficiency,
solubility in polyol, gave excellent skin quality and was solubility in the polyol, vapor pressure, chemical and
very successful as a blowing agent in integral skin foam. physical stability in PU foams, environmental and
After the phase-out of CFCs, CFC-11 was replaced by industrial hygiene characteristics, as well as the
HCFCs, such as HCFC-141b. HCFCs have been banned in flammability and explosion hazards. Table 1 gives a
the US from use in integral skin foam since January 1, summary of the physical and environmental properties of
1994, with the exception of automotive safety application, blowing agents commonly used in integral skin foam.
which was granted an extension until January 1, 1996. In Another replacement option currently used in Europe is
Europe, HCFC-141b has been banned from use in integral pentane, which gives better skin than C 0 2 (water) and
skin foam since January 1, 2000, except for safety HFC-134a. But due to its flammability, pentane is not
applications. Since the phase-out of HCFCs, C 0 2 (water) considered a preferred alternative blowing agent in the US.
and HFC-134a have dominated integral skin market in the Therefore, this paper will focus on the comparison of HFC-
US. However, foams blown with C 0 2 (water) and HFC- 245fa with water, HFC-134a, and CFC-11. A discussion of
134a are both characterized by poor skin formation. the limitation of water and HFC -134a is given in this
Properties of these foams have generally not been as paper. A laboratory evaluation of HFC -245fa based on a
desirable as integral skin foam blown with CFC-11 or CFC-11 formulation is presented. In addition, some initial
HCFC-141b. Some incremental improvements have been results of investigation of HFC-245fa/C02(water) blend is
developed to address this issue in the past several years, also reported.
but essentially the problem remains.
With the phase-out date of HCFC-141b in all PU foam C 0 2(WATER)
applications on January 1, 2003, approaching, Honeywell
is now building a world-scale HFC-245fa plant in Geismar, C 0 2 is typically generated by the reaction of water with
Louisiana. This plant is on schedule to be commercially isocyanate. Water blowing technology was first introduced
operational by mid-2002. HFC -245fa is well known as a in 1993 and has been mostly used in medium-high density
third generation blowing agent with zero ozone depleting integral skin foam applications^ 1] However, there are

1 5 2 / Zhong, Bowman„ Williams

several problems associated with the use of water as H FC -134a
follows: [1]
HFC -134a is a gaseous zero ODP physical blowing
■ Reduced skin formation, which results in poor agent with a boiling point at - 27°C (-15.1°F). It has been
abrasion performance and color dishomogeneity. used in integral skin foam since the phase out of HCFCs.
Therefore, coatings are strongly recommended if Compared to water blowing technology, it is known in
abrasion resistance is required by the application. industry that HFC -134a gives lower in-mold pressure and
■ High in-mold pressure, which leads to post-blowing as better skin densification but its cost is higher than
well as skin-peeling phenomena with the mold inserts. C 0 2(water). It should be noted that the skin with HFC-
■ Poor touch feeling, i.e., high Compression Load 134a is still much thinner than with CFC -11. The reason is
Deflection / Hardness ratio. probably its boiling point, higher than C 0 2 but much lower
than CFC -11. Similarly to water, the skin with HFC-134a
■ Poor foam quality at low densities, which is due to
is not from the condensation of HFC-134a but the
extra amount of isocyanate that is needed to react with
suppression of the blowing on the foam surface. From a
water and thus might make the foam too tough or rigid
technical perspective HFC -134a is typically characterized
due to the extra crosslinking. This is why water blown
by several problems, in particular by its problematic
foams must be run at greater than 15pcf with 20-30pcf
processing and low solubility in polyol. The problems
being more common. It should be noted that cost
efficiency of water is offset to some extent by the associated with HFC -134a are summarized as follows:[1,7]
extra isocyanate added.
■ Low boiling point and high vapor pressure, which
Many of these problems result from the fact that the C 0 2 make shipping, handling, as well as packaging
generated in the foaming process does not condense on the problematic. Usually pressurized and temprature-
mold surface. Since the introduction of water technology, a controlled tanks are needed for blending with polyol.
great amount of effort has been taken to address this issue. ■ Low solubility in polyol. Due to its higher blowing
Basically, there are two approaches that have been taken to efficiency than CFC -11, its resultant lower loading
address this issue. [2] may offset part of this problem. But this is still a big
One way is properly choosing raw materials. High concern with HFC-134a.
functional polyols or isocyanates may be used to adjust the ■ Better skin than water but much worse than CFC -11.
balance of gelling and blowing reaction. [3,4] MDI soft
segment prepolymer technology is a well-known example Skin quality is a critical parameter in integral skin foam.
under this approach, in which novel MDI is employed to The industry was developed based on the performance of
make water blown integral skin foam. CFC -11which is considered the best blowing agent in
The second way is selecting appropriate catalysts. It is terms of skin quality. Water and HFC -134a blowing
known that the catalyst used in a PU system varies in its technology have been used in industry for over five years.
selectivity to the water and polyol reactions. If catalysts But the deficient skin formation with these two
can selectively catalyze the water reaction in the foam core replacements has become a driving force for the industry to
while favoring the gelling reaction at the surface, the skin look for other alternatives. With the approaching phase out
formation on the surface might be improved. This requires of HCFCs, Honeywell is commercializing HFC-245fa as
a package whose selectivity varies with temperature in the next generation blowing agent. The evaluation results
such a way that the blowing reaction is suppressed at lower of using HFC -245fa in integral skin foam from both
temperatures.[l] For example, some heat activated blowing laboratory and industrial trials are very encouraging. It has
catalysts are intended for this purpose. [5] shown that HFC -245fa gave very similar skin quality as
These modifications do improve the foam properties in CFC-11 and has the potential to be an excellent blowing
some respect. However, those improvements are agent in integral skin foam.
incremental. Essentially the problem remains. This is
probably because it doesn’t change the fact that there is no H FC -245fa OVERVIEW
condensation of C 0 2 on the foam surface. The skin
obtained with water blown foam is virtually not a result of HFC -245fa is chemically identified as 1,1,1,3,3-
condensation of C 0 2, but a suppression of the blowing pentafluoropropane with a boiling point at 15.3°C
close to the foam surface by the low temperature on the (59.5°F). HFC -245fa is a non-ozone depleting,
surface and the selection of catalyst. nonflammable compound with an acceptable global
In addition to C 0 2 generated by water, gaseous and warming potential (GWP). It is not considered a volatile
liquid C 0 2 are also being looked into as blowing agent. organic compound (VOC) in the US. HFC-245fa is of low
However, with gaseous and liquid C 0 2 there is dosing toxicity with a AIHA Workplace Environmental Exposure
accuracy problems that lead to unacceptable density Level (WEEL) of 300 ppm.(TWA 8 hr.)
variations in production. [6] HFC -245fa is compatible with most of the metals,
plastics and elastomers currently used in foam industry. It
also shows a good solubility in polyols. The vapor pressure

, ,
Zhong Bowman Williams / 153
of the typical integral skin foam polyol premix with HFC- prepared by adjusting the loading of both CFC-11 and
245fa is very similar to the systems with CFC-11 at the HFC-245fa.
same free rise density level. The US EPA has listed HFC- Each of the polyol blends was added directly into a
245fa on the TSCA inventory. In addition, The US EPA plastic quart cup for foaming. Sufficient reaction mixture
has approved HFC-245fa under the Significant New was added to produce foam that rose above the lip of the
Alternatives Policy (SNAP) program for all foam blowing quart cup so that free rise densities could be measured. The
end uses including integral skin foam applications. Similar mixture was then stirred for 10 seconds with a 2" double
approvals have been or are in the process of being obtained blade at 3100 rpm. Cream, gel, rise and tack free times
globally for HFC-245fa. were recorded. After setting for 2 hours, the foam top was
Honeywell is now building a HFC -245fa plant in cut along the cup lip and the net weight of the foam in the
Geismar, LA. which is on schedule to be commercially cup was weighed. Foam free rise density was calculated by
operational in mid-2002. Between now and then, sufficient multiplying the foam weight and the cup volume. The
semi-commercial material is available to conduct the lab, resultant free rise densities and reactivity profiles are given
plant, and field trials needed to qualify HFC -245fa and to in Table 2. The blowing agent loading versus given free
develop the formulation and processing technology for rise density is shown in Figure 1.
launching a successful product containing HFC-245fa. Figure 1 shows that HFC -245fa has a much better
blowing efficiency than CFC-11. For a given density,
COM PARISON OF HFC-245fa and CFC-11 HFC-245fa loading is approximately half that of CFC-11.
This is consistent with the results we reported earlier on a
The purpose of this study was to compare the
performance of HFC-245fa to the historical industry
standard, CFC-11. HFC-245fa was substituted for CFC -11
in a commercial CFC-11 based integral skin foam
formulation. Foam performance was compared to the same
formulation blown with CFC-11.

Blowing Efficiency

CFC-11 and HFC-245fa foams were prepared by

handmix with polyol blend at 50°F and MDI at 70°F. A
master batch of polyol blend was prepared combining all
components except the blowing agent. Karl Fischer method
for water determination was performed and residual water
in the master batch was determined to be 923ppm. This
value was used to determine all polyol/MDI ratios. For Figure 1. Wt% (in polyol blend) blowing agent versus free
each foam, polyol blend was made by adding the liquid rise density for C F C -11 and HFC-245fa.
blowing agents (CFC-11 or HFC-245fa) to the polyol
system, and then mixing. A series of three foams, with
nominal free rise densities of 8, 12, and 16pcf, were

Table 2. Blowing efficiency of HFC ’245fa & C F C 11.

#1-1 #1-2 #1-3 #2-1 #2-2 #2-3
Master Batch (g) 111.7 168.7 234.6 111.7 170.9 234.6
CFC-11 (g) 28 33.1 37
HFC-245fa (g) 12.6 15.1 16.0
BA wt% in polyol blend (%) 20.0 16.4 13.6 10.1 8.1 6.4
Mondur MR (g) 32.7 49.4 68.7 32.7 50.0 68.7
lndex=100
Component Temp. (°F) 50/70 50/70 50/70 50/70 50/70 50/70
Cream Time (s) 20 19 17 22 20 19
Gel Time (s) 66 64 60 59 54 52
Rise Time (s) 92 90 86 84 80 76
Tack Free (s) 95 92 87 90 85 78
Overall density(pcf) 8.4 12.2 16.0 8.4 12.4 15.6

1 5 4 / Zhong, Bowman, Williams

different commercial formulation.[7] This phenomena is M olding Study
very interesting because both blowing agents have almost
the same molecular weight. There are two benefits coming A 6" x 6" x 3" aluminum mold and typical handmix
out of the resultant lower loading. The first is lower cost equipment were used to make molded foams. Both CFC-11
due to the lower blowing agent loadings. The second is a and HFC -245fa blown foams were molded with the same
reduction in polyol blend vapor pressure, minimizing free rise density, mold temperature, and pack factor. This is
shipping and handling issues. to ensure that the skin thickness is caused only by the
blowing agent condensing on the mold surface. Two free
V apor P ressure rise densities, 8pcf and 16pcf nominal, were employed, i.e.
systems #1-1 & #2-1, #1-3 & #2-3 shown in Table 2.
The vapor pressure of polyol blends of two systems, #1- Mold temperature was set at 40°C. Pack factor was held at
1 and #2-1 (see Table 2), which gave same density of 2. Demolding time was held constant at 5 min. Semi-
8.4pcf, were measured at 20, 30, 40, and 50°C permeable tape was applied to the four 1/4” dia. holes on
respectively. Static vapor pressure was taken without the mold lid before each molding to release air during
shaking the polyol premix prior to reading, while dynamic foaming while maintaining the inside pressure. Silicone
vapor pressure was taken after shaking by hand. The mold release was applied to the mold inside wall prior to
results are shown in Figure 2 & 3. each molding. Two hours after demolding and trimming,
Figure 2 & 3 clearly show that even though HFC -245fa the net weight of the molded foam was taken and overall
has a lower boiling point and higher vapor pressure than density was calculated. The foam skins obtained are shown
CFC-11, because of its lower loading in the formulation, in Figure 4.
the vapor pressure of its polyol blend is equal to or slightly The skin thickness was measured with a ruler capable of
lower than that of the CFC-11 containing polyol. measuring 1 mm. Since there is no clear border between
Particularly at 30°C in Figure 3, CFC-11 system shows a the skin and core foam, the error could be ±0.5mm. The
dynamic vapor pressure of 3psi while the reading on HFC- results were shown in Table 3.
245fa system is zero. This may be due to better solubility The results demonstrated that the skin thickness with
of HFC-245fa. The results also show that vapor pressure of HFC -245fa was very close to that of CFC-11, given the
both systems was far below 22psi, the maximum allowable error of ±0.5mm. In addition, HFC -245fa showed very
pressure of standard 16 gauge drums used in industry. This similar in-mold pressure to CFC-11.
will allow for shipping of integral skin polyol premix in
standard drums.

Figure 2. Static vapor pressure of polyol blends with CFC -

11 and HFC - 245fa.

Figure 4. Molded foam skin with CFC-11 & HFC-245fa.

Figure 3. Dynamic vapor pressure of polyol blends with

CFC-11 and HFC-245.

Zhong, Bowman, Williams / 155

 Table 3. Molded foam with CFC-11 & HFC-245fa
Low Density High Density
HFC- HFC-
CFC-11 CFC -11
245fa 245fa
Mold temp.(°C) 40 40 40 40
Free rise density 8.41 8.44 16.01 15.64
Pack factor 2 2 2 2
Molded foam
16.8 16.95 31.86 31.09
density (pcf)
Skin thickness
2 1.5 3.5 2.8
(mm) Figure 6. The impact of mold temperature on skin
thickness.
M old T em perature Study

Mold temperature is a very important factor affecting the The results showed that mold temperature had a very
skin formation. The impact of mold temperature on the similar impact on the skin formation of HFC-245fa and
skin thickness was investigated using both CFC-11 and CFC-11 blown foams. The lower the mold temperature is,
HFC-245fa. the thicker the skin. However, it should be noted that CFC-
Low density systems #1-1 and #2-1 were employed in 11 blown foam showed more shrinkage than HFC -245fa at
this study. Pack factor was kept at 2. Four mold 20°C even though they both recovered after being set
temperatures, 20, 30, 40, 50°C, were used in the molding. overnight. At 50°C, both foams showed higher in-mold
Figure 5 shows the skin formation. The skin thickness was pressure and “honey comb” was observed on both skins.
m easure and shown in Figure 6.
Packing Study

Packing level is another important factor affecting the

skin formation during molding. The impact of pack factor
on the skin thickness was investigated using both CFC-11
and HFC-245fa.
Low density systems #1-1 and #2-1 were employed in
this study. Mold temperature was kept at 20°C. Four pack
factors, 1.5, 2, 3, and 4, were used in the molding. Figure 7
shows the skin formation. The skin thickness was
measured and shown in Figure 8.
The results show that pack factor has a very similar
impact on both HFC -245fa and CFC-11 blown foams. At
each packing level, HFC -245fa gave a skin thickness very
similar to CFC-11. It demonstrated that with HFC-245fa, a
4-5mm thick skin could be achieved by packing a system
with free rise density at 8pcf and pack factor 4, which
gives a molded foam with a overall density of 32pcf.

Top layer- CFC-11, Bottom layer-HFC-245fa Figure 7. Impact of pack factor on skin formation.
Figure 5. Skin formation of CFC-11 and HFC-245fa blown
foams at different mold temperatures.

1 5 6 / Zhong, Bowman, Williams

 M oisture C ontent

In order to control the water content in the system, the

moisture content of all the components and the polyol
blend were measured by Karl Fischer moisture tester. The
results displayed that the total moisture content of all the
components in polyol blend before blending is 267ppm,
while the moisture content of the polyol blend is 922. This
means that the system is hydroscopic and most of the
moisture was absorbed into the system during blending and
processing. In the applications where high quality skin and
very low in-mold pressure are needed, exclusion of water
during the blending process or the use of moisture
scavengers, such as zeolite, might be needed to prevent this
moisture absorption.

V apor P ressure and Skin Form ation

Two water levels, 10% and 30% by molar ratio, were

chosen to substitute the HFC -245fa in both system #2-1
and #2 -2 shown in Table 2. The vapor pressure of the
resultant polyol blends has shown that HFC-
245fa/C02(water) gives even lower reading compared to
neat HFC-245fa. At 50°C, 10mol% water level for both
low and high density formulations gave readings lower
than 3.5psi, while 30mol% water level gave readings lower
than 3psi.
The HFC -245fa/C02(water) blown foams were molded
using mold temperature 40°C and pack factor 2. The skin
formation was shown in Figure 9.
Figure 8. Skin formation of CFC-11 and HFC-245fa blown
foams at different pack factors.

H FC -245fa/C 02(w ater) BLENDS

HFC -245fa has shown that it can give very good skin
formation, close to CFC-11 but much better than HFC-
134a and water systems. Because of this large performance
gap between HFC-245fa and HFC-134a or water, for those
applications that do not require the skin formation to be as
good as HFC -245fa but do need much better skin than that
with current blowing agent, a cost efficient option might be
a HFC -245fa/C02(water) blend.
The introduction of water into the blend gives several
benefits. First it reduce the blowing agent cost
significantly. 30mol% of water in the blend means 30%
reduction in blowing agent cost. Secondly it helps further
reduce the vapor pressure of polyol blend which is
favorable in production and handling.
The questions with HFC -245fa/C02(water) blend are
how water will affect the skin formation and how much
water can be added into the formulation and maintain
acceptable skin quality. An initial study had been
conducted for this purpose.
Figure 9. Skin formation of H FC - 245fa/C 02(water) blown
foams.

Zhong, Bowman, Williams / 157

 The results show that the skin formation with HFC- 2. Edmund J. Madaj. 1996. “New Developments In
245fa/C02(water) is less than neat HFC-245fa. But even at Water-Blown Integral Skin Polyurethanes,” presented
30% water level there is still an appreciable amount of skin at Polyurethanes Expo’96, 1996.
formation. A potential problem with HFC- 3. H. Ueda, M. Sakai, H. Inoue, etc. 1994. “All-Water
245fa/C02(water) blend is in-mold pressure. Addition of Blown Integral Skin Foam for Automotive Steering
water to the system increased the in-mold pressure that led Wheels,” presented at the 35th Annual Polyurethane
to post-blowing in some cases at the 30% water level. This Technical/Marketing Conference, Oct. 9-12, 1994.
in-mold pressure is still lower, however, than 100% water 4. A. Frangareggi, G. Guidetti, and L. Pedroni. 1991.
blown foams. It is believed that this can be solved by “New Approach To CFC Free Steering Wheels,”
further system optimization. presented at SPI Conference’91, 1991.
5. Michael A. Dimitroff. 1999. “Catalysts And
CONCLUSION Surfactants Developed For Water-Blown Polyurethane
Integral-Skin Foams Used In Footwear Applications,”
HFC -134a and water are the blowing agents currently presented at Polyurethanes Expo’99, 1999.
used in integral skin foam applications. But compared to 6. H. G. Weber, W. Thiele, M. Nakamura, etc. 1997.
the industry standard, CFC-11, the skin formation with “New C 0 2-Blown Foam Systems For Integral Skin
these two replacements is not satisfactory. Applications Including Low -Density Headrests,”
HFC -245fa is a zero ozone depletion, non-flammable presented at Polyurethanes World Congress’97, Sep.
blowing agent with an acceptable GWP. Commercial HFC- 2 0 -Oct. 1, 1997.
245fa will be available in mid-2002 supplied by 7. David Williams and Mary Bogdan. 2000.
Honeywell. The study has shown that HFC-245fa has the “Performance of HFC-245fa in integral skin foam
potential to be an ideal blowing agent for integral skin applications,” presented at Polyurethanes Conference
foam applications. Compared to CFC-11, HFC-245fa 2000, Oct. 8 -11,2000
showed a much better blowing efficiency. For a given
density, HFC-245fa loading is approximately half that of ACKNOW LEDGEMENTS
CFC-11. Polyol premixes containing HFC -245fa have
vapor pressure sufficiently low to allow them to be shipped The authors would like to thank Ron Riegal and Les
in standard drums. In molded foams, HFC-245fa gave a Bement for their assistance in the laboratory. We would
skin formation close to CFC-11, and demonstrated similar also like to thank Mrs. Mary Bogdan for her help and
behavior under the effect of mold temperature and packing advice.
level. In addition, HFC -245fa shows low in-mold pressure,
a processing advantage compared to water blown foams. BIOGRAPHIES
HFC -245fa/C02(water) blends are another cost-effective
option for the industry. The study shows that there is still B. Zhong
an appreciable skin formation even at 30mol% water level. Bin holds a B.S. degree in Chemistry
This allows customers to balance the cost and performance from Sichuan University in China and
based on the specific end use need. M.S. degree in Polymers and
Coatings Technology from Eastern
DISCLAIM ER Michigan University. Bin joined
Honeywell in July 2000. Currently he
All statements, information and data given herein are is a technical service representative in
believed to be accurate and reliable but are presented the Blowing Agent Technical Service
without guaranty, warranty or responsibility of any kind, and Applications Group at Honeywell’s Buffalo Research
expressed or implied. Statements or suggestions Laboratory and is responsible for technical support and
concerning possible use of our products are made without new product development in integral skin foam and rigid
representation or warranty that any such use is free of foam markets globally.
patent infringement and are not recommendations to
infringe any patent. The user should not assume that all J. M . Bowman
safety measures are indicated, or that other measures may
not be required. Jim holds a B.S. degree in Chemical Engineering from
Iowa State University, Ames, IA, a M.Eng. degree from
REFERENCES McNeese State University, Lake Charles, LA, a MBA from
Northern Illinois University, DeKalb, IL, and is a
1. G. Guidetti and L. Pellacani. 1997. “Advances In registered professional engineer in the State of Louisiana.
Polyurethane Semi-Rigid Integral Skin Technology Jim joined Allied Corporation (AlliedSignal, now
For The Automotive Industry,” presented at Honeywell ) in 1984 and has worked in sales, marketing,
Polyurethanes World Congress’97, Sep. 20 - Oct. 1, and technical service capacities. He is currently a principal
1997. engineer in Honeywell’s Blowing Agent Technical Service

1 5 8 / Zhong, Bowman, Williams

and Development Group with primary responsibility for
appliance industry applications of fluorocarbon products.

D. J. W illiam s
Dave holds a B.S. degree in
Chemistry from The University of
New Haven, New Haven, CT, USA.
From 1975 to 1994, he worked as a
Research and Development Chemist
and Technical Service Representative
for the Upjohn Company and later
The Dow Chemical Company in a wide variety of rigid
polyurethane and polystyrene foam application areas.
Dave joined AlliedSignal (now Honeywell) in 1994 and is
currently the manager of the Blowing Agent Technical
Service and Development Group. In this capacity, he is
responsible for technical service and product development
of Honeywell’s line of CFC, HCFC, and HFC foam
blowing agents. Dave is a member of the UNEP Flexible
and Rigid Foam Technical Options Committee.

Zhong, Bowman, Williams / 159

 Advances in Polyurethane RRIM for Automotive Exteriors

MICHAEL R HURLEY
Bayer Corporation
100 Bayer Road
Pittsburgh, Pennsylvania 15205

ABSRTACT

Elimination o f postcure in the manufacture o f RRIM established for rocker panels, fascia, and trim in Europe
automotive body panels would result in significant and in North America.4"7 The capability o f polyurea
savings. B ayflexl90 polyurea, which attains virtually all RRIM to withstand high temperature paint bake ovens
properties at demold, is a candidate for manufacturing enables it to be used in such applications as pickup truck
without postcure. Several very sensitive analytical fenders.911 Advances in the green strength o f RRIM
methods have been employed to characterize the degree panels have yielded better release and cleaner mold
of cure at demold. Differential scanning calorimetry surfaces. These improvements in processing lend
(DSC) shows no exothermic chemical reaction up to themselves to improved cycle times. Finally, the
200°C. Thermal gravimetric analysis (TGA) shows no application o f a computerized Expert Process Control
C 0 2 loss from unreacted isocyanates. And Fourier System®8,11 has brought new levels o f quality control and
transform infrared spectroscopy (FTIR) scans indicate high productivity to RRIM molding.
only trace free isocyanate in freshly molded samples. The main driving force for RIM ’s growth is the weight
The ability o f an organic composite to anisotropically saving compared to parts made o f SMC or metal. For
match thermal expansion behavior o f metallic materials is large body panels the new Bayflex 190 polyurea
highly desirable for automotive body panels such as door formulation, allows high temperature post mold
skins. B ayflexl80 high performance polyurethane has processing. Bayflex 180, a high performance
been reformulated to accept higher loading o f reinforcing polyurethane, has been developed for use in thinner
filler. Using 25% by weight loading o f fiber fillers a wallstock. Both Bayflex 180 and the Bayflex 190
coefficient o f linear thermal expansion (CLTE) o f 26 x polyurethane formulations have sim ilar properties. Their
10"6 cm/cm/°C has been attained parallel to flow into the moduli are higher than earlier RRIM systems. W ithout
mold. This value is equal to that o f aluminum. filler both exhibit a flexural modulus o f about 600 MPa.
-
Both exhibit rapid cure and excellent self release from the
INTRODUCTION mold. Especially important, both systems exhibit
excellent green strength at demold and excellent
Reaction Injection M olding (RIM) is a process for durability in use.
molding thermoset materials. In this process two liquid Bayflex 180 formulation with 20% reinforcing filler
monomers, a polyisocyanate and a polyol slurry yields a flexural modulus o f 2100 MPa. Because o f the
containing filler, are metered at a controlled volume ratio excellent toughness o f this m aterial, thin wallstock parts,
under high shear. The components mix and react as they such as rocker panels, are molded from 1.5 to 2.5 mm. At
flow into a closed m old.1 Temperatures, pressures, and higher thicknesses o f 3.0 to 3.5 mm passenger car fenders
flow are carefully controlled. Figure 1 below provides a are molded.
simple schematic for this process. An additional benefit o f the rapid cure and green
Reinforced RIM has been exploited for 25 years for the strength is found in the improved release o f molded parts
manufacture o f fascia and body side moldings.2,3 It at demold. The mold remains cleaner longer. The flash is
commands a high first run painted part capability and tougher and tends to demold with the molded part. Thus
exhibits good in p act performance. More recently weight the mold maintenance frequency is improved. This can
savings via thin wall stock technology have been significantly improve the cycle time for molding and the.

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 161

 Figure 1. Schematic of the RIM process.

economics of RIM part manufacture For these reasons temperatures ranging from 120°C to 190°C. Indeed, all
reinforced reaction injection molding (RRIM) is re- values tend to improve slightly as temperature or time of
emerging as an important method in the production of cure increase. Modulus increases from 1800 to 2200 Pa
automotive exterior applications. Today’s applications and impact test values improve.
demand higher performance in processing and part These results suggest that Bayflex 190 already has a
performance. Recent improvements in RRIM processing high degree of cure directly from the mold. It might be
include: stability at higher heat to endure E -coat oven possible to omit an oven bake prior to prime for
bake, improvements in fillers yielding easier processing at automotive body panels since the material is already
high loading, and improved toughness at high modulus. cured enough for safe handling. Subsequent high
Higher productivity RRIM processes have already been temperature bake steps during paint application would of
realized in Europe and NAFTA.8' 10 This paper describes course strengthen the material as indicated in Table 1.
two new breakthroughs in RRIM for automotive body This would provide several benefits for
panels. These are the elimination of postcure with manufacturability. These benefits include lower energy
Bayflex 190 and achievement of metal-like, isotropic consumption, decreased handling, and lower capital
thermal expansion (CLTE) with Bayflex 180. investment for new high heat body panel applications.
Overall this would increase productivity and reduced
NO POST CURE R IM manufacturing costs.
Certain questions then arise. Can the degree of cure be
In the past RRIM molded parts were required to be measured? Would subsequent high heat exposures in the
baked at 120°C and above to complete chemical reactions, assembly plant promote paint defects of un-postcured
attain complete physical properties, and de-gas parts prior parts via outgassing?
to painting. More recently RRIM has been used for pick To answer these questions a series of analytical tests
up truck fenders where it replaced SMC or steel. In these were conducted. Their purpose was to determine 1) if any
cases the RRIM material is required to be cured at 190°C unreacted isocyanate is present in un -postcured parts, 2) if
prior to prime and assembly. This was necessary to further chemical reactions are carried out in postcure, and
improve properties such as modulus and sag8 and to 3) if gases are evolved. This was followed by the
prevent out-gassing during paint processing. molding and painting of actual automotive fenders. The
Bayflex 190 is a polyurea RIM material used for results of these investigations are described below.
automotive exterior body panels such as car and truck
fenders. It is capable of replacing SMC and metal due to Differential Scanning C alorim etry
its stability at high temperatures. It also exhibits
remarkable toughness. In addition it is nearly completely This analysis was made to determine the extent of cure
cured at demold. Thus it seemed possible to eliminate or of the molded panels. Typically, unreacted isocyanate
significantly reduce the need for postcure. This was first can combine with itself, water or unreacted polyol.
evident from the physical properties measured after Reaction of isocyanate with itself or with water generates
various post cure times and temperature. Table 1 shows carbon dioxide. The carbon dioxide immediately diffuses
the properties achieved with Bayflex 190 with 20% out of the substrate. This is often referred to as “out-
Wollastonite filler. This data indicates that mechanical gassing.” If parts were primed without postcure, and
properties of Bayflex 190 do change with postcure

162 /Hurley
 Table 1. Effect of Postcure Time and Temperature on Parallel Properties of Bayflex 190 with 20% wollastonite.13

PostcureTemperature (C) 120 120 160 160 175 175 190 190 190

PostcureTime (min) 30 45 30 45 30 45 30 45 90

5 mph impact @ 23°C (J) 10.6 9.8 11.5 11.3 12.2 11.7 15.5 11.4 12.6

5 mph impact @ -18°C (J) 4.3 4.7 5.3 5.6 5.6 6.0 6.0 5.4 6.8

Density (g/cc) 1.26 1.25 1.25 1.25 1.25 1.25 1.25 1.25 1.22

Flex mod @ 23 °C (MPa) 1836 1793 1923 1966 1984 2144 2146 2108 2243

Tensile A (MPa) 39 40 40 41 40 41 39 40 41

Elongation (%) 39 30 36 25 34 34 34 44 34

Heat sag 120°C (mm) <1 <1 <1 <1 <1 <1 <1 <1 <1

Heat sag 190 °C (mm) 6 4.8 5.3 3.2 4.5 2.2 3.5 1.7 2

Notched Izod (J/m) 76 76 73 85 85 91 98 89 85

Hurley / 163
 Figure 2. DSC of Bayflex 190 after molding and before postcure.13

further isocyanate reactions ensued, gases could be This data suggests that the Bayflexl90 polyurea is
trapped in the paint causing defects. completely cured at demold since no exothermic peaks
Figure 2 below shows the result of differential scanning are evident.
calorimetry (DSC) performed on a sample of Bayflex 190
with 20% mica filler directly after molding. This data Therm al G ravim etric Analysis
reveals no further detectable chemical reaction in the
sample when heated up to 200°C. Were unreacted Gravimetric analysis determines weight loss of a
isocyanate present, an exothermic peak (downward) material during heating. Release of gas would indicate
would be evident between 150°C and 200°C. In this evolution of C 0 2 from isocyanate reaction.
region isocyanates react to form carbodiimide and Figure 3 shows the weight loss (solid line) for a sample
isocyanurates. There does not seem to be any exothermic of Bayflex 190 with 20% mica reinforcement from 0 to
peak for any other reaction under 200°C. 600°C. The dotted line is the first derivative of the weight
A small endothermic peak is observed around 57°C. loss. Remarkably, only the loss of 0.3% moisture,
This is due to the small amount o f moisture being present. coincidental to the DSC above, is found between 50 and
Post-cured samples exhibit somewhat less moisture. This 100°C. No weight loss is found between 150 and 200°C
peak is easily removed by warming to 100°C briefly. where unreacted isocyanate would generate C 0 2.
Thus the amount of unreacted isocyanate present must
be very small indeed.

Figure 3. TGA of Bayflex 190 after molding and before postcure13

1 6 4 / Hurley
Figure 4. FTIR of Bayflex 190 after molding and before postcure.13

Figure 5. DMA of Bayflex 190 after molding and before postcure.13

Figure 6. Effect o f post molding heating on Bayflex 190.13

Hurley / 165
F ourier T ransform In frared Spectroscopy sample of Bayflex 190 with 20% mica reinforcement.
Shown are three traces - the storage modulus (G’), the
To determine whether unreacted isocyanate is still present loss modulus (G* ’), and their ratio Tan a.
in parts after demold, analysis via infrared spectroscopy The curves show a glassy state below -50°C and a
was carried out. Isocyanate groups when present exhibits broad rubbery region from 0°C to 200°C. No melt
a very strong absorption peak in the infrared spectrum transition is seen beyond 200°C because the material is a
near 2300 cm'1. Using FTIR, each molded surface and thermoset network.
the interior of a molded plaque was analyzed (Figure 4). These data are nearly identical for un -postcured and
Only very weak signals for isocyanates were detected. It postcured samples. Postcured samples show a slight
is known that other moieties containing functionality increase in modulus below -50°C and a small increase in
similar to isocyanate are found in this region such as modulus between 100°C and 200°C. This is consistent
carbodiimides. Such probably, account for a portion of with a re -ordering of hard and soft segments within the
the peak observed polymer. After high temperature treatment the hard and
Thus from FTIR analysis it has been determined that soft segment separate further. This causes a more
little or no isocyanate remains. Generation of C 0 2 during crystalline state at low temperature and thus higher
paint processing is expected to be minimal. Paint defects modulus below 50°C. There is also a slight flattening in
due to out-gassing are not expected the rubbery region consistent with improvements to
impact values shown in Table 1.
P aint Line Testing
CONCLUSIONS I
At Venture Industries in Grand Rapids, MI automotive
front fenders were molded and painted using a variety of The complete kinetics of Bayflexl90 RRIM polyurea
heat treatments. This included post cure at 190°C, post result in a virtual cure of molded parts upon demold. This
cure at 120°C, and no post cure. Parts were washed via is evident from the fact that no detectable NCO was
conventional 5 step acid wash of RRIM parts in detected by spectroscopy, no release of gases occurred
production. All parts were painted with a conductive during strong heating as in TGA, and no further
prime, base coat and clear coat. Paint bake from 120°C to detectable chemical reaction took place. Further heating
170°C was applied, and a post bake of finished parts at up to 200°C yield improvements in mechanical and
190°C was included for some parts. In all cases no thermal properties such as flexural modulus, ultimate
defects such as porosity, pinholes, yellowing or blistering elongation, high temperature sag, and impact strength.
were found. These subtle changes are possibly due to hard and soft
These results are consistent with the analytical results segment domain separation in the polymer. These
discussed above. Since no significant reaction occurs changes also have some effect on the post -mold shrink.
after demold, parts could be safely painted without
postcure LOW CLTE RRIM

M old S hrink One challenge to molded composite material has been

the area of door skins. Unlike fenders or fascia, doors are
It is normal for thermoset composites such as Bayflex constrained with regard to thermal expansion. There is no
190 with 20% filler to shrink at demold. The shrinkage is where for the door to shrink along its length, so the CLTE
the result of chemical reactions between the isocyanate in that direction must match that of the supporting metal
and polyol. Upon heating further shrinkage can continue frame. Conventional injection molding technology is
as chemical reaction progresses to completion. generally not acceptable for manufacturing automotive
Figure 6 shows the change in size with heating from door panels due to excessive expansion and contraction as
120°C to 190°C for plaque samples of Bayflex 190. Parts ambient temperature changes.
continue to shrink upon heating up to about 175°C. RRIM can control the thermal expansion characteristics
Heating above that level (up to 200°C) resulted in no via type and level of fillers in the composite RRIM part.
further change in dimensions. Since this shrinkage cannot Using plate like fillers as found with mica or of flake
be due to further chemical reaction, a subtler re -ordering glass, a coefficient of linear thermal expansion (CLTE) of
of the hard and soft segment of the polymer must cause it. about 7.0 x 10"5 mm/mm/°C can be obtained. The CLTE
This is discussed further below. measured in the flow direction is slightly lower due to
orientation with the material flow. If fiber fillers are
Dynamic M echanic Analysis employed the CLTE in the direction of flow is reduced
and the CLTE perpendicular to the flow increases. For
An important tool in the characterization of Bayflex example a loading of 20% high aspect ratio Wollastonite
190 is the mechanical behavior as a function of can yield CLTE values of 4.0 x 10'5 mm/mm/°C with the
temperature as found in dynamic mechanical analysis flow and 12.0 x 10'5 mm/mm/°C perpendicular with the
(DMA). Figure 5 above shows the DMA trace of a flow.

166 / Hurley
Figure 6. M old flow analysis for door skin w ith gate located at the w indow line.

Figure 7. M old flow analysis for door skin w ith gate located at the door hinge line. Fiber filler
orientation and thermal expansion characteristics are anisotropic.

Hurley /1 6 7
 Table 2. Mechanical Properties for 25% High Aspect Ratio wollastonite Filled Bayflex 180.
Flow direction Parallel Perpendicular
Density (g/cc) 1.24 1.24
Flexural Modulus (MPa) 1530 553
Tensile Strength (MPa) 31 24
Ultimate Elongation (%) 23 47
Tear Strength, D (kN/m) 100 67
Charpy Impact (kJ/m2) 8.2 10.5
Rheometric Dart Impact (J) 4.8
CLTE (x 10"6 m m /m n/C ) 23.2 114

The approach described in this paper is the use of high When high aspect fiber fillers such as wollastonite,
fiber filler loading in a resilient RIM system matrix. The carbon fiber, or refractory fibers such as milled glass fiber
filled system is molded such that a laminar like flow is are employed, isotropic thermal expansion values similar
maintained during filling of the mold. The fibers are to metal are realized. By employing special gate design
thereby aligned in the final molded articles. Gate design fillers are oriented to minimize CLTE in a door skin
and location can be selected to maximize this effect and during. This minimizes changes in gaps between body
control the direction o f flow. panels.
Figures 7 and 8 show the detail of the approach. In
each figure a mold flow analysis is given for a door skin. ACKNOWLEDGEMENTS
In Figure 7 a door skin is shown for which the gate
location is at the windowedge or the top of the door. The The author would like to thank Bob Rowley, Dan Allman,
small lines indicate fiber orientation during filling, as and Dell Pifer of Venture Industries for their help in
predicted by the mold flow model. The fiber directions molding and painting of fenders from un-postured
are somewhat random and the overall CLTE is the Bayflex 190; Dr. Chu Shih -Chien, Joe Dettorre, George
average of the parallel and perpendicular values. Pastor, and Dr. Farhad Cama for their analyses of Bayflex
In Figure 8 the gate has been widened to nearly the 190 samples; Matt Nagy, Nathan Anderson, and Dave
entire leading edge of the door. The gate is placed 90 Rocco for their assistance in developing flow modeling
degrees relative to the initial location shown in Figure 7. and gate design for low CLTE Bayflex 180; and Bill
The material develops a laminar like flow through the Keller, Les Hamilton Jr. and Rosemary Quader for
mold. The fiber orientation is significantly different molding and determination of physical properties.
compared to Figure 7. For Figure 8 the majority of fibers
align with what would be on the car a horizontal REFERENCES
direction. This is the most critical dimension to control
since it is confined by the front fender and the rear door. 1. Oertel G., Polyurethane Handbook. Carl Hauser
Normally filler content of 20% by weight is used in Verlag, Munich, Vienna, (1985).
RRIM. By reformulation of Bayer’s Bayflex 180 a 25% 2. Ludwico W.A. and Taylor R.P., ’’The Bayflex® 110
filler loading could be attained. Using a high aspect ratio Series - The New Generation of RIM Materials,"
wollastonite, a CLTE of 23 x 10'5 mm/mm/°C is realized paper no. 770836, SAE Automotive Engineering
in the direction o f flow. This is similar to aluminum. Congress and Exposition, Detroit, MI, Sept. 26-29,
Remarkably, this material molds easily and retains 1977.
considerable elastomeric qualities and is suitable for 3. Lau C.J. and Suddaby B.R., "Opportunities for
automotive door skins. Table 2 shows mechanical Mineral Fillers in Polyurethane/Urea RIM for
properties for Bayflex 180 with 25% high aspect ratio Automotive Applications," Functional Fillers,
wollastonite reinforcing filler. Intertech Conferences, Chicago, IL, Sept. 22-3, 1993.
4. Kogelnik H.J., Rains R.C., Haradzin R.L., and
CONCLUSION II Groenen J., "Novel Polyurethane RIM Approach for
Cost Reduction and Productivity Enhancement," SAE
Increasing filler content to 25% and above yield RRIM International Congress and Exposition, paper no.
polyurethane material suitable for automotive door skins. 930541, Detroit, MI, March 1-5, 1993.

1 6 8 / Hurley
5. Kratz H.S. and Milewski J.V., Handbook of Fillers
and Reinforcements for Plastics, Van Nostrand BIOGRAPHY
Reinhold Co., New York, 1978. pp301-10.
6. Hurley, M.F., et. al.., “New Competitive
Polyurethane Materials for Automotive Fascia,” M ichael F. H urley
Proceedings of the SPI 36th Annual
Technical/Marketing Conference, (1993). Joined Bayer Corporation,
7. Hurley M.F. et. al., “Advances in Competitive then Mobay, in New
Polyurethane Materials for Automotive Fascia,” Martinsville West Virginia in
950551SAE International Congress and Exposition, 1981, working as Process
Detroit, MI (1995). Improvement Chemist in MDI
8. Expert Process Control is a trademark of Bayer AG. production. Since 1987 he has
9. James, A., 1999, “ELPO Capable Polyurea RIM: A worked in the RIM
Breakthrough Material for Composite Body Panel Automotive Exteriors Group
Application,” presented at the SAE International at Bayer’s Pittsburgh,
Congress and Exposition, February 1999-01-3225. Pennsylvania headquarters.
10. McLaren, J., and Arnold, K. 1998, “Infrared Post- During this tenure he has
Cure of RIM Body Panels: A Significant Cost developed high and low density SRIM systems,
Savings,” technical paper presented at the SAE lightweight RRIM fascia technology, and high heat RRIM
International Congress and Exposition, February 23- body panel systems. He received a Ph.D in Organic
26,1998 (980988). Chemistry from the University of Tennessee in Knoxville
11. Hurley M.F., Meiners H-J. and Doemer K-H., “High in 1980 and a B.S in Chemistry from Fairfield University
Performance Reinforced Reaction Injection Molding in Connecticut in 1974.
for Body Panels,” 99IBEC20, International Body
Engineers Conference, Detroit MI (1999).
12. Postcure at 375°F for 45 minutes prior to these
experiments.
13. These items are provided as general information only
and are not part of the product specification

Hurley / 169
 Next Generation E-Coat Capable Exterior Body Panels

COLIN BAMBRICK KEN ARNOLD

Dow Chemical Canada The Dow Chemical Company
1086 Modeland Road 2301 North Brazosport Blvd.
Sarnia, Ontario N7T 7K7 Freeport, Texas 77541
Canada

ABSTRACT primer (E-coat) cure environments. The temperatures range

from 175 - 200°C in the oven. This configuration of the
Continued development of Reinforced Reaction Injection RRIM panels, riding with the steel assembly, results in a high
Molding (RRIM) polyurea polymers for toughness, blister quality and consistent finish. RRIM panels also produce a
resistance and large-part processing as exterior vertical body high gloss and DOI product with long-term durability and
panels has launched E-Coat-compatible exterior outers into ease of in-line repair. Significant development work to meet
automotive assembly -line operations. This allows automotive these diverse demands of the design, manufacturing and
OEM design to take advantage of the unique molding shapes assembly process has resulted in polyurethane polymer
for side outers and fenders while reducing weight, assembly advancements. This paper discusses both die learning process
(DFA) and time/operations costs (DEM). Polyurea RRIM and advancements to meet future body panel requirements.
body panels have been successful in meeting the demanding
auto industry requirement for lightweight, damage -resistant INTRODUCTION
exterior outer panels as an economical alternative to steel.
Design freedom advantages, low prototype cost and tooling Polyurea RRIM body panels have been successful in
savings through predictive modelling have allowed the meeting the automotive industry demand for a lightweight,
commercial use of RRIM body panels. Having these exterior damage-resistant and economical alternative to steel, in low-
body panels pass through the E-Coat lines with the complete to-moderate volume platforms. These properties make die
steel assemblies also avoids costly space and time panels ideal for low -volume platforms with less than 200,000
requirements, as well as “add -on” body shop space. vehicle builds. Advantages in design freedom and cost have
resulted in the commercial use of these body panels in the
Since late 1998, SPECTRIM* HH390 has had outstanding Corvette, Camaro & Firebird, vehicles from GM as well as
success as a SMC and steel replacement for exterior body the Truck programs of Ford and Chrysler Dual Wheel, Chevy
panels on two GM truck programs. The experience of Silverado Sportside truck box outer panels and previously the
bringing a part from CAD (computer aided design) drawings GM APV Mini -van front fenders. The latest application are
through molding of the part, to body-in-white assembly, the 2001 / 2002 GMC & Chevy Dually Truck as well the
through the steel treatment process including e -coat ovens Fleetside Composite Box & Quad Steer Outer panels, which
which reach 205°C, and the painting and assembly process are the commercial application of this E-coat-capable
gave us the direction for product enhancements that would polymer. These applications include both E-coat-capable and
benefit both the OE and Tier 1. This high-temperature- non-E-coat-capable SPECTRIM products from Dow.
- painting,
resistant polyurea RRIM composite allows on line
including passing through the steel corrosion protection This E-coat-capable RRIM material, can be processed in
existing assembly plants, which require the plastic body
panels to not only be painted on -line before general assembly,
*Trademark of The Dow Chemical Company

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 171

but also to pass through the E-coat process required for the Image) [8]. RIM products and specifically the Body Panel
steel and metallic parts. This process requires the products, meet tough performance requirements for vertical
electrodeposition of a corrosion-resistant coating over the panels, including high impact resistance, part stiffness and a
steel surface. The subsequent curing of this primer coat in a durable high quality surface finish after topcoat [5].
E -coat oven requires an elevated temperature above 175°C
and achieves peak ambient air temperatures higher than BP80 BP80 HH400 HH390
21% 22% 20% 21%
204°C. The total oven time is more than 45 minutes and peak Mica RG Mica RG
temperatures may persist for more than 8 of those minutes. Flexural Modulus para 1,900 1,900 1,700 1,870
While these RRIM panels do not require corrosion protection, (MPa) P«P 1,600 930 1,500 1,030
attaching the plastic panels along with steel panels before the CLTE para 57 36 57 55
OO"6 in/in -'C ) perp 60 132 61 141
E -coat process (Body -in -White stage) avoids costly assembly ­ Heat Sag (150 mm) 121°C 4 4 1 1
line logistics and factory floor space. The new polyurea RRIM (mm) 191°C NR NR 5 2
systems were developed specifically to flow through these Tensile Strength para 27.6 29.5 31 34
(MPa) perp 30
existing assembly -line logistics. A typical assembly process Elongation para 7.5 96 30 53
schematic (Figure 1), illustrates the two possible locations j m _______________ perp 74
where plastic body panels could be attached to the vehicle.
Table 1 Body Panel Materials

CLTE and Heat Sag

The dimensional performance of a polyurea RRIM body

panel in a high temperature environment is related to two
thermo-mechanical properties of the molded material, its
CLTE (Coefficient of linear thermal expansion) and its high
temperature heat sag.

The CLTE indicates the degree of expansion and

contraction of a material during heating and cooling cycles. A
typical CLTE for steel would be below 10 x 10E -06 mm/mm
°C versus RRIM averages of 50 to 60 x 10E -06 mm/mm °C.
The differences in material expansion and contraction
properties are important when considering overall
dimensional change ofthese panels at high temperatures. The
gap required to accommodate the part growth is important to
avoid any interference during elevated temperature processing
steps in the assembly plants, such as E -coat.

The Heat Sag of a material is a measurement of the

Figure 1 - Exterior Panel Assembly Process
deflection of a specified, unsupported horizontal sample
length after a specified temperature exposure. A low value
In order to meet these stringent requirements and extreme enables the part design to have elements of curvature or
conditions for plastic panels, considerable product horizontal sweep. This characteristic can allow a great degree
advancement was done in an accelerated time frame. The of design freedom in creating brand distinctiveness in truck
highlights of this development cycle will be discussed further. outers or front body panels.
NEW HIGH -HEAT POLYUREA BODY PANEL Low heat sags 5mm or less, tested at 191°C, can be
CHARACTERISTICS achieved with the new high -heat material. The postcure
temperature of 185°C allows the part to retain shape without
SPECTRIM HH400 is the next generation chemistry that distortion during the subsequent E -coat temperature exposure
meets additional performance objectives listed in the Product on the final assembly process.
Evolution section. The typical mechanical properties of the
new High -Heat and other Polyurea RRIM body panels are Blister Resistance
listed in Table 1, whether the parts must be E -coat capable or
can bypass this step. Mineral fillers comprise 20% by weight -
The E coat capability of a plastic also requires the moisture
of the final molded part, and fall into two highly refined and contained within the molded part, constantly in equilibrium
classified commercial mineral products. Treated wollastonite with its environment, to diffiise out of the part during the
and mica are used and each mineral filler provides its unique -
heating within the E coat oven. This must be accomplished
challenges to overcome. These mineral fillers are preferred without the degradation of the mechanical properties of the
over glass fillers, both milled and flaked, due to the lower cost polymer by the superheated moisture. The new high heat -
and improvements in Painted Part DOI (Distinctiveness of

172 / Bambrick, Arnold

polymers have passed rigorous testing, including 10-day process for this material begins. This process sets the stage
water soaks of the polymer parts, both at room temperature for the development of two materials that could meet the
and 38C. With this methodology, die sample is then put into major design criteria for truck side outers, which are highly
an oven, pre -heated to 199°C, and the sample must be kept at shaped and curved panels, as well as the more conservative
that temperature for 1 hour without blister defects forming on and flatter panels associated with the average truck buyer.
the surface.
Performance Objectives ( O riginal)
Panel Design and Attachments
• Stiffness^ 1000 MPa in any flow direction
The dimensional stability of body panel outers depends on • Tensile Strength 27 MPa psi inall directions
the panel design and attachment methods, as well as the • Toughness equal to non-E -coat BP 80
material character. An RRIM materials supplier has provided • Meet the heat sag criteria of <5mm at 191°C
significant support to ensure a number of Truck Outer panel • Capable ofusing standard rigid paints
designs meet the designers requirements of being functional, • DOl and gloss meets OEM standards
attractive and secure. • Shrinkage after postcure equals non -E coat BP80
• E -coat meets OEM specification
PRODUCT EVOLUTION
Performance Objectives ( Additional)
The evolution of body panel materials can be seen in
Figure 3 which highlights the first commercial application • Platelet Fillers must pass 2 day water soak testing
and then subsequent product evolution. The selection process and E-coat Simulation to 204°C
can be aided by the use of the Performance Triangle, which • Dynatup Toughness Target of 12 Joules at -30°C
aids in the design process to determine the right materials • Mold Release(s) between EMR Application to increase
(Figure 4) once the performance objectives have been
five (5) fold
established, and then the design criteria for the part can be
used to folly define the part’s material character.
Product Development

The more highly curved parts, developed using high-heat

RRIM with wollastonite filler, must be put through FEA and
cooling analyses in order to ensure a distortion free part.
Flatter surfaces must use a functional mica filler while parts
with moderate shape may use a combination of the two.
Development requires several iterations to achieve timely
success with minimal rework. The iterations to success
include:

• Finite Element Analysis (FEA) simulations using global

scale factors in three dimensions to understand and
minimize distortion due to material.
Figure 3 Evolution of Body Panel Materials • Diffraction analysis of Beta, Prototype and Production
molds to ensure all distortions or irregularities are
benched out the tool before it’s first use to manufacture
parts.
• Diffraction analysis of Beta, Prototype and Production
molded parts throughout its processing to ensure no
distortions or irregularities are appearing. Parts are
imaged using Diffraction techniques after molding, post -
cure, priming, and shipping to the assembly plant trials.
This imaging system shows the lack of distortion or
flatness of an object and stores it for comparative
measurements. Then the parts are topcoated and
measured again and the records are kept for future
reference.
• Standardized and certified postcure fixtures of aluminum
that are meant to support the part and offer some heat
Figure 4 Performance Triangle sink capability.
• Standard Production and CMM fixturing measurement
The performance objectives step sets the performance metrics at each stage to ensure fit and flush requirements
hurdle for the vertical outer panel and then the development are met or corrected in the next production stage.

Bambrick, Arnold / 173

 Thermoset Labs for part production. David Rock, Mark
The more conservative and traditional designs, which tend Felzien & Thomas Worswick (General Motors Truck)
to be flatter , must use a non-directional mica filler with the Design/Release Engineers, Allan James, Dow Automotive
HH 365 grade and be supported in the E-coat and assembly Thermosets Lab, and Martin Cornell, Dow Automotive
process, if E -coat oven temperatures exceed the maximum Scientist for their guidance. We also thank Dow Automotive
value. & the Dow Chemical Company for their investment &
The new high -heat body panels were also tested for support of the RIM program, through people and technology
optimum postcure and E -coat exposure temperature to support.
determine the combination that retained key physical
properties while maintaining the blister resistance that is REFERENCES
required to ensure part appearance and integrity. The E -coat
simulation exposed the part to 190°C for 30 minutes, after 1. Schrott, B.D., and Berg, J.W., “Improvements in RIM
soaking in room temperature water for 2 days (an equivalent Processing for Automotive Fascia” technical paper
of greater than 1% moisture absorbed), a intensely more 900425 SAE International Congress and Exposition,
severe test than the actual assembly line process. Parts are Detroit, MI, February 26 - March 2, 1990.
tested in the assembly plants throughout the development 2. Best, James R , et. al., Automotive Plastics Report - 88,
process, so that manufacturing concerns and capabilities are Market Search, In., 1988.
continually being met. 3. Cornell, M.C., “Improving RIM Economics” technical
paper SAE International Congress and Exposition,
Flat Part Development Detroit, MI, February 25 - March 1,1991.
4. Hemphill, J.J., “High Performance Polyurea RIM for
Design for more conservative brand or product lines often Automotive Body Panels”, technical paper 900285 SAE
require the designer to follow the limited steel design, which International Congress and Exposition, Detroit, MI,
results in a part, with less cross-car dimension. These ‘flat’ February 26 - March 2,1990.
parts require a filler that has a much more uniform shrink in 5. Martinez, E.C., and Sato, T.J., “RIM - A Process for the
three dimensions. The best mineral filler for this application 90’s”, technical paper 890698, SAE International
are the monoclinic prismatic micas (Potassium / Alumina / Congress and Exposition, Detroit, MI, February 27 -
Silicon Complex) with, again about 20% by weight in the March 3, 1989.
final polymer. Since the parallel-to-perpendicular properties 6. Macosko, C.W., “RIM Fundamentals of Reaction
are much more uniform in three dimensions, there is no Injection Molding”, Hanser Publishers, New York, NY,
longer a strong need for a FEA analysis. All other iteration 1989.
steps to success (listed in previous section) are necessary to 7. Hemphill, J.J., “High Performance Polyurea RIM for
ensure the parts are as tough and equal or superior to steel in Automotive Body Panels”, technical paper 900285, SAE
overall quality of appearance. International Congress and Exposition, Detroit, MI,
February 26 -March 2,1990.
CONCLUSION 8. Rettmann, K.J., and White, D.P. “High Temperature
Performance of RRIM Polyurea Body Panels”, technical
RRIM body panels have become a competitive alternative to paper 910859, SAE International Congress and
steel and SMC composite side outers due to their favorable Exposition, Detroit, MI, February 25 -March 1,1991.
production economics and their performance on vehicles with 9. Bambrick, C.R et al, technical paper 98098, SAE
less 200,000 total builds. A high thermal stability polyurea International Congress and Exposition, Detroit, MI,
RRIM body panel system has been in use & advanced, which February 23 -26,1998.
not only shows a strong balance of mechanical and thermal 10. James, A. and Miller, T., " Performance Comparison of
properties, improves productivity on Tier 1 molding presses, Plastic Composites with Metals for Vertical Body Panel
but also demonstrates minimum distortion through the Applications ", technical paper 1999-01-0848, SAE
assembly process. International Congress and Exposition, Detroit, MI,
March 1 -4,1999.
The development an improved product is geared to serve
two distinct design philosophies. The rounded and highly DEFINITIONS, ACRONYMS, ABBREVIATIONS
curved designs, as well as the conservative or ’flatter' designs
can use the HH 400 grade with a platelet mica filler or DFA Design for Assembly
proportions of a highly directional filler (wollastonite), with DFM Design for Manufacture
no effect on appearance or performance and which passes DOI Distinctiveness of Image
through the assembly process unsupported. E-coat Electro-lytic Phosphate Oven
EMR External Mold Release ( Soap Based)
ACKNOWLEDGMENTS FEA Finite Element Analysis
RRIM Reinforced Reaction Injection Molding
The authors would like to gratefully acknowledge the CLTE Coefficient of linear thermal expansion
following people for their support of this project: Ken Arnold, OEM Original Equipment Manufacturer
Rick Burr and Scott Derbyshire from the Dow Automotive

174 / Bambrick, Arnold

BIOGRAPHIES Ken Arnold

Colin Bambrick Ken Arnold is the Dow

Automotive Focal Point for
Colin Bambrick is an Application Automotive Polyurethanes in our
specialist for Dow Automotive Texas Lab. He joined Dow in 1978
RIM Thermosets based at our and has spent 21 years in
Heavy Automotive Laboratory in Polyurethane RIM Technical
Sarnia, Ontario. He joined Dow in Service.
1978 and has spent 10 years in
Polyurethane RIM Technical
Service. Colin has been involved in
the launch of each of the new
Truck platforms using Spectrim
HH Products.

Bambrick, Arnold / 175

High Performance Structural RIM for Large Composites:
Beams, Boxes, and Beyond

GENE SYMOSKO, RON CAGEAO, DAVE WALTER GUARNIERI

D’ERRICO a n d BIN LEE Bayer AG Leverkusen
Bayer Corporation PU-SM-AUT-EX
100 Bayer Road B-103 Leverkusen
Pittsburgh, Pennsylvania 15205 Germany

PAT DEPALMA
Meridian Automotive Systems
1890 Riverfork Dr
Huntington, Indiana 46750

ABSTRACT combination with the strength of fiberglass provided a

composite material that could be used to mold shaped
Bayer Corporation has been engaged in extensive body components at a relatively lower cost than steel.
development o f high -density structural RIM (HDSRIM) The sleek design of the car provided a showcase for
materials during the past several years. This paper "fiberglass parts" that were tough and dent resistant.
addresses HDSRIM technology from its infancy as Since that time, almost 50 years ago, structural
butnper beams to its present maturity as full-size pick-up composites have expanded into a wide variety o f uses
including automotive, agricultural, industrial, aerospace,
truck composite boxes and tailgates. This paper also
and recreational applications.
discusses implementation of the HDSRIM technology to
For many years, polyester resin based structural
those areas beyond the present applications.
composites formed the staple of the technology.
Recent technology developments have enabled the
However, during the mid-to-late 1980’s a new line of
manufacture of large composite parts, and this paper
fiberglass composite products came into existence for use
focuses on the development work surrounding the
in the automotive industry. The new products combined
composite pick -up truck box. Other current applications
conventional urethane reaction injection molding (RIM)
are touched upon and there is a detailed comparison of
with fiberglass reinforcement to form structural materials
HDSRIM to competing technologies via technical process
known as SRIM. The SRIM products could be used in a
descriptions. Also, new technologies for the production
wide range of applications from light structural interior
of HDSRIM parts are presented and compared. An
automotive door panels to highly structural bumper beam
HDSRIM market growth potential is provided as a
supports. This breadth of applicability was due, in part, to
culmination of the paper.
the relative ease by which the percentage and type of
glass could be varied. The overall composite density of
INTRODUCTION
the finished SRIM part could be tailored to meet the need
of the specific application.
With the advent o f the Chevrolet Corvette in 1953 the
public was introduced to structural composite materials.
HISTORY
While development work on composite materials had
taken place prior to this in other industries, this was the
Early applications of high -density structural RIM
first application where the technology was available to the
(HDSRIM) materials involved relatively small parts with
general public. The formability of polyester resins in

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 177

 simple designs. The general belief was that the urethane factors. All three of these materials are presently used for
reaction was too fast for molding large parts and that pick-up truck boxes and other large structural parts.
fiberglass preform technology was not suitable for HDSRIM offers the highest durability and the greatest
complex part designs. The first major breakthrough for weight savings of the three.
the technology involved the production o f bumper beam The ability o f HDSRIM to perform well in the
supports. The part was moderate in size and did not manufacture of a pick-up truck box is a direct result of
require an overly complicated preform. These factors several years of concentrated development. Improved
coupled with advances in urethane chemistry and longer flow polyurethane chemistry and robotic controlled
processing techniques allowed production of good quality preform technology are the two keys that make it possible
parts. This success led to work on other automotive to produce large complex parts with consistent glass
applications such as package shelves, floor pans, and loading. The high performance polyurethane SRIM
seatbacks. In nearly all of the cases the work was targeted system, Baydur®425IMR, allows longer shot times (10-12
at parts which were non-show. The parts acted as seconds) and improves initial wet-out of the glass
structural support but were not be seen directly by the preform. Processing the urethane via a squeeze-pour
consumer. molding technique is also extremely important. Squeeze
A major step was undertaken when the idea of molding involves dispensing the mixed polyurethane
producing a composite pick-up truck box was presented. chemicals into the tool while the tool is partially open
This was an exterior part of very large size that could (usually 0.25" to 0.50"). After the material is completely
potentially be subjected to severe abuse over several years dispensed, the tool is closed and the urethane is pushed
and would still need to look good. A multi-corporate through the glass mat using the clamp tonnage o f the
taskforce was formed to address the questions inherent in press rather than the injection pressure of the material.
a project of this size. After analyzing several competing This process evenly distributes the material within the
technologies HDSRIM was chosen as the one with the tool and prevents the movement o f the glass fibers around
greatest chance o f success. Consequently, General the mixhead sprue area of the part.
Motors made the decision to spearhead the production of
a pick-up truck box and tailgate for the GMT800 vehicle. HDSRIM PRODUCTION PROCESS
They assembled the necessary raw material suppliers in
order to turn the idea into a reality. Many lessons were High density SRIM is molded using machinery that is
learned along the way and as a result the SRIM process similar to other two component urethane processes.
has proven capable of producing large complex parts. Pump or cylinder style metering units mix the polyol and
isocyanate components at the mixhead through high-
HDSRIM AVANTAGES pressure impingement. The metering unit and mixhead
must be properly sized in order to accommodate the high
High-density structural RIM technology provides a mix-ratio requirement, the low viscosity of both
higher strength, lighter weight material alternative to steel isocyanate and polyol components, and the low chemical
or other composites for automotive structural throughputs of the process.
components. Table 1 below provides a comparison Utilizing HDSRIM, large parts can be molded using
between HDSRIM, Steel, and SMC (sheet molding clamping units that are significantly smaller and develop
compound) considering several cost and processability less force than competing technologies. Press tonnage
requirements are dependent on the part size and geometry.
Single cavity flat panels, such as a tailgate, typically
Table 1. Automotive Structural Component Options require a press capable of producing 400-600 tons. A
truck box requires a press capable of producing 2,500-
Parameter HDSRIM Steel SMC
3,000 tons of clamping force.
Tooling costs 0 ++ ++ The fiberglass reinforcement is a key component of
Material costs ++ 0 + the composite and is responsible for its inherent strength.
Equipment costs ++ 0 + Different glass types, amounts, fiber lengths, and binders
can significantly alter the physical property characteristics
Press requirements 0 ++ ++
of the part. Glass percentages can vary from 25% for a
Weight savings ++ 0 + part with moderate physical property requirements to 55%
Recyclability - 0 - for a part with higher property demands. Flat panels such
Durability ++ 0 + as a tailgate can be easily molded using glass mat. Glass
mat can be purchased in large rolls and cut on site or die
Large parts - single Y N N
cut by a glass supplier. The glass mat is usually more
operation
costly than directed chopped glass preforms but it can be
faster, easier to handle, and will eliminate the need for
Key: 0 = nominal; + = higher; - = lower preform equipment. However, molding complex parts
can be challenging using glass mat, particularly at the
higher glass loadings. Wrinkling and glass retracting

1 7 8 / Symosko, Cageao, D ’Errico, Lee, Guarnieri, DePalma

from the part edge are the typical problems encountered. Early in the truck box program there were
Consequently, directed chopped glass preforms are a problems with the composite blistering at postcure
necessity in order to manufacture complex parts with temperatures. An SEM (Scanning Electron Microscopy)
higher glass loadings. analysis o f the blisters produced at elevated temperatures
Chopped glass preforms are made by spraying glass indicated clearly that only dry glass (i.e., glass not wet-out
onto a preform screen that mirrors the part's geometry by urethane) existed inside the blisters. Figures 1 and 2
using a robot controlled chopper gun. A vacuum is pulled show clearly that the glass fibers are not coated with any
through the preform screen directing the cut fibers onto urethane and the glass fibers have not pulled out of the
the screen. A series of baffles help regulate the flow o f air urethane matrix. This behavior is independent of which
through the screen. Emulsion and powder binders are side of the blister is viewed - substrate side or underside
applied to the glass during this chopping process. The of the blister. The blisters were undoubtedly a
robot (or robots) assures that the glass is consistently and consequence of voids in the composite caused by poor
evenly sprayed onto the screen from part to part. Binders urethane wet-out o f the glass fibers.
must be oven-cured after the preform spraying is It was recognized that a molded composite was
complete. This is accomplished by placing the preform unlikely to be void free even when molded in a highly
into an oven at 350° - 400°F. Curing parameters can vary controlled environment. However, it was also recognized
dependent upon the binder type and amount. that larger voids would produce a lower blistering
The tooling used in HDSRIM applications is often temperature than very small voids. The goal was to
equipped with shear edges (i.e., by-pass cutters along the minimize the void size through chemistry and processing.
part perimeter) to effectively trim the fiberglass just The HDSRIM system, Baydur® 425IMR, was developed
before complete tool closure. This can be done with with a slow gelling characteristic to allow the urethane to
either glass mat or directed chopped preforms. The shear fill the entire truck box before building significant
edges not only pre -trim the part to make the final molecular weight. This means that the urethane would be
trimming operation simplier, but also act as a partial seal liquid longer and thereby more likely to wet-out the
to prevent leakage of the low viscosity urethane material. fiberglass surface. The long flow was coupled with a fast
Once the HDSRIM part is demolded and trimmed it cure characteristic that allowed parts to be demolded
is postcured in an oven. The postcure temperature is quickly.
dependent upon the maximum temperature that the part Two processing aids that were studied were vacuum
will be subjected to in any of the postmolding operations. application and higher molding pressures. Applying
In the case of the tailgate and truck box this temperature vacuum to the tool just prior to closure reduced the
is fixed by the requirements of the bonding operation. amount of trapped gases, assisted the urethane flow, and
After postcure the part is washed in order to prepare it for produced parts with superior physical properties due to
painting. better fiber wet-out. The higher molding pressures
HDSRIM can be top-coated with normal automotive favored urethane penetration and reduced the void sizes in
finishes and does not require any special operations the finished part.
beyond wash. In the case of a tailgate and truck box the
tool surface is grained so the part is produced with a non
class-A surface. A spatter type coating is applied to the
parts to enhance the grained surface appearance.

COMPOSITE OPTIMIZATION

One of the key lessons learned in the HDSRIM truck

box development was the importance of the interaction
between the urethane matrix and the fiberglass
reinforcement. The tailgate and truck box composite
material is targeted at 50% by weight each of fiberglass
and urethane. Each component is important to the
functioning o f the composite, and if either is not
compatible, the final composite will not be good.
Basically, the issue is fiber wet-out.
The glass fiber preform plays a very important role
in determining the ultimate physical properties and the
heat performance of the SRIM composite. Factors that
can affect the performance of the preform include glass
type, binder type, and glass distribution. In turn, these
factors can also affect the ability of the urethane to fully
wet-out the surface of the glass fibers. Figure 1. Underside of blister

Symosko, Cageao, D yErrico, Lee, Guamieri, DePalma / 179

 NEW TECHNOLOGIES

While Bayer Corporation as a company is concerned

with providing optimized polyurethane systems, it is also
intimately involved in new technologies for the
processing of those systems. One area of particular
interest for HDSRIM systems is in chopped fiber injection
(CFI) technology. A CFI technology is offered by each of
the three major urethane equipment suppliers, Cannon,
Hennecke, and Rrauss-Maffei, and each is capable of
producing parts with 50% glass loading. The CFI process
provides a way for glass fibers to be simultaneously
injected with die urethane material during part production.
The glass is fed to a chopper near the mixhead from
standard glass rolls.
The CFI process offers a molder some distinct
advantages. This process eliminates the handling issues
Figure 2. Substrate under blister for molders that use glass mat. If the molder is using
preform technology and wishes to expand there is no
longer the need to install another preformer. New
Also, the effect of surface treatment on the glass and molders wishing to start up an HDSRIM program do not
the effect o f binder chemistry were studied. These two need to purchase preform equipment. In each case there
variables control the interfacial adhesion between the is significant monetary savings to be realized by
glass surface and urethane substrate. The commonly used incorporating a CFI technology. Lastly, the use of a CFI
method to study interfacial adhesion in composites is to process can make the economics of HDSRIM more
perform SEM analysis on cryogenically fractured attractive.
surfaces. Visual urethane deposits on the glass fibers are There are various limitations on the CFI technology.
a sign o f good interfacial adhesion. An example of this Certain parts may be too complicated due to deep draw
behavior is shown in Figure 3 for the combination of sections or steep sidewalls, to be produced consistently
Baydur® 425IMR and OCF 8610 glass mat. The via the process. Other parts may simply be too large to
complete opposite behavior is seen in the SEM of Figure pour in the requisite amount of time. To address the
4 which shows Baydur® 425IMR in combination with a potential shortcoming of part size, Bayer Corporation has
chopped glass preform containing liquid polyester binder. developed an HDSRIM system with a greatly extended
Polyester binder is commonly a peroxide cured material gel time. The experimental chemical system has been
that is highly crosslinked by design to impart good heat successfully poured via the CFI process in several
resistance. The study found that completely cured binder different parts. The system offers a long pour time yet
is required to inhibit its migration to the part surface and maintains a quick cure time. Parts are currently in testing
cause blistering. to determine the viability of the process, but initial
indications are good. The process is ideal for parts with
minimal curvature and weights below 30 pounds.

Figure 3. Baydur®4251MR / O C F8610 Figure 4. Baydur* 425IM R / Chopped glass preform

1 8 0 / Symosko, Cageao, D fErrico, Lee, Guarnieri, DePalma

MARKET POTENTIAL BIOGRAPHIES

At the present time a major focus for Bayer Gene Symosko

Corporation is in the area of HDSRIM for the light truck
market. This material is an excellent fit given the Gene Symosko began his career
manufacturer's need for extreme durability and weight with Bayer Corporation in 1976
reductions in comparison to steel. The chance for as a Laboratory Technician. In
expansion to other OEMs is a distinct possibility given 1989 he joined the Automotive
the success of the GMT800 truck box program. SRIM group as a Technical
Production figures for General Motors, Ford, and Service Representative. In 1993
Dodge indicate an excess of two million light trucks he was promoted to Process
manufactured per year. Using the current production Chemist and began working on
scenario and amounts of urethane usage per truck box, a the large composite project soon
total urethane potential market of over 100 million pounds afterwards. Since 1995 he has
can be estimated. This is indeed a significant market for
worked exclusively on the development o f large
composite materials. The small truck market is another
composite technologies. He is currently Launch Team
untapped potential for HDSRIM applications.
Coordinator with responsibility for coordinating Bayer
Conversion o f these vehicles to an HDSRIM box and
Corporation's technical service coverage in this area.
tailgate can increase the market by at least 50%.
Since HDSRIM components have a proven
Ron Cageao, Ph.D.
production experience it is possible to work with the
automotive engineers to specify the material earlier in the
Dr. Ron Cageao received his
design process. This can lead to many different parts
Ph.D. in Macromolecular
using HDSRIM as the material o f choice and thereby
Science from Case Western
expand the market well beyond initial estimates.
Reserve University in 1988.
Information gained from this experience with the
Shortly thereafter he joined
GMT800 box program has lead to a streamlining of the
Bayer Corporation and began
manufacturing process. This information can be used
working in the Automotive RIM
along with new production technologies to expand
group. He has held positions of
HDSRIM to other markets. Opportunities exist in the
increasing responsibility and is
recreational, agricultural, and industrial sectors that are
presently a Development
similar in scope to those in the automotive sector.
Scientist in the Automotive
SRIM group with a focus on low density systems and new
CONCLUSIONS
process technologies.
Important information has been gained from the
Dave D 'Errico
experience of the GMT800 box program Studies of the
interaction o f the materials in the manufacturing process
have advanced the understanding of HDSRIM technology Dave D'Errico joined Bayer
Corporation in 1986 as a
significantly. The incorporation of an HDSRIM
Laboratory Technician. After
technology into the manufacturing process of large
complex parts has become a reality. HDSRIM has proven gaining knowledge in several
to be a higher strength, higher durability, and lighter aspects of materials and
processing he moved into the
weight alternative for automotive structural components.
technical service area of
Automotive Polyurethanes in He
has been with the Automotive
SRIM group since 1994 and is
presently a Technical Service
Specialist with expertise in both low-density and high-
density SRIM.

Symosko, Cageao, D ’Errico, Lee, Guarnieri, DePalma / 181

Bin Lee, Ph.D. the Central Research Department as a Research Scientist.
After more than 3 years with several different research
Dr. Bin Lee received his Ph.D. projects completed, in 1999 he joined the PU Marketing
in Material Engineering Science Group as head of the Automotive Baydur/Baypreg team.
in 1987 from the Virginia His responsibilities included all marketing and
Polytechnic Institute and State development issues associated with these products. Since
University. Since he joined July 2000, he was promoted to the Global Strategic
Bayer Corporation in 1987, he Marketing Automotive group. He is currently the Team
has held many development Leader for theBaydur/Baypreg Marketing and
positions in the Rubber Development Group for all of Europe.
Department, Polyurethane
Research, Central Research P at DePalm a
(Bayer AG, Germany), and
Corporate Polymer Research. He is presently a principal Pat DePalma received his B.S. in Chemical Engineering
scientist in the Polyurethane Automotive SRIM Group from Lawrence Technology University in 1974. He
active in the truck box development. became involved in the Composite Industry in 1975 as a
Material Development Engineer with GenCorp
W alter G uarnieri, Ph.D . Reinforced Plastics. Since, he has held several positions
in engineering and manufacturing areas. Currently, he
Dr. Walter Guarnieri received his Ph.D. from the has Manufacturing and Engineering Operations
University of Muenster (Germany) in Organic Chemistry responsibility for the production launch o f the GM
in 1995. He then joined Bayer AG and began working in Composite Box Program at Meridian Automotive
Systems.

182 / Symosko, Cageao, D ’Errico, Lee, Guarnieri, DePalma

 New PU Cavity Filling Foam for Sound Abatement in
Car Body Shells

SVEN MEYER -AHRENS a n d GARY S . KARAS a n d MICHAEL LUPINI

CHARLES M. MILLIREN Orbseal LLC, Inc.
Bayer Corporation 14700 Helm Court
100 Bayer Road Plymouth, Michigan 48170-6062
Pittsburgh, Pennsylvania 15108
ABSTRACT

Sophisticated noise control techniques are in great material temperature (175 °F) required to lower the
demand for modem automotive interiors. The increasing viscosity into a range that allows the use of low pressure
use o f cell phones and newly emerging voice control mixing equipment. Both drawbacks lead to production
applications as well as the greater perceived comfort of a difficulties.
quiet vehicle interior fuels this demand [1]. Classical noise This paper introduces a new low MDI acoustic foam
control measures such as dash insulators, carpeting, seating system for barrier applications in the automotive industry.
and other acoustically absorptive interior parts are already Physical, acoustical and industrial hygiene data will be
in wide use. However, cavity filling PU foams are a presented and compared to previously released data of
relatively new class of sound abatement materials [2], [3]. other classical and low MDI systems.
These foams reduce the propagation of noise through the
hollow spaces in the vehicle body shell [4]. DESIGN GOALS

Bayer Corporation has developed a PU foam system, Low MDI foams are, by definition, foams which exhibit
which will be marketed by Orbseal to Automotive OEM’s. low levels of residual airborne MDI monomers after the
It expands to 5000% o f its initial volume. The low-density foaming process is complete. The airborne monomers can
foam resulting from the high expansion rate allows for a result in industrial hygiene issues and associated costs for
complete acoustic seal and provides material savings of up the manufacturer in auxiliary equipment, such as
to 30% while still meeting OEM requirements on Insertion downdraft ventilation, required to reduce the airborne
Loss. The new foam system was specifically designed to monomer concentrations to acceptable levels. The
meet OEM demands for robustness and safety when American Conference of Governmental Industrial
dispensed on an assembly line in a paint shop general Hygienists (ACGIH) has adopted a Threshold Limit Value-
assembly application. As a result, this foam provides a low Time Weighted Average (TLV -TWA) for methylene
foam surface temperature (<137 °F) and a rapid viscosity bisphenyl isocyanate (MDI) of .005 parts per million
build up to overcome sagging problems. The emission of (ppm ), i.e. 5 parts per billion (ppb) or .051 mg/m3. This is
residual MDI during the foaming process is extremely low an airborne concentration for a normal 8-hour workday and
even though the foam is produced using a 1:1 mixing ratio. a 40-hour workweek and represents conditions under
which typical workers can be exposed without adverse
INTRODUCTION effect. In addition, the federal Occupational Safety and
Health Administration (OSHA) has established a
Polyurethanes, including two -component water blown Permissible Exposure Level (PEL) for MDI of .02 ppm or
polyurethane foam systems, are a major contributor to .2 mg/m3as a Ceiling (C) limit. The Ceiling limit is a
minimizing interior noise levels in passenger vehicles. concentration that should not be exceeded during any part
Engine covers, carpet underlay, headliners, dashpad of the day. At present, neither an ACGIH TLV nor a
insulators, seating systems and, more recently, new so- federal OSHA PEL has been established for polymeric
called low MDI cavity filling foams have been successfully MDI.
used for vehicle noise abatement. One drawback to the use Therefore, one of our primary goals was to achieve an
o f these last foams is the extreme mixing ratio required to acoustic MDI based foam, which passed the ACGIH TLV
produce these foams (P/I = 1/24). Another is the high and the OSHA PEL requirements. A foam currently on the

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 183

market meets these requirements at mix ratio of P/I = 1/24 These process requirements include the following points.
[5]. Because o f the inherent difficulty in achieving a The acoustic foam must be applied after the E-coat and
reliable mix at a P/I ratio of 1/24 as seen in other low MDI cure at ambient temperature. Dimensional stability of foam
foams a goal was set of achieving a low MDI foam system was achieved under extreme conditioning of 24 hours at
with a P/I ratio of 1/1. Additionally, a goal was set to 250 °F. The foam must adhere to the E -coated sheet metal
develop a low viscosity system, which would allow the and be applied with a programmable shot size to meet the
user to run at or near room temperature and reduce the design specifications, shot repeatability, using
thermal hazard to personnel. Another goal was to further impingement mixing. The foam reactivity must be
reduce the density of the acoustic foam while still meeting sufficient to keep the foam where it is applied at a delivery
the manufacturers requirements for Insertion Loss. Further rate of 75-150 g/sec.
we needed to develop a foam system with sufficient
physical properties to meet manufacturing process DEVELOPM ENT RESULTS
requirements.
The results of our development process is a low MDI
acoustic foam, Bayfill® NV, which has the following
characteristics:

Table 1. Properties of New Low M DI Foam

Mix ratio (P/I) 1:1 by weight, 1.19:1 by volume

System temperature (F) 80-100
Gun temperature (F) 80-100
Cream time (sec.) at mixhead
Gel time (sec.) 3
Rise time (sec) 6
Tack free time (sec.) 6
Expansion (%) >5000
Surface temperature (F) 137
MDI emissions (ug, ppb) <.15, <1.1 (test method as described in [5]
Free rise density (pcf) 1.2-1.5
Adhesion (various cycles) pass
Compressive strength (kPa) 26.89
Dimensional stability (%)
20 min after foam manufacture
(following 15 min @120 C, 250 F) <2.5
24 hours @ 80 C (175F) <1
100 F @ 50% RH <1.5
168 hours @ 98-100% RH, 38 C
(100F) <1
Dispensed weight loss (%) <4
Heat aged weight loss (%) <2
Cold slam resistance pass
Flammability pass
Fogging pass

All properties measured in accordance with Daimler Chrysler MS-CD 643

1 8 4 / Meyer-Ahrens, Milliren, Karas, Lupini

 Figure 1. Insertion loss of test channel filled with Bayfill® NV, testing details as in /67-

PERFORM ANCE BENEFITS SUMMARY

This new acoustic foam, Bayfill® NV, is the first

member of a series of Bayer Corporation NVH foams,
which ORBSEAL will distribute under the marketing
umbrella “Engineered Polyurethane Foam Solutions”.
These foams are designed to be injected into selected body
side cavities, pillars, and roof structural beams of vehicles
as they move along the assembly line. Bayfill® NV will
fill any cavity shape or contour and flow into hidden areas
where it can expand and create an acoustic seal as shown
in figures 2 - 4 . This characteristic is far superior to the
performance offered by baffles (expandable inserts).
Because of this flow and gel time characteristic, redesign
of sheet metal or injection points may not be necessary to
enhanced design flexibility. The new product, Bayfill®
NV, is thus a cost and mass efficient route to designed
acoustic barriers in vehicle bodies.

AKNOWLEDGEMENTS

The authors would like to thank Michael Blaskiewicz,

Ph.D. and Timothy Kelly, Technician Specialist for their
invaluable contributions to the development of these
Bayfill NV Foams.

Meyer-Ahrens, Milliren, Karas, Lupini / 185

 REFERENCES

1. Markel, A., “Sound Judgement”, Inside Automotive

International, May 1998,: 14-17.
2. Gettys, G. 1993, “New Concepts in Acoustical
Baffles”, Proceedings of the Interior and Exterior
Systems, International Body Engineering Conference,
p p .9 0 - 9 3 (1993).
3. Goldman, A., and Lilley, K.. 1994, “NVH
Improvement with Two-Component Polyurethane
Foam”, Proceedings of the Automotive Body Interior
& Safety Systems, International Body Engineering
Conference, pp. 23-27 (1994).
4. Visintainer, A.P., and Saha, P. 1997, SAE Paper
971931 presented at the SAE Noise and Vibration
Conference, Traverse City, MI, Conference
Proceedings pp.503 - 509, 1997.
5. Pearson, B.A. 1999, “New Low MDI Polyurethane
Foam System, for Acoustical Barrier Applications in
the Automotive Industry”, SAE Paper 19999-01-1671
presented at the SAE Noise and Vibration Conference,
Traverse City, MI, Conference Proceedings pp.150 -
162, 1999.
6. Saha, P. 1995, “The need O f A Standardized
Acoustical Evaluation Method for Body Cavity
Figure 2 . Foamed section with injection opening Fillers”, Paper presented at International Body &
Safety Systems Conference 1995, Conference
Proceedings pp. 1 6 6 - 171, 1995.

Bayfill® is a registered trademark of Bayer AG.

Figure 3. Same section as in Fig. 2 after removing upper

shell.

1 8 6 / Meyer-Ahrens, Milliren, Karas, Lupini

 Michael Lupini
Mike Lupini serves as Product
Application Manager at
Orbseal LLC. He is currently
in charge of the Process
Development Group, working
on Dispense Technologies for
the Bayer Cavity Filling NVH
Foams. He joined Orbseal in
1988 after graduating from
Pittsburg State University in
Kansas with a Plastics Engineering degree. He maintains
membership in the Society of Automotive Engineers.

Charles M. Milliren
Charles Milliren is a Senior
Development Scientist for
Bayer Corporation in
Pittsburgh, Pennsylvania. He
received his B.S. degree in
Chemistry from Ursinus
College in 1969 and his Ph.D.
in Organic Chemistry from
the University of North
Carolina at Chapel Hill in
1976.He has worked in a variety of polyurethane foam
Figure 4. Expansion o f foam into contours of cavity
systems, primarily in the Automotive Business Group and
providing excellent fit. is currently exploring low-density acoustic foam systems.
Dr. Milliren is the author of a number of papers and patents
BIOGRAPHIES on polyurethane foams.

G ary S. K aras Sven Meyer- Ahrens

Sven Meyer-Ahrens
Gary Karas is the Business obtained his Ph.D. in organic
Development Manager for chemistry from the
Orbseal LLC, a global University of Kiel,
supplier of Adhesive and Germany, in 1989. He then
Sealant Technologies. He is joined the polyurethanes
currently the Project Leader application technology
for the commercialization of department of Bayer AG
Bayer's Cavity Filling NVH where he worked on the
Foams. Mr. Karas has over development of flexible
twenty years experience in slabstock foam. In 1993 he
Automotive Market Development, working with various transferred to the automotive applications group where he
Tier One Suppliers. He studied Business Management at was responsible for the development of MDI-based
Oakland University in Rochester, Michigan, and is a flexible molded foams for acoustic and upholstery
member of the Society of Automotive Engineers. applications in the automotive industry. Since 1998 he is
leading a team working on advanced systems and
processing technologies within the Polyurethanes
Automotive Group of Bayer Corp., Pittsburgh, USA.

Meyer-Ahrens, Milliren, Karas, Lupini / 187

 SESSION F
FUNDAMENTALS

TECHNICAL SESSION ORGANIZERS

Sue Clift, Chair

BAYER CORPORATION

Ken McDaniel
BAYER CORPORATION

Alan Schrock
THE DOW CHEMICAL COMPANY

TECHNICAL SESSION MODERATOR

Ralph Priester
THE DOW CHEMICAL COMPANY

TECHNICAL SESSION VICE - MODERATOR

Sue Clift
BAYER CORPORATION
Structure-Property Relationships of Poly(urethane-urea)s with
Soft Segments Consisting of Ultra-Low Monol Content
Poly(propylene glycol) with and without Tri (propylene glycol)

MATTHEW J . O’SICKEY a n d BRUCE D. LAWREY

GARTH L. WILKES Bayer Corporation
Virginia Polytechnic Institute and State University P.O. Box 38007
Department o f Chemical Engineering South Charleston, West Virginia 25303
Blacksburg, Virginia 24061

ABSTRACT INTRODUCTION

Segmented poly(urethane-urea)s have been synthesized, Segmented, microphase separated, poly(urethane-urea)

via a two stage polymerization, with mixed soft segments elastomers are important both commercially and
of ultra-low monol content polypropylene glycol) (PPG) scientifically. A wide range of chemical formulations are
and tri(propylene glycol) (TPG) which allows the used to fabricate materials with a range of thermal and
fabrication o f quality elastomers without crosslinking. The mechanical properties, though many o f the most significant
narrow molecular weight distribution o f the ultra-low industrial formulations have until recently relied on the use
monol content PPG polyols allows for the probing of the of the poly(tetramethylene ether glycol) as a soft segment
influence o f the low molecular components of the material. Polypropylene glycol) (PPG), most commonly
molecular weight distribution through the inclusion of low produced through a base -catalyzed process, has also been
molecular homologs o f PPG such as TPG. Structure- studied as a soft segment material for foams, and with less
property relationships for these materials were investigated spectacular results for elastomers[l-7]. However, in the
as average soft segment molecular weight was varied by early 1990s, advances in the catalysts used to produce PPG
blending 8000 g/mol PPG with TPG to achieve molecular diols, resulted in substantial improvements in the level of
weights o f 1500, 2000, and 2500 g/mol. Morphological difunctionality, molecular weights, and a narrowing o f the
features such as microphase separation, interdomain molecular weight distribution[8-9] use o f PPG in
spacing and interphase thickness were quantified and poly(urethane-urea) elastomers has been generally limited
revealed with SAXS. AFM was utilized to verify the largely due to the inability to generate PPG diols with high
microphase separation characteristics inferred by SAXS. difunctionality and appreciable molecular weight. The
The thermal and mechanical behavior was assessed impact of monol on the properties of urethane elastomers
through applications of DMA, DSC, and conventional has been previously documented, but in brief,
mechanical tests. It was found that as the average soft monofunctional PPG polyofs a d as terminators during the
segment molecular weight was decreased through the development of elastomer molecular weight upon chain
addition o f TPG, the interdomain spacing increased extension with either diols (urethanes) or diamines
contrary to the trend seen for decreasing soft segment (urethane-ureas). The elastomers thus produced have
molecular weight in PPG based systems without TPG. inferior mechanical properties. PTMEG in contrast, is and
Additionally, the inclusion o f TPG in the polyurethane - has been produced as a low monol content diol, though
urea) formulations resulted in the formation of larger hard with broader molecular weight distribution and has been
domains as evidenced by AFM. These results and used with considerable success in both urethane and
supporting evidence from DMA, DSC, birefringence, and urethane-urea elastomers applications.
mechanical testing led to the conclusion that TPG may act There has been little work to assess the impact o f the low
as a chain extender as well as or in contrast to a molecular diols, added during the prepolymer synthesis, in
conventional soft segment. the PTMEG molecular distribution due in part to
difficulties in obtaining a narrow molecular distribution
PTMEG for comparison purposes. Use o f ultra-low monol
content PPG, with its narrower molecular weight
distribution, presents an opportunity to make such a study

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 191

on a similar poly(urethane-urea) system. An artificially Table 1. Sample Nomenclature and HS Contents
added low molecular weight homolog to PPG, such as PPG SS NCO Wt% Sample
tri(propylene glycol), may be blended with the PPG diol MW OH HS Nomenclature
prior to prepolymer formation, thus providing a series of 2000 1.48 6.25 {(P2K)}-2K-6.3
materials to contrast against those with no appreciable low 2000 1.71 8.98 {(P2K)}-2K-9.0
molecular contribution. 4000 1.91 6.25 {(P4K)}-4K-6.3
4000 2.34 8.98 {(P4K)}-4K-9.0
8000 2.86 6.25 {(P8K)}-8K-6.3
EXPERIM ENTAL 8000 3.71 8.98 {(P8K)}-8K-9.0
8000/2500 1.87 8.98 {(P8K/T2.5K)}-2.5K-9.0
M aterials 8000/2000 1.70 8.98 {(P8K/T2.0K)}-2.0K-9.0
8000/1500 1.55 8.98 {(P8K/T1.5K)}-1.5K-9.0
4000/2000 1.48 6.25 {(P4K/T2K)}-2.0K-6.3
Ultra-low monol content polypropylene glycol) (Mn of
4000/2000 1.71 8.98 {(P4K/T2K)}-2.0K-9.0
2000, 4000, and 8000 g/mol) and Mondur M 4,4’-MDI
were obtained from Bayer Corporation. Tri(propylene
glycol) was obtained from Lyondell, and ethylene diamine,
propylene diamine, N,N -dimethyl acetamide (DMAc) were Thermal fluctuations were addressed via the procedure of
obtained from Aldrich. Vonk and the interphase thickness parameter, a , was
obtained through a graphical approach using the methods
Synthesis of PPG-based Poly(u ethane-urea)s of Bonart as described by Tyagi[12-13].
Confirmation of trends determined by SAXS was
A more detailed description >f the synthesis of the obtained from Atomic Force Microscopy (AFM), as
elastomers has been described previously and is only conducted with a Digital Instruments Scanning Probe
briefly recounted here[10]. Mi: ures of diols are made Dimension 3000 Microscope with a Nanoscope Ilia
from Acclaim™ PPG diols and tr.(propylene glycol) prior controller. Nanosensors TESP (Tapping Etched Silicon
to end-capping with 4,4’-MDI. These mixtures were end- Probe) type single beam cantilevers were used in this
capped with an excess of 4.4’-MDI to generate instrument in tapping mode to obtain phase images o f the
prepolymers. The prepolymers were then chain extended free surfaces of the poly(urethane-urea) films cast from
with the remaining MDI and a 80:20 mixture of ethylene DMAc upon Fischer brand glass microscope slides. The
diamine and propylene diamine in DMAc. Compositions cantilevers were approximately 125 jim in length, force
and nomenclature of the samples are shown in Table 1 constants of 35 ± 7 N/m, and were operated at ca. a
where PPG soft segment MW in the form of xOOO/yyyy frequency of 290 kHz. Operation conditions were held
indicate that a PPG polyol of xOOO g/mol was mixed with constant for all samples within this study.
TPG until the number average molecular weight of the A Seiko model 210 DMA operated in tensile mode was
mixture was yyyy. Films were cast from the 22wt% solids used to obtain the dynamic mechanical response o f the
solutions upon glass plates using a doctor blade to achieve materials. Data was taken at a heating rate of 1°C / min at
final (post-drying) film thickness ranging from 0.1 to 0.3 1 Hz. Materials of cross-sectional area between 0.7 and
mm. Immediately upon casting, the coated glass plates 2.0 mm2 with length no less than 20 mm were mounted
were placed within a circulating hot air oven heated to with a grip-to-grip distance of 10 mm and initially cooled
80°C for a period no less than 2 hours. A similar to ca. -100°C with a liquid nitrogen cooling system.
procedure was used to coat thin films upon glass Fischer Thermal properties were assessed using a Seiko model 220
brand microscope slides for purposes of AFM analysis. DSC. Sample loading varied between ca. 7 mg and ca.
15mg. In all cases, samples were cooled to ca. -100°C and
Instrum entation heated to 100°C at a rate of 10°C / min, cooled again to -
100°C and heated again to 100°C at a rate of 10°C / min,
employing a nitrogen purge gas. Results o f the second
Small Angle X -ray Scattering (SAXS) was performed to
heating run were normalized on a weight basis.
ascertain the average spacing between hard domains, and a
relative measure o f the interphase thickness. A Phillips
RESULTS AND DISCUSSION
model PW1729 generator was operated at 40 kV and 20
mA in conjunction with a slit-collimated Kratky camera
(0.03 x 5 mm) with nickel filtered CuKD radiation It had been previously seen, for formulations without
(wavelength = 1.542A). A Braun OED 50 position TPG, that as the molecular weight o f the soft segments
sensitive platinum wire detector was employed with this decreased, the interdomain spacing as determined from
system. Slit-smeared data was obtained and utilized as is; SAXS also decreased (Table 2)[10]. In striking contrast to
the interdomain spacings reported are therefore somewhat these results, when the average soft segment molecular
greater than in actuality. Data normalization with a weight is decreased via blending with the low molecular
Lupolen standard and correction for parasitic scattering weight homolog, TPG, it was seen that the interdomain
was performed as described by Kratky’s procedure[l 1]. spacing increased appreciably. When a 8000 g/mol PPG
was blended to soft segment molecular weights of 2500,

192 / O ’Sickey, Wilkes, Lawrey

 Table 2. Interdomain SAXS Spacings and Interphase additional of TPG, the results were in stark contrast to that
_______________ thicknesses previously observed trend (Figure 1). In the AFM
Sample Interdomain micrographs shown, the brighter regions corresponds to
Designation spacing, Interphase areas of the sample with higher modulus (i.e. the hard
d (A ) <J(A) domains)[14]. It is clearly seen that as the average soft
{(P2K)}-2K-6.3 120 6.0 segment molecular weight is decreased from 8000 g/mol
{(P2K)}-2K-9.0 120 5.0 (Figure 1A) to 1500 g/mol (Figure 1C) through the
{(P4K)}-4K-6.3 140 5.5 addition of TPG the grain size dramatically increases. The
{(P4K)}-4K-9.0 145 5.6 increase in the amount of bright region in the micrographs
{(P8K)}-8K-6.3 160 6.1
is an indicator that there is a greater amount o f hard
{(P8K)}-8K-9.0 175-180 4.8
{(P8K/T2.5K)}-2.5K-9.0 160 6.8 segment or hard domain material. However, recalling that
{(P8K/T2.0K)}-2.0K-9.0 175 5.1 the hard segment content was kept nominally fixed at 9
{(P8K/T1.5K)}-1.5K-9.0 190 7.1 wt%, this is at first surprising until one considers the
{(P4K/T2K)}-2.0K-6.3 150-155 na possibility that a significant portion of the TPG may be
{(P4K/T2K))-2.0K-9.0 155 na acting less as a soft segment and more as a chain extending
diol. The validity o f this claim will be discussed in more
detail after other experimental results have been discussed.
2000, and 1500 g/mol it was seen that the interdomain DMA has previously revealed for PUUs, that as the
spacing varied from 160 to 190 A (Table 2). These values average molecular weight o f ultra-low monol PPG soft
of interdomain spacings are more comparable to spacings segments decreases, for systems which do not contain
seen for formulations based upon an 8000 g/mol PPG TPG, there is a rise and broadening of the soft segment
when no TPG was included (175-180A). A similar trend glass transition (Table 3)[10]. There is also a slight
was seen for formulations based upon 4000 g/mol PPG increase in the modulus o f the material above the soft
diols (Table 2)[10]. It should also be mentioned that when segment glass transition temperature. These phenomenon
TPG is included in the formulations that there exists a were explained by recognizing that a hard domain will
distinct sharpening of the peaks in the scattering curves influence the mobility of some terminal portion o f the soft
and an increase in the scattering intensity. The increase in segments it is attached to. Hence, longer soft segments
peak sharpness is an indication of a narrower distribution will possess more similar freedom to that o f the
of interdomain spacings and the increase in the intensity homopolymer PPG since they experience less restriction
may imply better phase separation of the hard and soft due to attachment to the hard domains. When the soft
domains. segment molecular weight is decreased via blending with
Additionally, an interphase thickness analysis was TPG there is again a rise in the glass transition
performed on the SAXS data using a graphical approach temperature, but not nearly to the extent that would have
based upon Bonart’s method[13]. It was seen that, for been predicted from the behavior of the soft segment
systems based upon an 8000 g/mol PPG containing TPG, without TPG (Table 3). In fact, the soft segment glass
there was an apparent increase in the thickness of the transition temperature of systems containing TPG more
interphase region of ~ 2A when compared to analogous greatly reflects the glass transition temperature o f the 8000
systems without TPG (Table 1). While the interphase g/mol PPG soft segments than the 2000 g/mol soft
thickness numbers should not taken as absolutes, they are segments. The glass transition region o f materials
valuable as indicators of trends. containing TPG are slightly broader than that o f those
Atomic force microscopy had previously shown that as based upon systems with the same starting PPG molecular
soft segment size decreased, a finer grain structure was weight, but without TPG (i.e. 8000, 4000 g/mol). There’s
seen in the AFM micrographs[10]. However, when the
average soft segment length was decreased through the
Table 3. Soft Segment Glass Transition Temperatures
Sample DMA# D SC *
Designation (°C) +/- 0.3°C (°C) +/ 0.3°C
{(P2K)}-2K-6.3 -47.9 °C -48.9 °C
{(P2K)}-2K-9.0 -48.1 °C -49.7 °C
{(P4K)}-4K-6.3 -59.2 °C -58.4 °C
{(P4K)}-4K-9.0 -58.0 °C -59.1 °C
{(P8K)}-8K-6.3 -65.9 °C 62.3 °C
{(P8K)}-8K-9.0 -63.4 °C -62.3 °C
{(P8K/T2.5K)}-2.5K-9.0 -59.0 °C -61.2 °C
Figure 1. 500 nm Phase image A FM micrographs for 8K {(P8K/T2.0K)}-2.0K-9.0 -57.6 °C 60.8 °C
PPG based PUUs with TPG added to achieve average soft {(P8K/T1,5K )H .5K-9.0 -56.0 °C 60.6 °C
segment M W of A) 8000, B) 2500, and C) 1500 g/mol. # Peaks of the tan delta curves,
White bars represent 200 nm widths. •Inflection of Ta region of DSC curves.

O 'Sickey, Wilkes, Lawrey / 193

also a slight modulus rise above the glass transition 8. Seneker, S., N. Barksby, and B. Lawrey. 1996.
temperature as TPG content increases/soft segment MW presented at SPI Polyurethanes *96, Las Vegas, NV,
decreases, which is consistent with the trends previously October 1996.
seen for decreasing soft segment molecular weight and for 9. Barksby, N., S. Seneker, and G. Allen. 1995.
increased hard segment content (Table 3). presented at the Fall PMA Meeting, Pittsburgh, PA,
The trends seen in DMA were confirmed and expanded November 1995.
upon using differential scanning calorimetry[10]. It was 10. O'Sickey, M., B. Lawrey, and G. Wilkes. 2001. J.
seen again, that for systems based upon 8000 g/mol PPG, Appl. Poly. Sci. (accepted 02/01).
blending with TPG to achieve a final SS MW of 2000 11. Kratky O., I. Pilz, and P. Schmitz. 1966. J. Coll. Int.
g/mol (Table 3), resulted in a broadening of the glass Sci. 21:24.
transition and increase to higher temperatures though not 12. Tyagi D., J. McGrath, and G. Wilkes. 1986. Poly.
nearly as great as would have been predicted by the Engr. Sci. 26:1371.
influence of soft segment molecular weight alone. 13. Bonart R., L. Morbitzer, and E. Muller. 1974. J.
Macromol. Sci-Phys. B9:447.
CONCLUSIONS 14. Sauer B. 1997. Macromolecules^ 30(26):8314.

It has been suggested that a significant portion of TPG is BIOGRAPHIES

incorporated into the hard domains as a diol chain M atthew J. O ’Sickey
extender, rather than in the soft domains as a short soft Matt O ’Sickey is pursuing a
segment. Atomic force microscopy clearly showed an Ph.D. degree in Chemical
increase in the hard domain content and size with Engineering at Virginia
increasing TPG content. Dynamic mechanical analysis and Polytechnic Institute and
DSC showed that the TPG containing systems tended to State University working in
mimic the behavior of those with shorter soft segments the area of polymer
and/or greater hard segment content. Undoubtedly is it processing-structure-property
some combination of both effects. Dynamic mechanical relationships of polyurethane-
analysis and DSC showed that the soft segment glass urea elastomers. He received
transition temperature behaved more like the unblended his B.S. in Chemical Engineering from Purdue University
PPG diol, implying that the behavior of the 8000 and 4000 where he spent two years investigating polymeric materials
g/mol PPG soft segments is unchanged. This may imply for non-linear optical device applications.
that though a considerable portion of the TPG acts as a Bruce D. Lawrey
chain extender in the hard domains, the long narrow
Bruce Lawrey is a Senior Scientist in the Step-outs and Specialties
distribution PPG soft segments still behave as flexible coils
between the hard domains. Thus if the soft segment Division of Bayer Corporation’s Global Polyurethanes Research
continue to separate the hard domains by essentially the Department. He is currently assigned to their South Charleston,
same amount, and the hard domains are clearly larger, then West Virginia Technical Center and is responsible for R&D
the interdomain spacing as determined by SAXS would by programs related to development of new applications for ultra-
necessity have to be greater. low monol polyols. He received his B.S. Degree in Chemical
Engineering from the University of Pennsylvania. He received
ACKNOWLEDGMENTS his Ph.D. in Chemical Engineering from Princeton University.
The authors o f this paper would like to thank Bayer G arth L. Wilkes
Corporation for the financial support and materials making Garth L. Wilkes is co-director
this research possible. of the Polymer Materials and
Interfaces Laboratory at
REFERENCES Virginia Polytechnic Institute
1. Petrovic, Z. and I. Soda-So. 1989. J. Polym. Sci. (B) and State University where
Phys. 27:545. he is also a university
2. Pechhold, E. and G. Pruckmayr. 1981. Rubber Chem. distinguished professor at the
Tech. 55:76. Department of Chemical
3. Cooper, S. and A. Tobolsky. 1966. J. Appl. Poly. Sci., Engineering. His research
10:1837. interests are focused
4. Dong, S. and S. Cooper. 1971. Polym. Engr. Sci. primarily on the processing-
11:369. structure-property behavior of polymeric materials with an
5. Ishihara, H. 1983. J. Macromol. Sci-Phys. B22:763. accent on a molecular and experimental approach. Over the
6. Sung, C.S., T. Smith, and N. Sung. 1980. thirty odd years that his research personnel have been
Macromolecules 13:117. active, an extremely wide range of materials and types of
7. Wilkes, G., and S. Abouzahr. 1981. Macromolecules problems have been considered. Topics have included
14:456. many aspects of crystalline, rubbery and glassy materials.

1 9 4 / 0 *Sickey, Wilkes, Lawrey

Exploring Urea Phase Connectivity in Flexible Polyurethane
Foams Using Lithium Chloride as a Probe

ASHISH ANEJA a n d GARTH L. WILKES

Virginia Polytechnic Institute and State University
Dept, o f Chemical Engineering
Polymer Materials and Interfaces Laboratory
Blacksburg, Virginia 24061

ABSTRACT alters the reaction kinetics as well as the solubilization

characteristics of the different foaming components, and
Lithium chloride was used as an additive to thus leads to systems possessing varying levels o f urea
systematically alter the phase separation characteristics phase connectivity. In the present study, the addition o f LiCl
and was hence used as a probe to examine the urea phase to foam formulations in a systematic manner has been used
connectivity in a series o f formulations based on flexible to probe the different levels of urea phase connectivity.
slabstock polyurethane foams. Several complimentary It needs to be pointed out that the present study makes use
characterization techniques were employed to of ‘plaques’ based on slabstock foam formulations. The
characterize urea phase connectivity which was observed ‘plaque’ formulations differ from actual ‘foam’ recipes in
to be present at different length scales. SAXS, TEM, and that they do not incorporate a surfactant, since addition o f a
AFM were collectively used to study the size, presence, surfactant is not required in the formation of a plaque (as
and dispersion o f the urea aggregates commonly observed there is no desire to stabilize the cellular structure).
in slabstock foams, thus giving insight into the macro- Kaushiva et. al. in a recent study on slabstock flexible
connectivity o f the urea phase. WAXS was used to assess polyurethane foams observed that the presence o f a
the regularity in packing of the hard segments constituting surfactant does have an influence on the solid state
the urea microdomains, or the urea phase micro- morphology of the polymer [4]. However, in the past,
connectivity. DSC and DMA indicated that the addition workers have noted that the trends observed in the
of LiCl led to dispersing the urea microdomains in a morphological features o f plaques strongly resemble those
manner which restricted the soft segment mobility as o f actual foams with similar compositions, which
observed by a broadening in the soft segment glass rationalizes the present investigation o f ‘plaques’ as a
transition. means of indirectly studying foam morphology [5, 6]. Thus,
the key results obtained from this study are expected to be
INTRODUCTION similar for comparable flexible foam materials containing
surfactant - also an area of investigation of the authors’
The physical properties of flexible polyurethane foams present research.
are greatly influenced by the phase-separated morphology
which arises due to an incompatibility between urea hard EXPERIMENTAL
segments and polyol soft segments [1]. The urea hard
segments are known to form partially ordered hydrogen- M aterials
bonded microdomains which in turn aggregate at
relatively higher urea contents, resulting in urea rich Three series of plaques based on slabstock foam
regions termed as urea balls/aggregates [1]. The presence formulations, containing 2.0, 3.5, and 4.5 water pphp, were
of certain additives in foam formulations has been shown prepared at Dow Chemical in Freeport, Texas. For each
to disrupt the hydrogen-bonding characteristics of these series prepared the LiCl content was varied as 0.0, 0.1, 0.5,
materials, often resulting in an alteration of the foam and 1.0 LiCl pphp. The polyol utilized was Voranol 3137
morphology at the microdomain as well as aggregate level which is a random EO/PO heterofed polyol (13% EO) with
[2, 3]. One such additive is lithium chloride, which does a molecular weight of ca. 3000 and an average functionality
not react into the covalent network of the polymer, but of 2.79. The isocyanate employed was an 80:20 mixture of

,
POLYURETHANES EXPO 2001 SEPTEMBER 30 - OCTOBER 3, 2001 195
2,4/2,6 TDI and was used at an index value of 100. The Phillips 420T scanning transmission electron microscope
catalyst was added at a level of 0.2 pphp. No surfactant operating at an accelerating voltage o f 100 kV. Tapping
was utilized. The quality of network formation for mode atomic force microscopy (AFM) was carried out on
materials with and without LiCl was found to be similar the cryogenically microtomed surfaces of the samples.
by carrying out swelling and sol% analyses. Since a wide Phase images were collected using a Digital Instruments
range o f materials with varying water and LiCl content scanning probe microscope employing a Nanoscope Ilia
were prepared and analyzed, only a portion of the results controller and Nanosensors TESP (Tapping Etched Silicon
which are representative of this study are presented in this Probe) type single beam cantilevers. The cantilevers had a
paper. nominal length of 125 pm, with force constants in the range
The preparation of the plaques was carried out in a lab of 35 ± 7 N/m, and were used at oscillation frequencies of
scale cup-foaming setup at Dow Chemical in Freeport, ca. 295 kHz. The samples which were cryo-sectioned
Texas. Water (with a known amount of LiCl dissolved in smooth for TEM were examined by AFM. The ‘height’ as
it), and the polyol, were added to a cup and the mixture well as the ‘phase’ images were collected, but since the
was stirred for 25 seconds at 2000 rpm with a 1” diameter samples were microtomed smooth, the height images were
stirrer. The TDI and the catalyst were then added and the disregarded.
mixture was stirred for another 15 seconds. The foam Dynamic mechanical analysis (DMA) was carried out in
formation was suppressed by forcing the foam to collapse the tensile mode using a Seiko model 210. Samples with
by stirring. Just prior to gelation, the reacting mixture was dimensions of approximately 15 x 4.5 x 1.2 mm3 were
quickly poured on Teflon sheets supported by steel plates heated from - 120 °C to 250°C at a rate of 2.0 °C/min from
and placed in a hot-press operating at 100 °C and 20,000 which the storage modulus and tan8 were collected at a
lbf for 1 hour. A picture-frame mold ca. 0.05” thick was frequency of 1 Hz. The grip-to-grip distance was set at 10
utilized. At the end of 1 hour, the plaque was cut out from mm.
the picture-frame and allowed to cool at ambient Differential scanning calorimetery (DSC) experiments
conditions. were conducted using a Seiko DSC 220C under a nitrogen
purge and at a heating rate o f 10 °C/min. The DSC curves
Instrum entation were normalized to a 1 mg sample mass. DSC was carried
out to observe any changes in the soft segment glass
Small angle x -ray scattering (SAXS) experiments were transition position and breadth on addition of LiCl.
carried out using a slit collimated (0.03 x 5 mm2) Kratky
camera with nickel filtered CuKa radiation having a RESULTS AND DISCUSSION
wavelength o f 1.54 A. A Philips model PW1729
generator operating at 40 kV and 20 mA was utilized. The SAXS profiles for the 4.5 water pphp series with varying
detector used was a Braun OED 50 position-sensitive LiCl content are displayed in Figure 1. It is noted that the
platinum wire detector. The raw data was corrected for addition of LiCl leads to a drop in the scattering intensity at
parasitic scattering and normalized using a Lupolen lower values of the scattering vector ‘s’. This observation is
standard. noted to be most pronounced for the highest LiCl containing
The local ordering of the hard segments at the 1-10 A plaque. Since larger structures are known to contribute to
level was investigated using wide angle x-ray scattering the scattering intensity at lower angles, it is believed that
(WAXS). A Warhus camera with pinhole collimated (ca. addition of LiCl suppresses the formation of the urea
0.02 in. diameter) nickel filtered CuKa radiation with a
wavelength o f 1.54 A was used. Samples having a
thickness of ca. 1.2 mm were exposed to x-rays for 5
hours, with a sample to film distance of 5.5 cm.
Transmission electron microscopy (TEM) was used to
examine the urea particulates and their level of
aggregation. The urea phase has a higher electron density
as compared to the softer polyol phase and hence shows
up as darker regions on the micrographs. Small samples
were cut from the plaques and embedded in epoxy which
was allowed to cure overnight at ambient conditions. The
samples retained their original dimensions (no swelling)
even after the surrounding epoxy had cured, suggesting
that there was no chemical interaction between the epoxy
and the samples. A diamond knife was used to
cryogenically microtome the samples at ca. - 90 °C on a
Reichert-Jung ultramicrotome. Sections with thickness of
ca. 80 nm were collected on 600 mesh copper grids using Figure 1. SAXS profiles for the 4.5 water pphp series
ethanol as the solvent. Micrographs were taken using a with varying LiCl content

1 9 6 / Aneja, Wilkes
aggregates. This observation is also supported by TEM
and AFM micrographs (discussed later). At an ‘s’ value of
ca. 0.01 A’1, a first order interference in the form of a
shoulder was observed for all formulations. This indicates
that all the plaques displayed microphase-separation with
an interdomain spacing of ca. 100 A, in conformity with
earlier results obtained for slabstock foam formulations
[1, 5]. This first order interference, which appeared as a
weak shoulder in the plaque containing no LiCl
transformed into slightly sharper shoulders as the LiCl
content was increased. This observation can be explained
as follows. In the plaque with no LiCl, since the urea
microdomains reside both in the urea rich aggregates as
well as in the polyol phase, the distribution of the
interdomain spacing would be wider and hence the first
order interference appears as a weak shoulder. On the Figure 3. Effect of LiCl addition obsen/ed via TEM for
other hand, in formulations containing LiCl, since the the 3.5 water pphp series a) 0.0 LiCl pphp b) 0.1 LiCl
urea rich aggregates are suppressed, this leads to a more pphp c) 0.5 LiCl pphp d) 1.0 LiCl pphp
uniform distribution of the urea microdomains which
‘sharpens’ the SAXS shoulder. Similar results were
obtained for the 2.0 and 3.5 water pphp series, and are aggregation for this formulation is observed to be irregular
thus not presented here. in nature, and aggregates are noted to be smaller than 0.4
TEM images for the 2.0, 3.5, and 4.5 water pphp pm in size. At the next higher level of LiCl addition, i.e. at
plaques with no LiCl are presented in Figure 2. The urea 0.5 LiCl pphp, the urea phase shows no aggregation
phase has a higher electron density as compared to the behavior, as seen in Figure 3c. This systematic decrease in
polyol phase and hence shows up as darker regions on the the urea aggregation, which was also suggested earlier by
micrographs. It can be observed that as the water (or urea SAXS profiles, clearly suggests a reduction in the urea
phase) content is increased, the urea aggregates tend to phase macro-connectivity - aggregates which appeared to
increase in both number, and size. These three systems have physical associations at 0.0 LiCl pphp are observed to
would be expected to possess varying levels of urea be comparatively isolated from each other as LiCl is
aggregate connectivity, or macro-connectivity. The introduced in the formulations. These observed changes in
addition of LiCl to each of these three plaques would then the urea aggregation behavior were also supported by noting
give insight into the different levels of urea phase that the plaque without LiCl was opaque, and the plaques
connectivity which might be present while maintaining a became clearer or more transparent as the LiCl content was
constant urea phase content. The effect o f addition of LiCl increased. Finally, at the highest dose of LiCl, i.e. at 1.0
observed via TEM for the 3.5 water pphp plaque is shown LiCl pphp, no aggregation of the urea phase is observed, as
in Figure 3. The image corresponding to the plaque with seen in Figure 3d. The TEM images for the other two series
no LiCl displays urea precipitates of the order of 0.2 - 0.6 examined, i.e. the 2.0, and 4.5 water pphp series, led to
pm in size. It can be observed that the addition of LiCl, similar conclusions with regard to the macro-connectivity of
even at a small amount o f 0.1 LiCl pphp, leads to a the urea phase, and are thus not presented here.
noticeable reduction in urea aggregation. The urea In recent work by Kaushiva et. al., AFM was
demonstrated as a useful technique to study flexible
polyurethane foam morphology [4]. In phase images
obtained via AFM, a higher modulus material typically
induces a higher phase offset and appears lighter as opposed
to a softer phase which appears darker. Thus for the
polyurethanes imaged in this investigation, the urea rich
regions appear lighter where as the darker regions
correspond to the polyol phase. The AFM images obtained
for the 3.5 water pphp series with varying LiCl content are
presented in Figure 4. These images strongly reinforce the
findings obtained so far from the SAXS and TEM results,
and also give insight into the phase-separation behavior at
the relatively higher LiCl concentrations. It is observed that
as the LiCl content is increased (Figures 4a to 4d), there is a
Figure 2. TEM images of slabstock plaques as a function clear reduction in the urea phase macro-connectivity. It can
of water content: a) 2.0 water pphp, b) 3.5 water pphp c) also be observed from Figures 4a and 4b that die urea
4-5 water pphp aggregates are not purely based on urea but have some

Aneja, Wilkes / 197

 Figure 6. Soft segment glass transition obsen/ed
Figure 4. Effect o f LiC! addition observed via AFM for
via DSC as a function of L id content
the 3.5 water pphp series a) 0.0 L id pphp b) 0.1 L id
pphp c) 0.5 L id pphp d) 1.0 L id pphp

thus suggest that the addition of LiCl not only leads to a

polyol residing in them [7]. At relatively higher LiCl reduction in order at the urea aggregate level, but also alters
concentrations (Figures 4c and 4d), noticeable differences the /wcroconnectivity or the regularity in packing order
in the phase-separation behavior can be observed via between the hard segments themselves.
AFM. A decrease in the size of the urea rich as well as DMA and DSC were used to ascertain the effect of LiCl
polyol rich phases is observed as the LiCl content addition on the breadth and position of the soft segment
increases from 0.5 to 1.0 LiCl pphp. The addition of LiCl, glass transition, Tg. Figure 6 presents the DSC scans for the
therefore, results in the distribution of the urea series with 4.5 water pphp series. A slight but systematic
microdomains in a more homogeneous manner in the soft broadening of the soft segment Tg was observed on LiCl
polyol phase. addition. The relative position of the soft segment Tg,
Previous studies have shown that the hard segments however, was found to be nearly unchanged. The observed
which constitute the urea microdomains possess ordering small Tg broadening can be possibly due to the urea phase
of a somewhat paracrystalline nature [5]. This ordering is imposing greater restraints on the soft segment mobility on
detected using WAXS in the form of a weak but distinct increasing the LiCl content. Further support o f this
4.7 A diffraction peak. WAXS patterns for the 4.5 water observation was obtained via DMA where a broadening o f
pphp plaques with 0.0 and 1.0 LiCl pphp are presented in the Tan8 peak was observed on addition of LiCl (data not
Figure 5. The characteristic 4.7 A reflection is observed to
presented).
be present in the plaque without LiCl where as it is found
to be absent in the plaque containing 1.0 LiCl pphp.
C O N C LU S IO N S
Plaques with intermediate LiCl contents show a
systematic loss in the peak intensity as the LiCl content is This study has utilized LiCl as an additive to alter and thus
increased and are thus not presented here. The patterns gain insight into urea phase connectivity in formulations
based on slabstock flexible polyurethane foams. Several
techniques such as SAXS, WAXS, TEM, AFM, DSC, and
DMA were utilized to characterize urea phase connectivity
which is found to occur over different length scales. SAXS
was used to show that the addition of LiCl did not prevent
the plaques from undergoing microphase separation. SAXS,
TEM, and AFM were collectively used to probe the
presence, size, shape, and physical associations o f the urea
rich aggregates, thus giving insight into the macro level
connectivity of the urea phase. WAXS was used to
demonstrate that the regularity in the hard segment packing,
or the micro-connectivity of the urea phase, was found to be
disrupted on addition of LiCl. The soft segment Tg was
Figure 5. WAXS patterns for plaques containing 4.5 found to be slightly broadened on addition of LiCl,
water pphp a) 0.0 L id pphp b) 1.0 L id pphp suggesting that as the urea microdomains became more

1 9 8 / Aneja, Wilkes
uniformly distributed in the polyol matrix, they imposed BIOGRAPHIES
some further restrictions on the soft segment mobility.
Ashish Aneja
ACKNOWLEDGEMENTS
Ashish Aneja received his B.S.
The authors would like to acknowledge The Dow in Chemical Engineering from
Chemical Company for financial support of this research. the Indian Institute of
Technology, Delhi. In 1998,
REFERENCES he joined the research group of
Professor G. L. Wilkes at the
1. Herrington, R. and K. Hock. 1998. Flexible Chemical Engineering
Polyurethane Foams. 2nd ed., Dow Chemical Co., Department at Virginia
Midland, MI. Polytechnic Institute and State
2. Kaushiva, B. D. and G. L. Wilkes. 2000. “Influence University. He is presently
o f Diethanolamine (DEOA) on Structure-Property working as a doctoral student
Behavior o f Molded Flexible Polyurethane Foams,” in the area of structure-
J. Appl. Polym. Sci., 77:202-216. property relationships of polymers.
3. Moreland, J. C., G. L. Wilkes, R. B. Turner, and E.
G. Rightor. 1994. “Characterization of Flexible G arth L. Wilkes
Slabstock Foams Containing Lithium Chloride,” J.
Appl Polym. ScL, 52:1459-1476. Garth L. Wilkes is co-director
4. Kaushiva, B. D., S. R. McCartney, G. R. Rossmy, of the Polymer Materials and
and G. L. Wilkes. 2000. “Surfactant Level Influences Interfaces Laboratory at
on Structure and Properties of Flexible Slabstock Virginia Polytechnic Institute
Polyurethane Foams,” Polymer, 41:285-310. and State University where he
5. Armistead, J. P., G. L. Wilkes and R. B. Turner. is also a university
1988. “Morphology of Water-Blown Flexible distinguished professor at the
Polyurethane Foams,” J. Appl Polym. Sci., 35:601- Department of Chemical
629. Engineering. His research
6. Moreland, J. C., G. L. Wilkes, and R. B. Turner. interests are focused primarily
1994. “Viscoelastic Behavior o f Flexible Slabstock on the processing-structure-
Polyurethane Foams: Dependence on Temperature property behavior of polymeric materials with an accent on
and Relative Humidity. I. Tensile and Compression a molecular and experimental approach. Over the thirty odd
Stress (Load) Relaxation,” J. Appl. Polym. Sci., 52: years that his research personnel have been active, an
549-568. extremely wide range of materials and types o f problems
7. Ade, H., A. P. Smith, S. Cameron, R. Cieslinski, G. have been considered. Topics have included many aspects
Mitchell, B. Hsiao, and E. Rightor. 1995. “X-Ray of crystalline, rubbery and glassy materials.
Microscopy in Polymer Science: Prospects of a
‘New’ Imaging Technique,” Polymer, 36:1843-1848.

Aneja, Wilkes / 199

 Fundamental Investigation of the Factors That Influence the
Properties of Slow Recovery Polyurethane Foams

SUJATHA NARAYAN a n d AMIE BERUBE

Rogers Corporation
One, Technology Drive
Rogers, Connecticut 06263

ABSTRACT It is well known that the location and shape o f the glass
transition plays a critical role in the recovery behavior of a
PU foam [1,2]. Slow recovery foams typically have a
Microcellular slow recovery polyurethane foams (the
broad glass transition around room temperature. Changes
PORON®* brand) from Rogers Corporation, were studied
in environmental parameters during use, such as,
in the context o f use as insole cushions in shoes. The effect
temperature and humidity can therefore have a strong
of environmental factors inside a shoe, such as
effect on properties. This paper attempts to elucidate the
temperature, humidity, compression strain rate and
effect that environmental parameters encountered in a shoe
compression time, on properties such as stiffness, impact
have on critical performance properties in a foam insole
attenuation and recovery profile were evaluated. Two slow
and how these relate to structural changes in the polymer.
recovery grades with low and high modulus were
compared with their analogous high resiliency grades of EXPERIMENTAL
equivalent density and thickness. Dynamic mechanical
thermal analysis and FTIR spectroscopy were used to PORON® microcellular polyurethane foams o f 0.125"
correlate the effect of environmental parameters on the thickness and 20 pounds per cubic feet (pcf) density were
ultimate properties, morphology and thermal transitions of studied. Table 1 shows the sample identification used.
these foams.
Table 1: Sample Identification for this Study
INTRODUCTION
ID USED RECOVERY MODULUS
Slow recovery, or visco-elastic, foams have gained RATE
popularity in the recent years due to their conformability, SR -LM SLOW LOW
softness, shock absorption, vibration dampening, sound HR -LM FAST LOW
SR - HM SLOW HIGH
dampening and other favorable properties. These foams are
HR -HM FAST HIGH
typically used in applications such as mattresses, pillows,
SR = Slow Recovery; HR = High Resiliency
sports and medical equipment, ergonomic cushions, etc.
LM = Low Modulus; HM = High Modulus
Their use is growing significantly in the footwear industry
as well.
Standard sample preparation involved conditioning the
Poron® microcellular polyurethane (PU) foams are used
foams at 25°C and 50% relative humidity (RH) for 24
extensively in the footwear markets due to their excellent
hours. To evaluate the effect of temperature samples were
cushioning properties and range of available moduli along
with excellent compression set resistance and impact conditioned in ovens at 4°C and 50°C for 24 hours at 50%
attenuation. A slow recovery series of the product family is RH, transferred to insulated containers and tested
gaining popularity in the footcare market through the immediately. For the effect of humidity samples were
combined advantage of low compression set and low conditioned in dessicators at 0%, 50% and 100% RH at
energy return. 25 °C until just before testing. These temperature and
humidity levels were picked based on the conditions that
* Poron is a licensed trademark of Rogers Corporation the foam can be subjected to during use in a shoe.

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 201

 Dynamic mechanical analysis was conducted on a TA contact with the disc. When the air cylinder is released, the
Instruments DMA 2980 Dynamic Mechanical Analyzer LVDT or optical encoder stays with the foam, as it
using a heating rate of 3 °C/minute and 1Hz test frequency. recovers and records the position of the surface of the
Loss modulus, storage modulus and tan delta were foam. With appropriate calibration, this data is shown as
measured. Samples conditioned at 0%, 50% and 100% percent of original thickness as a function of time. The
relative humidity were also tested. 90% recovery time is used as a measure of the recovery
FTIR spectroscopy study was conducted on a Nic -Plan rate of the foams.
Infrared microscope with an ATR crystal attachment, 128
scans were run to obtain data for each spectrum. RESULTS AND DISCUSSION, -
The performance properties tested were compression
force deflection (CFD) curves for both loading and Figure 2 shows the recovery profile of all four foams
unloading, impact attenuation and recovery rate. The CFD after compression to 50% for 15 seconds. Both the high
of the foam determines the hardness or stiffness that the resiliency foams have 90% recovery time of less than 0.15
end user feels while wearing the foam insole. seconds, whereas, the SR-LM foam takes longer at 0.9
An Instron was used to measure CFD on 2" x 2" samples seconds and SR-HM even longer at 5 seconds.
stacked to 0.375" thickness. A strain rate of 0.2
inches/minute was used to compress the sample to 75% of
its original thickness, followed by unloading back to the
undeformed state. CFD hysterisis curves were then plotted.
To study the effect of compressive strain rate on stiffness,
0.02, 0.2 and 2 inches/minute strain rates were used.
Impact attenuation was evaluated according to a
modified version of ASTM D1596. Peak g ’s were
measured by dropping a five pound weight from 2", 4" and
6" heights on to 3" disc of the sample. The effect of strain
rate was captured from the varying drop height data.
Recovery rate was tested using equipment designed by
Rogers Corporation (patent pending) to test slow recovery
Poron® foams. Figure 1 shows a diagram of the recovery
tester. A 1" disc o f the foam sample is placed in the sample
retainer area (1) with the appropriate gauge blocks
Figure 2: Recovery Profile of SR-LM, HR-LM, SR-HM and
underneath it. A light aluminum disc (for load spreading) is
HR - HM Foams Measured on Recovery Tester
placed on top of the foam sample. The extent of
compression and the time of compression can be varied.
Settings used for this study were 50% compression and 15- Dynamic Mechanical Analysis (DMA)
second compression time. When the test is started, the air
cylinder (2) and the compression fixture (3) press down on Figures 3 & 4 show tan delta as a function of
the disc load spreader and foam at a very high compressive temperature for the low modulus (LM) and high modulus
strain rate, bringing the LVDT (4) or optical en c o d er^ ) in (HM) materials respectively. The LM materials in figure 2,
due to the low hard segment content, exhibit only one glass
transition corresponding to the soft segment. The SR-LM
has a higher Tg (5°C) then HR-LM (-16°C), thus
exhibiting a slow recovery response at room temperature,
as seen in Figure 2. For the high modulus materials (Figure
4), there appears to be a separate hard segment glass
transition temperature along with the soft segment
transition. This indicates a phase separated morphology
due to the higher hard segment content. The high modulus
foams - SR-HM and HR-HM materials differ significantly.
The very broad soft segment transition that occurs around
room temperature blends into the hard segment transition
(40 °C). This explains the slow recovery response o f this
material (Figure 2). The HR -HM however, exhibits two
distinct transitions at 0°C and 69°C. It is apparent that both
the high resiliency foams HR-LM and HR-HM are in their
rubbery state at room temperature and hence, highly
Figure 1: Schematic Diagram of Slow Recovery Tester resilient.

202 / Narayan, Berube

 Table2: Effect of Temperature, Humidity and Compression
Strain Rate on 25% CFD , CFD Hysterisis Index and
Recovery Time of SR-LM and HR-LM Foams

EFFECT OF TEMPERATURE
MATERIAL TEMP. 25% HYST. 90%
CFD INDEX REC. TIME
<°C) (psi) (sec)
SR -LM 4 1.5 0.53 2.5
25 1.0 0.43 0.9
50 1.5 0.18 0.9

HR -LM 4 3.3 0.21 0

25 3.1 0.21 0.1
Figure 3: Tan Delta Measured through DMA for SR - LM 50 3.6 0.17 0.1
and HR - LM foams (3 °C/min/1Hz)
EFFECT OF HUMIDITY
MATERIAL RH 25% HYST. 90%
CFD INDEX REC. TIME
% (psi) (sec)
SR -LM 0 1.6 0.40 1.4
50 1.0 0.43 0.9
100 1.1 0.30 0.4

HR -LM 0 3.3 0.21 0.1

50 3.1 0.20 0.1
100 3.5 0.18 0.1
EFFECT OF STRAIN RATE
MATERIAL COMP. 25% HYST.
RATE CFD INDEX
("/min) (psi)
SR -LM 0.02 1.0 0.24
Figure 4: Tan Delta Measured through DMA of SR - HM
and H R - H M foams (3 °C/min/1Hz) 0.2 1.0 0.43
2 1.5 0.63

Compression Force Deflection (CFD) Study 0.02

HR -LM 3.1 0.18
0.2 3.1 0.20
Table 2 shows the effect of temperature, humidity and
strain rate on the 25% CFD, CFD hysterisis index and 90% 2 3.5 0.25
recovery time. Only 25% CFD will be discussed in this EFFECT OF COMPRESSION TIME
section. The CFD hysterisis and recovery time will be MATERIAL COMP. 90%
discussed in subsequent sections. These effects are TIME REC. TIME
important because changes in the hardness of an insole can (SEC) (sec)
affect the comfort of a shoe. In the range tested, increasing
SR -LM 1 0.4
temperature has little effect on hardness of both low
3 0.6
modulus foams - SR-LM and HR-LM; increase in relative
humidity appears to soften the SR materials a little, with 15 0.9
little effect on HR -LM; and increase in strain rate appears
to harden both materials a little, more so in the SR-LM. HR -LM 1 0.1
These observations correlate well with the fact that the
3 0.1
storage modulus, which should represent the modulus of
the polymer (Figure 5), shows very little change in that 15 0.1
temperature range.

Naroyan, Berube / 203

 Table 3: Effect o f Temperature, Humidity and Compression
Strain Rate on 25% CFD , CFD Hysterisis Index and
Recovery Time of SR-HM and H R -H M Foams

EFFECT OF TEMPERATURE
MATERIAL TEMP. 25% HYST. 90%
CFD INDEX REC. TIME
(° C) (psi) (sec)
SR -HM 4 29.1 0.80 8.5
25 8.1 0.64 4.9
50 4.1 0.21 7.2
HR -HM 4 14.8 0.50 0.2
25 11.9 0.40 0.1
50 6.9 0.37 0.3
Figure 5: Storage Modulus Measured through DMA for EFFECT OF HUMIDITY
SR - LM and HR - LM foams (3 °C/min/1Hz)
MATERIAL RH 25% HYST. 90%
CFD INDEX REC. TIME
% (psi) (sec)
Table 3 shows the 25% CFD of the high modulus foams
- SR-HM and HR-HM and the effect of temperature, SR -HM 0 15.0 0.76 10.5
humidity and strain rate on the 25% CFD. It is apparent 50 8.1 0.63 4.9
that all three factors have a significant effect on both 100 4.3 0.35 0.6
foams, the effect being especially dramatic for the former. HR -HM 0 11.7 0.51 0.1
The sharp drop in CFD between 4°C and 25°C for SR-HM
50 11.9 0.40 0.1
foam correlates directly with the sharp drop in storage
modulus (Figure 6) in that temperature range. 100 9.0 0.43 0.1
EFFECT OF STRAIN RATE
MATERIAL COMP. 25% HYST.
RATE CFD INDEX
("/min) (psi)
SR -HM 0.02 5.6 0.51
0.2 8.1 0.64
2 16.1 0.80
HR -HM 0.02 9.4 0.41
0.2 11.9 0.40
2 13.7 0.47
EFFECT OF COMPRESSION TIME
MATERIAL COMP. 90%
TIME REC. TIME
(SEC) (sec)
SR - HM 1 1.1
Figure 6: Storage Modulus Measured through DMA for 3 2.1
SR -HM and H R - HM foams (3 °C/min/1Hz)
15 4.9
HR -HM 1 0.1
3 0.1
15 0.1

The fact that the high modulus materials are affected to a

much greater extent by relative humidity may be related to
a distinct hard domain. It is well known [3] that the hard
domain structure is characterized by extensive hydrogen
bonding between urethane groups of adjacent molecules in
the hard segment. When such a urethane is subjected to

204 / Narayan, Berube

high humidity, the moisture tends to plasticize the
hydrogen -bonded areas. In phase separated morphology,
the ultimate hardness of the foam is predominantly affected
by the intrinsic hardness of the hard domains. Hence,
plasticization of the hard domains can dramatically drop
the CFD. This is apparent from figure 7, which shows the
storage modulus of SR-HM foams conditioned at 0, 50 and
100% relative humidity. The storage modulus of the 100%
RH materials is significantly lower than 0 and 50% RH.
The existence o f hydrogen bonding can be observed by
FTIR spectroscopy. The N -H group participating in the
hydrogen bond typically displays a characteristic
absorption band around 3300 cm '1 [4]. The band of the
non-bonded N-H group appears at about 3446 cm'1 and
hydrogen bonded N -H groups appear at 3327 cm'1.
Similarly, bonded C = 0 groups appear at 1707 cm '1; non- Figure 7: Storage Modulus Measured through DMA for
bonded groups at 1731 cm"1. Figure 7, 8, 9 and 10 show SR - HM and HR - H M foams (3 ° C/min/1Hz) after conditioning
FTIR spectra of all four foams after conditioning at 0, 50 at 0 , 50 and 100% Relative Humidity
and 100 % RH. In all foams, subtle changes are seen as the
aging humidity increases, but the most dramatic changes
are seen in the N-H stretch region of - 3310 cm'1. The N-H
stretch in all the foams are closer to 3327 cm'1 than 3340
cm -1 suggesting that they are all hydrogen bonded. As the
humidity increases, this peak changes from a relatively
sharp, distinct peak to a broad, indistinguishable one. This
suggests clearly that the moisture from the humidity aging
has affected the N -H group region, which is the hard
segment. This effect is most obvious in the slow recovery
foams, which correlates well with the change in properties
with relative humidity.
Tables 2 and 3 also show the effect of strain rate on the
modulus of the foams. Both the HR foams are relatively
unaffected by strain rate due to the fact that they are in the
rubbery state at room temperature. The SR-HM foam is Figure 8: FTIR spectra of SR - LM foams conditioned to 0,
50 and 100% Relative Humidity
affected very strongly by the rate of compression and are
much firmer at higher compressive strain rates. This is
related to the dynamic nature of these foams (as seen in
Figures 2 and 6) wherein the structure and morphology of
the foam is changing or can be changed significantly
around room temperature by small changes in temperature
and other parameters. Using the concept of time-
temperature superposition, it can be inferred that testing at
higher frequencies is akin to testing at lower temperatures,
hence, the foams are harder.

Figure 9: FTIR spectra of HR - LM foams conditioned to 0,

50 and 100% Relative Humidity

Narayan, Berube / 205

Figure 10: FTIR spectra of SR - HM foams conditioned to 0,
50 and 100% Relative Humidity

Figure 12: CFD hysterisis curves o f SR-LM and HR-LM

foams

Figure 11: FTIR spectra of H R - H M foams conditioned to 0,

50 and 100% Relative Humidity

Compression Force Deflection (CFD) Hysterisis Study

Figure 13: CFD hysterisis curves of SR-HM and HR-H M
Figures 12 & 13 show CFD curves and the hysterisis foams
associated with loading and unloading of low and high
modulus foams respectively. This is an important
characteristic o f the foam in an insole since walking or Table 2 and 3 also summarize the hysterisis data for the
running causes a cyclic loading and unloading compression low and high modulus foams respectively. The “Hysterisis
on the insoles. The hysterisis measures the energy that is Index” of each foam is calculated from the area enclosed
returned from the foam during these loading and unloading by the loading and unloading curves.
cycles. It is clearly apparent that the hysterisis for the SR In all the foams, the hysterisis is reduced with increasing
foams at both moduli are higher than their corresponding temperature suggesting that the material gets more resilient
HR foams. This is expected for the slow recovery foams with temperature. The effect of temperature is quite
because lower rebound forces would be expected on dramatic on both the SR foams, where they become quite
decompression. Hence, the hysterisis data can be used as a resilient around 50°C. This may be an advantage in some
measure of the recovery rate of the foams. insole applications: the foam is slow recovery and hence,
highly conformable when a person puts their shoes on, but
becomes resilient at higher temperatures while running.
Humidity also decreases the hysterisis in these foams, the
effect being quite significant for the SR foams. Increase in
strain rate, however, increases the hysterisis o f these
foams.

206 / Naroyan, Berube

Recovery Testing

As mentioned above Figure 2 shows the recovery

profiles of the four foams after 50% compression for 15
seconds. Tables 2 & 3 also summarize the recovery testing
results (90% recovery time) as a function of temperature,
humidity and compression time for low modulus and high
modulus foams respectively. It should be noted that the
recovery tester compresses the foam at a very high
compression strain rate (~ 307min), which may be a more
accurate representation of the compression mechanism in a
shoe while jumping or running compared to the strain rate
of 27minute used for the CFD testing.
As seen from table 2, the recovery time of the low
modulus SR-LM foam drops significantly between 4°C Figure 14: Loss Modulus Measured through DMA for
and 25°C, whereas the HR-LM foam is unaffected. This SR-LM and HR-LMfoams (3 °C/min/lHz)
correlates directly with the loss modulus changes shown in
figure 14, where the SR-LM foam’s loss modulus
decreases significantly between 4 and 25°C. In the case of
SR-HM foam (Table 3) however, the recovery time
decreases from 4°C to 25°C and then increases from 25°C
to 50°C. This follows directly from the loss modulus
behavior o f the high modulus foams (Figure 15), where an
increase in loss modulus is observed between 4°C and
25°C, followed by a decrease between 25°C and 50°C.
Both the HR foams have fast recovery times and are
relatively unaffected by temperature.
Humidity has a very dramatic effect on the recovery time
of the SR foams. Samples that have been conditioned to
100% RH become very resilient. As mentioned above, it is
possible that the moisture plasticizes the hard domains.
This would decrease the molecular entanglements, which
may be responsible for the slow recovery behavior. One Figure 15: Loss Modulus Measured through DMA for SR -
reason for die large popularity of the slow recovery foams HM and HR - HM foams (3 °C/min/1Hz)
in footwear and footcare applications may be the fact that
when a user initially puts on the insole, it is highly
conformable and gives a very strong feeling of comfort. Im pact attenuation
However, during use, when the foot temperature is - 40°C
and the humidity is high, they behave as resilient materials Impact absorption properties of foam have a strong
with excellent energy return and impact absorption. correlation to the performance of a foam insole in a shoe.
Tables 2 and 3 also show that as compression time is Shock or impact absorption is required to prevent sports
increased, the recovery time increases in the slow recovery injuries. The test for impact attenuation uses a specific
foams. Compression rate has no effect on the high weight which is dropped on the foam from different
resiliency foams. This is due to the fact that the heights. The peak g force (gravities of deceleration) is used
morphology o f the slow recovery foams is highly dynamic as a measure of the impact/shock attenuation. The lower
around room temperature (Figures 3 and 4). As the foam is the value of peak g force, the better is the impact
kept compressed for longer periods of time, the molecules attenuation. The foot o f the tester hits the foam with a very
have enough mobility to rearrange their morphology, thus high compressive strain rate which decreases to zero as the
taking longer to recover to original form. foam attenuates the impact. The amount of attenuation
strongly depends on the hardness or CFD o f the foam.
Hence, if the CFD is dependent on compressive strain rate
as we have observed in Tables 2 and 3, then it can be
assumed that the attenuation will change as the foot
compresses the foam.
Tables 4 and 5 give the results of the impact attenuation
testing as a function of drop height, temperature and
relative humidity. Figures 16 and 17 show the effect of

Narayan, Berube / 207

temperature on the peak g force of the low and high
modulus foams respectively. It is observed that the slow
recovery foam (SR -LM) is better at shock attenuation that Table 4: Impact Attenuation Data as a Function of
the HR -LM foam even thought it is softer. This may be Temperature, Humidity and Drop Height on SR - LM and
either due to the fact that it exhibits a higher loss modulus HR - LM foams
or that at the compression strain rates involved in the
EFFECT OF TEM PERATURE Peak g 's
impact attenuation test, the SR-LM foam may be
significantly harder than the HR -LM foams. This would MATERIAL TEMPERATU 2" 4” 6"
yield improved impact attenuation. The strain rates RE
involved in Table 2 may not be high enough to show this 6" drop height (C)
effect. Conversely, the HR -HM foam is better than the SR- SR-LM 4 40 67 102
HM foam. As seen in Table 3, the CFD decreases 25 33 73 115
significantly as the temperature is increased, which is seen 50 43 96 142
in the impact attenuation data, where the SR-HM and HR-
HM are equivalent at 50 °C. Neither high resiliency foams
HR-LM 4 44 95 165
appear to be affected by temperature or humidity.
25 93 110 165
50 47 111 170
EFFECT OF HUMIDITY P e a k g 's
MATERIAL HUMIDITY 2" 4" 6"
6" drop height %
SR-LM 0% 32 69 114
50% R H 33 73 115
100% R H 38 88 138

HR-LM 0% R H 46 109 163

50% R H 93 110 165
100% R H 50 114 164
EFFECT OF DROP HEIGHT
MATERIAL DROP Peak
HEIGHT g 's
Figure 16: Effect of Temperature on Impact Attenuation of
SR-LM and HR-LM foams 6" drop height (Inches)
SR-LM 2 33
4 73
6 115
8 132

HR-LM 2 44
4 95
6 165
8 »

Figure 17: Effect of Temperature on Impact Attenuation of

SR-HM and HR-H M foams

2 0 8 / Narayan, Berube
 Table 4: Impact Attenuation Data as a Function of the slow recovery materials around room temperature.
Temperature, Humidity and Drop Height on SR - HM and Typically, for slow recovery foams, especially at higher
H R - H M foams modulus, the 25% CFD and recovery time decrease. The
SR foams become resilient at internal shoe use
EFFECT OF TEMPERATURE Peak g's temperatures (37°C-50°C). This is significant because it
MATERIAL TEMPERATU 2" 4" 6” gives an insight into a key advantage o f the SR foams in
RE relation to use as an insole foam. The foams are highly
6" drop height (C) conformable initially when the shoe is put on, but behaves
SR - HM 4 75 176 239 like a resilient material at shoe temperatures, which
25 95 152 185 provides energy return and cushioning properties.
The effect of humidity is large for die SR foams, while
50 96 110 113
the HR foams are relatively unaffected. It is hypothesized
that high levels of moisture in the foam affects the
HR -HM 4 67 97 118 hydrogen bonding in the hard segment, decreasing the
25 51 76 104 CFD and reducing the molecular entanglements, which
50 41 77 115 decreases the recovery time.
The effect of strain rate is quite dramatic in the SR
EFFECT OF HUMIDITY Peak g's foams, where both the CFD and recovery time (CFD
MATERIAL HUMIDITY hysterisis) increase with increasing strain rate. This is
6" drop height % explained by time -temperature superposition, higher strain
SR - HM 0% R H 95 169 200 rates are analogous to lower temperatures, which have
been shown to increase both CFD and recovery time.
50% R H 95 152 185
Impact attenuation results correlate strongly with CFD
100% RH 77 115 127 and the effect of strain rate on CFD. It is observed that for
the softer foams, the SR foam is better at impact
HR - HM 0% R H 54 77 106 attenuation than the HR foams, which may be due to the
50% R H 51 76 104 high loss modulus of the foam or significant increase in
hardness of foam at the very large compression strain rates
100% R H 43 66 109
used in impact testing. In the case of the slow recovery
EFFECT OF DROP HEIGHT Peak high modulus foams, low temperature hurts the impact
g's attenuation due to a very high CFD. At higher
MATERIAL DROP HEIGHT temperatures, the impact attenuation becomes equivalent to
6" drop height (Inches) the HR foam.
SR - HM 2 95
Hence, slow recovery Poron® foams are excellent for
use as shoe insole materials, because they are highly
4 152
conformable when the shoe is initially put on, but behave
6 185 as resilient materials with excellent shock absorption and
8 » recovery at internal shoe use conditions.

HR -HM 2 51
ACKNOWLEDGEMENTS
4 76
6 104 The authors would like to convey their sincere
8 139 appreciation to Rogers Corporation for sponsoring this
study and to the following people - Scott Simpson for his
help, guidance and useful discussions through the course of
CONCLUSIONS this study and writing of this paper; Claudette Hoffman for
the DMA work; Scott Derosier for help with FTIR
Slow recovery and high resiliency PU foams have been spectroscopy; Michelle Koss for help with the
studied for the effect of temperature, humidity, environmental chamber and Mike Bessette for
compression rate and compression time on the encouragement to complete this study and write this paper.
performance properties relevant to a shoe insole material.
It is found that the effect of temperature (4°C to 50°C) on
properties such as 25% CFD, CFD hysterisis and recovery
time can be correlated to morphological changes observed
through DMA. It is found that the high resiliency foams
are affected to a much lesser extent than the slow recovery
foams. This is attributed to the highly dynamic nature of

Narayan, Berube / 209

REFERENCES

1. Kageoka, M., Inaoka, K., Kumaki, T., and Tairaka, Y.

1998. “Low Resilience Polyurethane Slabstock Foam
with Microphase Separated Morphology”.
Polyurethanes Expo ’98, September 17-20, PI 15-120
2. Sperling, L.H., and Fay, J.J. 1991. “Factors that affect
the Glass Transition and Damping Capability of
Polymer”, Polymers for Advanced Technologies, 2(1)
P49 -56
3. Oertel, G., Polyurethane Handbook, 2nd Edition, 1993,
Hanser/Gardner Publications, Inc., P41
4. Szycher, M., Szycher’s Handbook o fPolyurethanes,
1999, CRC Press., PI 1-5

BIOGRAPHIES

Sujatha N arayan

Sujatha Narayan joined Rogers Corporation in 1998 after

receiving a Ph.D. degree in Polymer Science and
Engineering from Georgia Institute of Technology. Before
that, she received a Master’s degree in Polymer Science
and Engineering also from Georgia Institute of Technology
and a Bachelor of Technology degree in Textiles from the
Indian Institute of Technology, New Delhi, India. She
works in the Polymer Products group at the corporate R&D
department at Rogers, CT. She has worked on new product
development relating to polyurethane foams and silicone
foams and elastomers.

Amie Berube

Amie Berube joined Rogers Corporation in 1998. She

works in the Polymer Products group at the corporate R&D
department at Rogers, CT. She has worked on supporting
new product development relating to polyurethane foams
and test method development for slow recovery foams. She
is currently working towards a B. S. degree in chemistry
from Rhode Island College.

2 1 0 / Narayan, Berube
Polyurethane Elastomers with Multiple Chain Extenders

JOHN McCLUSKY, SUBIMAN GHOSH a n d MONICA A. POCOL -SAVAGE

XIN WEI Kaneka Texas Corporation
University of Texas at San Antonio 6161 Inderwood Road
6900 N. Loop 1604 West Pasadena, Texas 77507
San Antonio, Texas 78249

ABSTRACT INTRODUCTION

Polyurethane elastomers are a commercial success Polyurethane elastomers have been a commercial
because of their broad range of properties, low cost, and success because of their broad range of properties, ease of
ease of production. Nonetheless, developing new, less production, and low cost. They are used in a wide variety
expensive polyurethane elastomers with improved or of products including tires, industrial drive and conveyor
expanded properties is an important goal. Since hard belts, vibration damping materials, wheels, rollers, and
domains dominate the physical properties of these shoe soles [1]. Thus developing new, less expensive
polymers, we have investigated methods to disrupt hard polyurethane elastomers with improved or expanded
domain formation. The use of multiple chain extenders properties is an important goal.
was envisioned to inhibit hard domain formation by It is well known that polyurethane elastomers are not
disrupting the alignment between hard segments. In MDI- homogeneous polymers. They have a phase-separated
based polyurethanes, we observed up to a one magnitude structure made up of a flexible soft segment above its Tg
reduction in tensile modulus and a significant increase in covalently bonded to a rigid hard segment below its Tg or
elongation compared to polymers containing a single Tm. The hard segments are made from a chain extender
chain extender. The stress relaxation of these polymers (di- or trifunctional alcohol or amine) reacted with a
containing mixed chain extenders is lower than the control diisocyanate. In polyurethane elastomers there are usually
polymers, and there was no loss of high temperature between two and four chain extender/isocyanate pairs in
properties when multiple chain extenders were used. the hard segment. Depending on the isocyanate, chain
Furthermore, as the mole fraction of either of the two extender, and polyol used, the hard domains are frequently
chain extenders is increased from 0.5, the properties of the insoluble in the soft segment. As the polymer is formed in
resultant polymer approach those of the polymer with that these cases, the hard segments grow, become insoluble in
particular pure chain extender. Polymers containing TDI the soft matrix, and phase-separate to form hydrogen-
or HDI as isocyanate do not have such a dramatic drop in bonded hard domains [2]. The hydrogen bonding within
modulus when multiple chain extenders are used. HDI the hard domains is often thought to confer a highly
polymers are too strongly phase separated to significantly symmetrical organization (pseudo-crystallinity) to the
alter the morphology. TDI polymers produced with a hard domains; each hard domain binds many hard
strongly hydrophilic soft segment are not phase separated, segments together, thereby acting as a virtual crosslink.
while in less polar polyols the two isomers in TDI may Thus while the hard segment content usually accounts for
effectively disrupt hard domain formation independently less than half of the polymer mass, the physical properties
of the chain extenders. Nonetheless this technology is a of the polyurethane elastomers tend to be dominated by
simple and cost effective method for tailoring MDI-based these virtual crosslinks [3].
polymers’ properties over a wide range by only minimally There are three main approaches currently used in
changing the formulation. industry to control the hard domain formation. The first
involves increasing the polarity of the polyol, typically by
increasing the level of polyoxyethylene. In this case the

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 211

greater solubility of the hard segments in the soft segment oxide heterofeed with a molecular weight of
reduces the driving force for phase separation [4]. The approximately 3000, while VORANOL® CP - 1421 is a
second process uses bulky or branched chain extenders nominal triol polyether containing approximately 20%
such as 2-methyl-1,3-propanediol or diethanolamine [5], propylene oxide and 80% ethylene oxide heterofeed with
while the third involves using less symmetrical a molecular weight of approximately 4700. The polyols
isocyanates such as 2,4-toluene diisocyanate [6]. The were dried on a rotory evaporator at 80°C under 10-20
inefficient packing of hard segments containing branched mm Hg for 8-10 hours prior to use. Ethylene glycol (EG),
or asymmetric species inhibits hard domain formation, 1,3-propanediol (PDO), and 1,4-butanediol (BDO) chain
while linear hard segments fit together easily and allow extenders were purchased from the Aldrich Chemical
efficient hydrogen bonding and phase separation. While Company and distilled prior to use.
effective, these three methods tend to be expensive and
allow only a limited range of physical properties to be Instrum entation
obtained without major changes in the elastomer raw
materials. Dynamic Mechanical Analysis was carried out using a
A more versatile approach to modify polymer modulus Rheometrics DMTA-IV in tensile mode from -80 to 215
and elasticity is presented in this paper. Since the hard °C with a heating rate of 3°C/min, at 1 Hz. Infrared
domains dominate the physical properties of the polymers, spectra were obtained on a Mattson Polaris Fourier
methods to disrupt hard domain formation were transform spectrometer using attenuated total reflectance
investigated. The use of multiple chain extenders was (1x5x50mm ZnSe crystal). The stress-strain tests were
envisioned to inhibit hard domain formation by disrupting performed on a DDL 200R tensile tester at 50mm/min.
the alignment between hard segments. This paper Moduli were measured at 10% strain. Median values
describes the results of these experiments and the physical (n=5) are given.
properties of the resulting polymers.
Polymer Synthesis
EXPERIMENTAL
SYNTHESIS OF MDI-VORANOL® 3010 POLYMERS
M aterials
MDI polyurethane elastomers were synthesized with
4,4'-Methylenediphenyldiisocyanate (MDI) was donated VORANOL® 3010 using a one-shot method and 100
by the Dow Chemical Company. To purify it for polyme index, Table 1. Polymers all had approximately 35% hard
synthesis, it was melted in an oven at 45°C, filtered, and segment content; all polymers had exactly the same molar
poured into shallow foil boats under N2. After ratio of polyol, isocyanate, and chain extender. The MDI
solidification, it was broken into conveniently sized was weighed, placed in a 100-mL beaker and covered
fragments for weighing. Toluenediisocyanate (TDI, 20% with parafilm. A magnetic stirring bar was added and the
2,6; 80% 2,4) and hexamethylenediisocyanate (HDI) were beaker was placed in an oil bath at 65 °C. When the MDI
purchased from Aldrich Chemical Company and used as was melted (water-white liquid), the polyol and the chain
received. VORANOL®3010 and VORANOL® CP-1421 extender(s) were added and the solution stirred
were donated by The Dow Chemical Company. vigorously. The reaction was kept in the oil bath with
VORANOL® 3010 is a nominal triol polyether containing stirring until it stopped foaming or the viscosity increased
approximately 95% propylene oxide and 5% ethylene

Table 1. MDI-Based Polymer Formulations

Polymer MDI [g] VORANOL VORANOL EG [g] BDO [g]
3010 [g] 1421 [g]
MDI-3010-EG 24.6 52.35 - 2.98 -

MDI-3010-BDO 24.6 52.35 - - 4.32

MDI-3010-4:1 EG-BDO 24.6 52.35 - 2.38 0.86
MDI-3010-3:1- EG:BDO 24.6 52.35 - 2.24 1.08
MDI-3010-1:1-EG:BDO 24.6 52.35 - 1.49 2.16
MDI-3010-1:3-EG:BDO 24.6 52.35 - 0.75 3.23
MDM421-EG 25.46 - 50 5.38 -

MDI-1421-BDO 25.46 - 50 - 7.80

MDI-1421-EG-BDO 25.46 - 50 2.69 3.90

212 / McClusky, Ghosh, Wei, Pocol-Savage

 Table 2. Non -MDI Polymer Formulations
Polymer HDI TDI VORANOL VORANOL EG BDO Stanous % Hard
[g] [g] 3010 CP-1421 [g] [g] Octoate Segment
[g] [g] [ml]
HDI-EG 17.11 - - 50.00 5.38 - 0.07 31
HDI-BDO 17.11 - - 50.00 - 7.80 0.07 33
HDI-EG-BDO 17.11 - - 50.00 2.69 3.90 0.07 32
TDI-3010-EG - 20.06 50.00 - 5.59 - 0.02 34
TDI-3010-BDO - 20.06 50.00 - - 8.12 0.02 36
TDI-3010-EG-BDO - 20.06 50.00 - 2.60 4.06 0.02 35
TDI-1421-EG - 17.4 - 50.00 5.38 - 0.04 32
TDI-1421-BDO - 17.4 - 50.00 - 7.80 0.01 33
TDI-1421-EG-BDO - 17.4 - 50.00 2.69 3.90 0.04 33

visibly (1-5 minutes). The polymer mixture was poured RESULTS & DISCUSSION
into preheated and greased molds and cured at 90 °C for
24 hours. The glassware and syringes were dried at 125°C MDI/VORANOL® 3010 Polyurethanes
for a minimum of 24 hours prior to use. No catalyst or
other additives were used; gel times were typically 20 The tensile properties obtained for the MDI-based
minutes at 90 C. polymers are presented in Table 3. Several features are
worth noting. First is the dramatic reduction in modulus
SYNTHESIS OF ALL OTHER POLYMERS and increase in elongation realized by the use of the mixed
chain extenders. The modulus of the EG:BDO elastomer
All other polymers were synthesized using the one-shot was reduced over one magnitude compared to the EG
method at 100 index. Total moles of chain extender were polymer, and a factor of four compared to the BDO
kept constant, so percent hard segment varied from 32 to polymer. O f even greater importance is that the modulus
40 percent, Table 2. The isocyanate was measured into a can be tuned to any value between that of the pure chain
100-mL beaker and covered; MDI was placed into a 85C extender and that of the 1:1 mixture simply by adjusting
oven to melt. The polyol was weighed into a the ratio of chain extenders, Table 3. It can be assumed
polyethylene beaker, and the catalyst and chain that the covalent matrix is nearly identical in the polymers,
extender(s) added. The isocyanate was added with since a constant molar ratio of isocyanate, polyol, and
vigorous stirring under a stream of dry argon. The total chain extender was used throughout the series.
reacting mixture was poured into a polyethylene form, Swelling experiments were conducted for the mixed and
then degassed at 10-20 mm Hg for 15 minutes. After pure chain extender polymers in order to investigate the
gelling the polymer was placed into a 85 C oven to cure crosslink density of the elastomers. The polyurethane
for 24 hours. elastomers were swelled in three different solvents at
room temperature for four days. The swelling percentages
(by weight) are presented in Table 4 for 100%EG,
100%BDO and 1:1 EG:BDO polymers. The swelling
data for diethyl ether are indicative of total effective
Table 3. Physical Properties of MDIA/ORANOL 3010
Polymers
Chain Elastic Ultimate Elongation
Extender modulus strength [percent] Table 4. Solvent Swelling of MDIA/ORANOL 3010
[MPa] [MPa] Polymers
EG 36.4 75.5 185 Solvent 100% EG 100% BDO 1:1
BDO 11.4 39.1 145 EG:BDO

1:1 EG:BDO 3.40 30.9 275 Diethyl 17.5% 16.9% 21.0%

ether
3:1 EG:BDO 29.8 53.1 180
THF 44.4% 49.7% 54.4%
4:1 EG:BDO 32.2 36.0 75
DMAC 45.0% 43.5% 51.4%
1:3 EG:BDO 7.60 36.5 160

McClusky, Ghosh, Wei, Pocol-Savage I 213

 Stress relaxation measurements at 25% strain have been
completed on these elastomers. During the first 30
minutes, the EG based polymer relaxed 13.7%, the BDO
based polymer relaxed 10.0%, and the 1:1 EG:BDO based
polymer relaxed 6.2%. While the linear logarithmic
relationship between stress and time has no theoretical
basis, such relationships are commonly observed [7] and
allow one to calculate a stress decay rate using linear
regression. It is observed that the EG based polymer has
the greatest stress decay rate at 0.042 MPa/sec, the BDO
has an intermediate rate of 0.025 MPa/sec, and the
EG:BDO elastomer has the slowest decay rate of 0.017
F igure 1. In fra re d S pectra o f M D IA /O R A N O L 301 0
MPa/sec. The low stress relaxation in the EG:BDO
P o lyu re th a n e E lastom ers.
polymer was initially unexpected given the initial
hypothesis that polymer softening is caused by disruption
of the hard domains. However, it is likely that stress
crosslinking contributed from both the covalent triol and relaxation at 25% strain does not entail any chain scission
the virtual crosslinking from the hard domains. The data but rather arises from loss of chain entanglements in the
show a significant decrease in total crosslinking in the soft segment and rupture of hard domains. (This is in part
polymer with mixed chain extenders. Swelling with supported by infrared observations of these polymers 90
DMAC is more indicative of covalent crosslinking only, seconds following tensile failure: a 15% reduction in
and also shows a small reduction in polymer containing urethane hydrogen bonding was observed which
two chain extenders. recovered over eight minutes.) Since the hard domains in
Infrared spectra of these polymers demonstrate a strong the EG:BDO polymer contribute proportionally less to its
disruption of the hard domains, Figure 1. Note that the modulus, it is consistent that this polymer will have a
elastomer with two chain extenders had much more free correspondingly smaller stress relaxation since the stress
urethane (1725 cm-1) and less hydrogen bonded urethane relaxation is more concentrated in the elastic soft segment.
(1705 cm'1) exhibited in its spectra than either the BDO or The low stress relaxation bodes well for applications
EG polymer. These data suggest that the hard domain involving polyurethanes with multiple chain extenders.
formation is indeed being disrupted upon use of multiple Polyurethanes are known for their wide range of service
chain extenders. temperature; this is demonstrated in the dynamic
It is not very likely that in the case of the polymer using mechanical analysis (DMA) curves for the EG, BDO, and
both EG and BDO phase separation occurs to form three 1:1 EG:BDO based elastomers, Figure 2. All samples
types of hard domains: those containing only BDO hard have a relatively constant modulus over a temperature
segments, those containing only EG hard segments, and range of more than 150°C. Comparison of the tan 5
those hard segments containing both EG and BDO. A curves of the polymers is informative. The EG:BDO
normal statistical distribution does not favor individual polymer has a Tg five degrees higher than that of both the
separation of the chain extenders, and the existence of BDO and EG based polymers, while the width of the Tg at
such pure hard domains would still impose a higher half height is seven degrees narrower. This indicates that
ordering than it is seen in the elastomers investigated. the EG:BDO based polymer is somewhat more phase
Differences in polymer physical properties must then arise mixed and has a lower effective level of crosslinking [8].
from changes in the polymer morphology: extent of phase This is consistent with the swelling experiments and the
separation, domain size and content, and the interfacial
region. Furthermore, it is known that increasing levels of
hard domains (either by increasing the level of chain
extender or decreasing the phase mixing) raises the
polymer modulus and decreases the elongation. Thus, the
dramatic reduction in modulus from multiple chain
extenders must arise from changes in polymer
morphology, more specifically a reduction in hard domain
content. This is further supported by the observed
relationship between modulus and the molar ratio between
chain extenders, Table 3. The more random the hard
segments formed, the higher the kinetic barrier for hard
domain formation. This results in a lower modulus and a
higher elongation at break. F igure 2. M D I V O R AN O L 3 0 1 0 DMA data.

214 / McClusky, Ghosh, Wei, Pocol-Savage

 F igure 3. M D I V O R AN O L 1421 tensile data. F igure 4. TDI V O R AN O L 1421 tensile data.

lower modulus of this polymer, although the difference in DMA of these polymers is similar with those of the
moduli as measured by DMA is smaller than observed on VORANOL 3010 based polymers: the Tg of the 1:1
the large-scale deformations of the tensile test. EG:BDO polymer is 8 °C higher than the EG-based
Furthermore these results are consistent with our polymer and 18 °C higher than the BDO-based polymer,
hypothesis that the use of mixed chain extenders results in yet it has the sharpest glass transition.
an increase, although modest, in phase mixing in the
polymers. TDI - based Polyurethanes
It is interesting to note that the BDO and 1:1 EG:BDO
polymers have very similar high temperature softening The use of mixed chain extenders with TDI and
temperatures Tm, while the EG polymer begins softening VORANOL® 1421 obeyed the rule of mixtures: there was
roughly twenty degrees lower. This suggests that the hard little change in any of the physical properties, and the
domains in the EG polymer may actually be less ordered values for the elastomer produced with mixed chain
than those in either the BDO or EG:BDO polymer. This extenders was intermediate between those of the pure EG
is contrary to what is expected given the higher flexibility and pure BDO based polymers Figures 4, 5. This was not
of the BDO polymer and the multiple chain extenders in unexpected: these polymers are transparent, suggesting
the EG:BDO sample. It is likely that at least some of the little or no phase separation. The IR spectra show only a
differences in Tm are caused by the kinetics of gelation weak shoulder for hydrogen bonded urethane, further
and phase separation. supporting the low phase separation, Figure 6. This then
is why the rule of mixtures is followed: there are very few
MDI/VORANOL® 1421 Polyurethanes hard domains to disrupt, so use of multiple chain
Polymers synthesized using VORANOL 1421, a highly extenders results in a very minimal change in morphology.
hydrophilic soft segment, were very similar to those Elastomers produced with TDI and VORANOL® 3010
synthesized with VORANOL 3010. Again there was a also obeyed the rule of mixtures: the moduli and DMA
large reduction in modulus upon using the mixed chain spectra all indicated that the elastomer with mixed chain
extenders Figure 3, and the corresponding reduction in extenders exhibited properties intermediate between those
hydrogen bonded urethane in the infrared spectra. The of pure EG and pure BDO. The infrared spectra also

F igure 6. In fra re d S pectra o f TDI V O R AN O L 1421

P o lyu re th a n e E lastom ers.
Figure 5. TDI VORANOL 1421 DMA data.

McClusky, Ghosh, Wei, Pocol-Savage / 215

 F ig u re 7. H D I V O R A N O L 1421 tensile data.
F igure 8. In fra re d S pectra o f H D I VO R AN O L 1421
P o lyu re th a n e E la sto m e rs

showed an intermediate amount of hydrogen bonding in which minimizes the driving force for phase separation.
the mixed elastomer. It is not at first apparent why this The drive for phase separation in these polymers is so
should be so. The amount of hydrogen bonding in the strong that changes in hard segment interactions cause
MDI and TDI/VORANOL® 3010 polymers are very only a minimal change in hard domain formation. Thus
similar, suggesting similar thermodynamic driving forces mixed chain extenders are not a reasonable method for the
towards phase separation. It is therefore unexpected that control of HDI based polyurethane elastomers.
the changes in hard segment size and shape do not inhibit
hard domain formation. We believe that this behavior CONCLUSIONS
may be a result of the 80:20 isomeric mixture inherent
within the TDI itself. If there is already difficulty in MDI-based polyurethane elastomers showed a dramatic
aligning the urethane groups during phase separation due reduction in tensile modulus and an increase in elongation
to the TDI isomers, then any additional barriers to phase when two different chain extenders were used in the same
separation may have only a limited effect. polymer. It is hypothesized that the use of multiple chain
extenders disrupts hard domain formation and slightly
HDI -based Polyurethanes increases the phase mixing between the soft and hard
segments. The maximum change in properties occurs at
The polyurethanes were synthesized with hexamethylene equal molar ratio of the two chain extenders. As the
diisocyanate, VORANOL® 1421, and either EG, BDO, or polymer is enriched in one of the chain extenders, the
1:1 EG:BDO. While the elastomer with mixed chain properties of the resultant polymer approach those of the
extenders showed a somewhat lower modulus at polymer with that particular pure chain extender. Thus it
elongations less than approximately 20% and above is very easy to produce polymers with a wide range in
approximately 100%, the modulus was higher at properties by only a minimal change in composition. The
elongations inbetween, Figure 7. Infrared spectroscopy stress relaxation of polymers containing mixed chain
indicates that all of these polymers have extremely high extenders is lower than the control polymers; this may be
phase separation; however, the EG:BDO polymer has a due to the decreased contribution of the hard domains to
wider hydrogen bonded peak indicative of less ordered
hard domains, Figure 8. DMA results indicate that the
mixed chain extender elastomer has a slightly lower Tg
than the BDO-based polymer and much lower Tg than the
EG-based polymer; the maximum of the tan 5 peak is
intermediate between the two. The modulus at room
temperature is intermediate between the EG and BDO
polymers, and it has a broader Tan 5 peak, Figure 9. Note
that the transition at ca. - 15 °C is due to polyethylene
oxide crystallization in the soft segment
These data indicate that there is only a minimal change
in polymer morphology upon use of mixed chain
extenders when HDI is used as isocyanate, resulting in a
minimal change in properties. Note this is true even with Figure 9. HDI VORANOL 1421 DMA data.
the very hydrophilic VORANOL® 1421 as soft segment

216 / McClusky, Ghosh, Wei, Pocol-Savage

the polymer modulus. DMA and swelling data are BIOGRAPHIES
consistent with an increased level of phase mixing and
decreased crosslink density in the EG:BDO polymer John M cC lusky
compared to either the EG or BDO polymers; moreover
John McClusky received his
the EG:BDO elastomer has similar high temperature
BS in Chemistry from Texas
properties to the control polymers. The modification of
Lutheran University and his
properties for HDI based polymers was minimal due to
Ph.D. in Chemistry from Yale
the very high phase separation present in these polymers.
University. He spent five
TDI based polymers followed the rule of mixtures,
years at Dow Chemical
possibly due to the isomeric mixture of TDI disrupting
working on polyurethane
phase separation independent of the chain extenders. This
polymers, after which he
technology, then, is a simple and cost effective method for
joined the faculty of the
controlling MDI-based elastomer properties over a wide
University of Texas at San
range. Work is continuing to understand how multiple
Antonio where he is an
chain extenders change the physical properties of
Associate Professor.
polyurethane elastomers.
M onica Pocol - Savage
ACKNOWLEDGEMENTS
The authors acknowledge the Robert A. Welch Monica Pocol-Savage
foundation grant AX -1474. This material is based in part received her BS in
upon work supported by the Texas Advanced Research Chemistry from Politehnica
Program under Grant No. 003658-0059b-1999. The Univeristy of Bucharest,
authors acknowledge Dow Chemical Company for MDI Romania, and Masters
and VORANOL® polyols. Degrees in both Chemical
Engineering from
REFERENCES Politehnica Univeristy and
1. Gribble, MY. 1996. Proceedings o f the Utech ’96 in Chemistry from the
Polyurethane Conference, March 26-28, 1996, pp. 45. University of Texas at San Antonio. She is presently an
2. Lee, KJ; Wnag, YK; Macknight, WJ; Hsu, SL, 1988. Advanced Chemist at Kaneka Texas Corporation.
Macromol, 21:270-273. Priester, RD Jr.; McClusky,
JV; Cortelek, DI; Carleton, PS; Porter, JR; S ubim an G hosh
deHaseth, JA, 1994. J. Reinforced Plastics and Subiman Ghosh obtained his
Composites, 13:68-86. BS in Chemistry and MS in
3. Lee, PC; Chen, LW; Huang LLH. 1996. Polymer Physical Chemistry from
International, 41:419. Peffley, RD; Spells, JH; Presidency College, Calcutta
Vanderhider, JA. 1987. US Patent 4,689,356. University, India. He
4. Lan, PN; Comeillie, S; Schacht E.; Davies, M; Shard, obtained his Ph.D. from the
A. 1996. Biomaterials, 17:2273. Zetterlund, P; Rubber Technology Centre,
Hirose, S; Albertsson, AC. 1997. Polymer Indian Institute of
International, 42:1. Mitzner,E; Kennedy, JP. 1997. Technology in Kharagpur,
Pure Appl. Chem., A34:165. Martin, DJ; Meijs, GF; India. Since that time he has
Renwick, GM. 1996. J. Appl. Polym. Sci., 60:557. been a postdoctoral associate at the University of Texas at
5. Priester, RD Jr.; Strojny, EJ; Stutts, DH. 1990. US San Antonio.
Patent 4,931,487. Saito, J; Wada, H; Shigehito, S;
Watabe, T; Tanabe, K; Kunii; N. 1989. US Patent Xin W ei
4808636.
6. McClusky, J.V.; O’Neill, R.E.; Priester, R.D.; Turner, Xin Wei received both his BS
R.B. 1990. Journal o f Cellular Plastics, 23:346. and Ph.D. degrees in Polymer
Creswick, M.W.; Lee, K.D.; Turner, R.B.; Hubner, Chemistry and Physics from
L.M. 1988. Proceedings o f the Polyurethane World Nanjing University.
Congress, Oct. 18-21, 1988, p. 11. Following a postdoctoral
7. See for example Moorland, JC; Wilkes, GL; Turner, assignment at the University
RB. 1994. J. Appl. Poly. Sci., 52:569. Sung, CSP; of Oklahoma, he joined the
Wu, CB. 1980. Macromol. 13:111. Smith, TL; University of Texas at San
Dickie, RA. 1969. J Polym Sci., Part A-2, 7:635. Antonio as a postdoctoral
associate where his is
8. Kumar, A.; Gupta, R.K. Fundamentals o f Polymers, presently employed.
McGraw-Hill Companies, Inc., 1998.
McClusky; Ghosh, Wei, Pocol-Savage / 217
The Effects of Temperature/Pressure Gradients in the Rigid
Polyurethane Foaming Process

BERNARD E. OBI, IKE LATHAM a n d LINDA J . BROWN

The Dow Chemical Company
2301 North Brazosport Blvd.
Building B-1608
Freeport, Texas 77541

ABSTRACT gradients, were too small to cause de-mold splits in the

foam.
The United States appliance industry must meet the U.S. The authors’ recommended increased green strength at
Department of Energy (DOE) requirement for energy the same de-mold time and percent overpack. To achieve
reductions of approximately 30% (from the 1993 standard) this, it was suggested that a higher functionality polyol
by July 2001. In addition, the manufacture of HCFC-141b, would lead to a foam with higher green strength at the time
the predominant blowing agent used for rigid foam in of de-mold.
North America will be discontinued in the United States on
January 1, 2003. These two directives have driven INTRODUCTION
appliance OEM’s to evaluate new cabinet designs in
conjunction with alternative blowing agents. In response On the one hand, energy scarcity and need for
to these needs, The Dow Chemical Company began a study conservation have mandated governments to pass
to evaluate the effects of shorter de-mold times and regulations that impose more stringent energy efficiency
increased foam thickness, as OEM's are driven to improve requirements on both the appliance and construction
productivity by decreasing cycle time, while driving industries. On the other hand, environmental
towards reducing energy consumption. considerations (i.e. ozone destruction concerns) have also
It was hypothesized that pressure and temperature necessitated governments to abolish the use of energy
gradients observed throughout the foam, could lead to efficient CFCs/HCFCs blowing agents, which are known
unbalanced stresses across the part promoting de-mold to be global warmers and/or ozone depleting agents. To
splits in a typical Polyurethanes appliance formulation. It meet these juxtaposed requirements to these industries,
was further suggested that increased green strength would OEMs have been forced to evaluate new cabinet designs
be necessary at the time of de-mold, so the foam could (often requiring larger wall thickness), in conjunction with
resist greater expanding forces. alternative blowing agents. In response to these needs, The
First, modeling of the temperature profile across the Dow Chemical Company began a study to evaluate the
foam thickness was performed. The corresponding gas effects of de-mold times and increased foam thickness, as
pressures and pressure drops as a function of position OEM's are driven to improve productivity by decreasing
across the foam thickness were calculated using the cycle time, while continuing to drive towards reducing
assumption that all gases were ideal. The second model energy consumption.
estimated the critical gelation point (conversion), as a In carrying out laboratory studies on machine runs, it
function of both the isocyanate and polyol functionalities. was noticed that de-mold splits occurred more frequently
Thirdly, from calorimetric measurement, the time- with larger molds than with smaller ones. Also, de-mold
conversion profile for the foaming polymerization reaction splits were more frequent with shorter de-mold times and
was obtained. at the higher over-pack levels necessary to get at density
Examination of the modeling and experimental data targets required to maintain dimensional stability in the
reinforces the hypothesis that pressure differences within cabinet.
the foam coupled with inadequate green strength caused by To evaluate the nature of these problems, it was
de-molding at conversions near or below the critical hypothesized that there exist sufficient temperature and
gelation point, would lead to de-mold splits. From the hence corresponding pressure gradients across the foam
analysis, it was also observed that at a two inch foam matrix, from the center of the foam to the surface. It was
thickness, the temperature and consequently pressure further inferred that these pressure gradients could lead to

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 219

unbalanced stresses across the part promoting de-mold BACKGROUND
splits. It was also suggested that with insufficient green
strength at short de-mold times, the foam could not resist A typical polyurethane appliance system was used to
greater expanding forces due to the pressure gradients thus explore the effect of foam thickness, de-mold time, and
leading to stress cracks and splits on de-mold. over-pack level on foam stability (de-mold splits). Figure
1 shows some experimentally determined and calculated
OBJECTIVE process parameters for the foaming reaction.
From figure 1, it can be observed that the time to reach
The objective was to use simple mathematical models maximum foam height was around 60 seconds, which
combined with targeted experimentation to elucidate the correlates closely with the tack-free time. The observed
nature of de-mold splits in rigid polyurethane foams. temperature rise and the corrected temperature rise profiles
are also plotted as a function of foaming reaction time in
figure 1. The corrected temperature versus time profile
was calculated by accounting for the overall rate of heat
loss in these systems during the foaming process. Based
on an estimate of the heats of reaction for the water and
polyol reactions with isocyanate1, and an estimate of the
foam average heat capacity, the overall time-conversion
profile for the foaming reaction was estimated as shown in
figure 1.

EXPERIM ENTAL W ORK

Evaluation of the Effect of Foam Slab Size and De -mold
times on Foam Splits
Table 1 tabulates the results from experiments carried
out to evaluate the effects of foam thickness and de-mold
times on post expansion splits. This table contains data
from foam made with Formulation A (FA) which had an
Figure 1. Typical Appliance Foaming Reaction Parameters.
average polyol functionality of 3.9. These experiments
were carried out in a Brett mold (2” thick) and in specially
Modeling of the temperature profile across the foam
designed Jumbo molds with thicknesses of 3 and 4 inches.
thickness was performed. The corresponding gas pressures
Mold temperature (foam surface temperature) was about
and pressure drops as a function of position across the
125 °F. De-mold times from 2.5 to 4 minutes were
foam thickness was then calculated making the assumption
examined in combination with the mold sizes.
that all gases inside the foam cells were ideal. Secondly, a
It can be clearly seen that foams made in larger molds,
model estimation of the critical gelation point (critical
or with shorter de-mold times, showed higher levels of post
conversion to form a polymer network), as a function of
expansion splits. The worst post expansion splits were
both the isocyanate and polyol functionalities was carried
observed for run #6, involving a 3 inches jumbo mold and
out. From the temperature measurement, the time-
shorter de-mold time of 3 minutes. Run #5 had the second
conversion profile for the foaming polymerization reaction
worst split. Slab size and de-mold time were similar for
was estimated. Finally, analysis of the combined results
runs #5 and 6, however, run # 6 had a much higher over­
from both the experimental and modeling work was used
pack than run # 5. Higher over-pack translates into more
to validate the hypothesis on de-mold splits in rigid PU
blowing agent as well as material for the same foam
foams, as well as suggest ways to design formulations that
volume, hence higher density foam will result, as well as
will mitigate this problem.
higher internal pressures.
Table 1. Effect of Foam Thickness and De-mold times on Post Expansion Splits - Formulation A (FA)
Run # & Mold Type Mold Density Over-Pack De-mold De-mold Splits
Thickness % Time (Qualitative)
( Inches) (PCF) (minutes)
Run#1 - Brett 2 2.17 10 2.5 0
Run#2 - Brett 2 2.24 14 2.5 0
Run#3 - Jumbo 3 1.95 10.7 3 2
Run#4 - Jumbo 3 1.99 12.8 3 2
Run#5 - Jumbo 3 2.04 15.8 3 3
Run#6 - Jumbo 3 2.17 22.6 3 4
Run#7 - Jumbo 3 2.17 22.6 4 1
Run#8 - Jumbo 4 1.92 10.9 4 2
Legend: 0 = No Split, 1 = Very Slight, 2 = Slight, 3 = Moderate, 4 = High

220 / Obi, Latham, Brown

 Figure 2. Schematic showing hypothetical temperature and pressure profiles in a foam
matrix as functions of position and time.

green strength at the time of de-mold to withstand these

Analysis stresses would not split, while those with poor green
strength at de-mold would have significant splits.
A significant temperature gradient exists along the width
of the foam from the center to the surface. A Simulation of the Tem perature and Pressure Profiles in
corresponding pressure gradient exists along the width of a Foam M atrix
the foam due to this temperature gradient. The pressure
gradient results in differential stresses perpendicular to the The development of the temperature profile in a reacting
rise direction in the foam. The magnitude of the pressure foam matrix is a complex event involving polymerization
gradients, hence the magnitude of the stresses pushing on reactions, which are exothermic, blowing agent
adjacent foam layers along the thickness direction is a evaporation which is endothermic, and heating, followed
function of the foam thickness. The larger the foam by cooling via the mold surface. With limited kinetic
thickness, the greater the pressure gradients and the parameters to properly account for the energy liberation
stresses for a given foam. In addition, higher gas loadings rates during the foaming reaction, it is difficult to develop
in the foam lead to higher pressure gradients and a rigorous process model that accurately predict the
corresponding stresses. Figure 2 shows a schematic temperature profile of a foaming matrix. However, with
depicting a hypothetical temperature and pressure profile some simplifying assumptions, one can reasonably model
in the foaming matrix along the foam thickness and as a the cooling process, and hence generate information that
function of time. closely simulate the true temperature profile in a foam
matrix. In this work, the following assumptions were
Hypothesis made:

It is hypothesized that these differential stresses along 1. The foam is fully expanded in about 60 seconds.
the foam thickness provide the mechanism for foam splits 2. From the measured temperature profiles for appliance
when the restraining force of the mold wall is removed on formulation foams, it can be generally observed that
de-mold. The extent and severity of the splits on de-mold the center of the foam reaches a temperature of
depend on the level of green strength developed in the approximately 140 °C in 2.5 minutes.
foaming matrix before de-mold. It can be surmised that a 3. The heat losses are due to the temperature gradients
foam made from a formulation which builds adequate relative to the surrounding, as well as evaporative

Obi, Latham, Brown / 221

 action of the organic blowing agents. (The carbon
dioxide production was assumed not to contribute to
heat loss from vaporization.)
4. Even though the foam slab is cellular, it is treated as a (2)
continuous solid slab for the purposes of modeling w hen B i —» oo
heat conduction. The thickness of the foam is so much
less than the other two directions (height and length) The general transient heat conduction equation at
that the effective path for the transport of heat is constant pressure with a heat source term can be found in
essentially across the thickness only. many texts on heat transfer 2,34,5 and is given in equation 3
5. Although the polyurethane reaction is only about 50 % below.
complete at 3 minutes, we do not have the kinetic
parameters to properly solve the transient heat (3)
conduction equations with generation terms and thus
this heat generation is ignored. However, with the rate
of temperature rise approaching zero due to a where S dot is the heat source term, and V is the linear
continuously slowing reaction rate and heat loss vector operator del. When terms are collected assuming a
through cooling, this assumption is reasonable. constant value for the thermal conductivity we have
6. In order to model the temperature profile in the foam equation 4.
slab, the initial temperature (T0) will be estimated
iteratively from knowledge of the center point
temperature at the time of de-molding the foam. (4)

M athematical Equations for the Transient Heat The slab is taken to be infinite in the y- and z- directions so
Conduction in a Finite Slab our governing differential equation of heat conduction in
the x- direction takes the form of equation 5.
Figure 2, is a schematic of the finite foam slab showing
the geometric arrangement for the conduction problem.
The temperature distribution, Tx„ at any time t and position
x is found from dimensional analysis to be: (5)

Equation 5 can be rewritten as equation 6 when the

(1) source term is neglected.

where Twis the temperature of the surface of the slab, T0 is

( 6)
the temperature at the center of the slab, h is the heat
transfer coefficient, 2L is the thickness of the slab, k is the
thermal conductivity, a is the thermal diffusivity Analytical solutions exist for equation 6 subject to the
(a = k/pCp), p is the density and Cp is the heat capacity. constraints of equation 2 and the initial and boundary
The following dimensionless parameters can be defined: conditions below.
Biot number, Bi-hLIk; Fourier number, F 0l=(^ t/L2 and the
dimensionless distance from the center of the slab, X=x/L. Initial and Boundary Conditions (Bi—>oo):
For a heat barrier material, like polyurethane foam
inside a metal mold, the foam’s thermal conductivity is
very much smaller than the surface heat transfer
coefficient. As such, the Biot number is very large
indicative of the fact that the barrier to energy transport in
the system is strictly within the foam slab. Therefore, the
temperature distribution is not dependent on the external
resistance, and the first term in equation 1 drops out, and
equation 1 reduces to equation 2 below.

Equation 7 below is a solution of equation 6 subject to

the boundary conditions above.

222 / Obi, Latham, Brown

 (7)

where

gradient from the center outward to the surface. Figure 5

A spreadsheet was used to obtain solutions for equation below show plots of the change in pressure between
7 as a function of the slab thickness position and time. adjacent vertical plains along the foam slab thickness as a
T{0to) is a hypothetical foam initial maximum temperature function of position and slab size.
that the foam will be at, at the center of the foam and time The pressure drops shown in figure 5 were calculated
zero. T(0,o) is iteratively obtained from knowledge of based on a constant spacing of 0.05 dimensionless distance
T(o,i50) (i-e. temperature at the center of the foam at time units between adjacent plains. Thus for a 2 inch foam slab,
equal to 150 seconds). Tw is the mold wall temperature pressure drops were calculated between plains separated by
before de-mold (51.5 °C), and is the ambient temperature 0.05 of an inch. Similarly, the pressure drops for the 6
after de-mold (25 °C). inch foam slab were estimated for plains separated by 0.15
of an inch.
Results of the Simulation Figure 6 below shows the overall pressure drop across
the foam thickness at the time of de-mold of 180 seconds.
Figure 3 is a plot of the temperature versus time and It is worthwhile to point out that for each slab size, the
position from the center (0”) to near the surface (1.98”), as bulk of these pressure drops appear to be concentrated in a
shown in the legend, for a 4 inches thick foam slab. The small region about midway between the center and the
discontinuity in the plot is indicative of the effect of surface of the foam as can be seen in figure 4. These
opening the mold and the change in surface temperature. positions will experience the maximum stresses that must
The de-mold time was set at 3 minutes (180 seconds). be overcome by the strength developed in the polymer at
With the approximation of ideal gas behavior, the gas the time of de-mold.
pressures inside the foam cells were estimated from the
temperatures shown in figure 3 as a function of time and Interplay of Internal Stresses and the Foam Structures
position. Estimated pressure curves are shown in figure 4 at De -mold
below. These gas pressures were calculated assuming a
typical packing level and uniform cell sizes throughout the Rigid polyurethane foams develop their rigidity relatively
foam. early in the polymerization due to molecular weight build­
When one considers that adjacent foam cells interact up in the cross-linked system. The level and extent of
with each other via their cell walls, one finds that variation cross-linking during the rigid polyurethane foaming
in cell pressure across the foam thickness results in a reaction is related to the functionalities of the isocyanate

Figure 3. Cell temperatures versus time and position (0” Figure 4. Cell gas pressures versus time and position
=center of slab) & (1.98” = near surface of slab) for a 4 inch (0”=center of slab) & (1.98”= near surface of slab) for a 4
thick foam slab. Estimated T(0,o) = 151.7 °C inch thick foam slab.

Obi, Latham, Brown / 223

 and polyol sides, as well as the rates of reactions. As is
seen in figure 1, the reaction is far from complete at the
time of de-mold (3 min.). It will be useful to understand
how much conversion has occurred at the time of de-mold
in order to relate that value to observed post-expansion
splits.
There is abundant literature in polymer science on the
topic of cross-linked systems6,7,8,9. A simple model that
describes gelation in cross-linking systems was used to
model the critical conversion for gelation in the
polyurethane system. Critical conversions for gelation
during polymerization were evaluated as a function of the
polyol and isocyanate functionalities. Figure 7 below
shows a plot of these results. A band is shown in figure 7
indicative of the range of conversions typically achieved in
the polyurethane polymerization in the 3 minutes time
before de-mold.
Figure 5. Pressure differences between adjacent positions
It is clear from figure 7 that the higher the functionalities
across the foam thickness at the time of de-mold of 180
of both components, the lower the critical conversion for
seconds versus position as a function of slab thickness
(X=0; center of slab) & (X=0.50; surface of slab) gelation. Thus, foams made with higher functionality
polyol and isocyanates will reach gelation much sooner
and therefore be more resistant to stress that might cause
failure in the foam material. Critical gelation is by
definition the point in the polymerization of a cross-linked
system at which an infinite network is achieved. While a
noticeable exponential increase in viscosity occurs near the
critical gelation point, structural strength (modulus) is only
beginning to develop at this stage. Figure 8 below is a
simulated display of the relationship between the viscosity
(p), Modulus (E) and soluble weight fraction (fsoi), as a
function of conversion (P) for a gelling system. From this
figure, it can be seen that on approaching the critical
conversion for gelation (Pc) (dotted line), significant
changes take place in the reacting mass. One can clearly
see that at the critical gelation point, the soluble fraction is
Figure 6. Overall pressure-drop across the foam thickness disappearing at an exponential rate while the viscosity goes
at the time of de-mold of 180 seconds as a function of slab to infinity. Although the foam can be considered a solid at
thickness. this point, the green strength or modulus has barely begun
to grow. The growth in modulus is sigmoidal and so the
polymer toughens-up some time after gelation had been
attained.

Validation of Hypothesis

It can be seen from figure 1 that full expansion of the

foam is achieved in about one minute after mixing - long
before de-molding the foam. Modeling has shown that
large temperature, and consequently pressure gradients
exist across the foam thickness very early in the foaming
reaction, when only a modest conversion had been
achieved. The pressure gradients result in stresses across
the foam thickness, which on de-molding, have to be
overcome by the foam’s green strength or foam failure will
result. Table 1 has shown that significant foam splits can
result when de-molding takes place before adequate foam
Figure 7. Critical Gelation Conversion Points versus Polyol modulus (green strength) is developed. This happens
and Isocyanate Functionalities (fNco)- either when the de-mold time is very short as in faster

224 / Obi, Latham, Brown

 Formulation A (FA) was modified to design formulation
B (FB) whose data is shown in table 2 below. This
modification was made in order to validate the hypothesis
that higher green strength could overcome the pressure
gradient induced stresses within the foaming mass that lead
to splits on de-molding. The major change in the
formulation was increased polyol functionality. Whereas
the average functionality of the polyol mixture in FA was
3.9, the average functionality of the polyol mixture in FB
was 4.7. The average functionality of the isocyanate was
held constant in both systems at 2.7.
From table 2, it is obvious that the degree of post
expansion splits in the foams of (FB) is very small to non­
existent. The experimental conditions for Run #5 of table
1 (FA) are comparable to run #7 of table 2 (FB). However,
the data showed very clearly that in run #5 (FA) moderate
amounts of splits were observed, while run #7 (FB) had no
to a very slight amount of splits.
Figure 8. Physical characteristics of a gelling system From figure 7 it can be seen that (FA) foaming reaction
showing the relationship between the viscosity (p), the gets de-molded just in the time frame of gelation and thus
soluble fraction (fSOi) and the modulus (E) as a function of may have very low green strength. On the other hand,
conversion (Pc). (FB) foaming reaction gets de-molded some time after
gelation had taken place at a stage of higher green strength.
production environment, or when there is need to have a These observations support the data reported in table 1
larger wall thickness which will exacerbate the pressure (FA) and table 2 (FB) respectively.
gradients. It appears that OEMs have a need for both faster
production capabilities (faster de-mold), as well as larger CONCLUSIONS
wall thickness (to meet energy requirements). These two
conditions are directly juxtaposed to produce defect free We have attempted to use selected experiments, some
foam. In consideration of the information provided in simple mathematical modeling of heat transfer processes
figures 7 and 8, the only means of reaching the goals of and polymer rheological behavior to validate a hypothesis
faster de-mold and larger wall thickness is by building on how post expansion splits occur in rigid foams. In short
higher modulus at the time of de-mold. To achieve this, we think that post expansion splits occur in rigid foams
figure 7 would suggest that foam formulations should when faster de-mold times result in poor green strength
maximize the system functionality. However, the need for and high interlayer stresses due to temperature/ pressure
faster modulus development must be balanced by the need gradients within the foam matrix. The situation becomes
for other foam attributes, particularly foam flow. Hence progressively worse with larger foam thickness (larger
the system functionality needs to be optimized taking into gradients) and higher over-pack levels (more gas for same
consideration requirements for faster de-mold and larger volume).
walls on the one hand (higher functionality) and foam flow Therefore, in the face of needs for increased productivity
which might be adversely impacted by higher (faster de-mold times) as well as larger wall thickness,
functionalities. formulations must be designed taking into account the need

Table 2. Effect of Thickness and De-mold time on Post Expansion Splits Formulation B (FB)
Run # & Mold Type Mold Density Over-Pack De-mold De-mold Splits
Thickness % Time (Qualitative)
( Inches) (PCF) (minutes)
Run#1 - Brett 2 2.26 11 2.5 0
Run#2 - Brett 2 2.30 14 2.5 0
Run#3 - Jumbo 3 1.98 10.7 3 0
Run#4 - Jumbo 3 1.99 11.0 2.5 0-1
Run#5 - Jumbo 3 2.03 13.6 3 0-1
Run#6 - Jumbo 3 2.03 13.2 2.5 0
Run#7 - Jumbo 3 2.08 16.2 3 0-1
Run#8 - Jumbo 3 2.08 16.2 2.5 0-1
Run#9 - Jumbo 4 1.99 12.6 4 0
Run#10 - Jumbo 4 1.99 12.6 3 0-1
Legend: 0 = No Split, 1 = Very Slight, 2 = Slight, 3 = Moderate, 4 = High

Obi, Latham, Brown / 225

for rapid modulus development to counter the larger The Dow Chemical Company in July 1985 in the research
internal stresses. The first line of action is the design of assignments program, and in 1987 began a permanent
formulations which utilizes higher functionalities to assignment in the Saran™ polymers R&D organization. In
achieve rapid modulus development, while recognizing the 1995 he earned a Ph.D. in chemical engineering from
potential for flow problems. We recognize that many other Michigan State University. That summer he transferred to
factors not discussed in this paper might affect the ultimate the Polyurethanes R&D group in Freeport, Texas. He is
strength of the foam, and thus the need for further research. currently working on technology development for rigid
foam applications.
ACKNOWLEDGEMENTS
Ike Latham
The authors would like to especially thank Steve Kahla,
Perry Brooks and Bob Kuklies for their assistance in Ike Latham received his B.S.
making the runs to produce the foams used to validate this degree from Washington College
work. in Chestertown, Maryland and his
Ph.D. in physical chemistry from
REFERENCES Texas A&M University. After
obtaining his degree, he spent two
1. Wood, G. Flexible Polyurethane Foam - years as a postdoctoral fellow at
Chemistry and Technology. Applied Science the University of Southern
Publishers Ltd. (1982) Mississippi. He joined The Dow
2. Bird, B. R., Stewart, W. E. and Lightfoot, E. N. Chemical Company in 1989 as a
Transport Phenomena. John Wiley & Sons, Inc., member of the material science
New York (1960) group in the Polyurethane Product
3. Crank, J., The Mathematics of Diffusion. 2nd Research & Development Laboratory. He is currently
Edition, Oxford University Press, New York, working on technology development for rigid foam
(1975) applications.
4. Ozisik, M. N. Basic Heat Transfer. McGraw-Hill,
Inc, (1977) Linda J. Brown
5. Whitaker, S. Fundamental Principles of Heat
Transfer. Pergamon Press Inc., New York (1977). Linda Brown joined the Dow
6. Flory, P. J. Principles of Polymer Chemistry. Chemical Company in Freeport,
Cornell University Press, 13th Printing, Ithaca Texas, in 1986. She received her
(1986) B.S. in Biological Sciences from
7. Burchard, W. and Ross-Murphy, S. B. Physical the University of Houston, Clear
Networks Polymers and Gels. Elsevier Science Lake. Prior to joining the
Publishers Ltd. New York (1990) Polyurethane Department, she
8. Stepto, R. F. T. Polymer Networks Principles of spent four years in the Health and
their Formation, Structures and Properties. Environmental Sciences
Blackie Academic & Professional, 1st Edition, Department, specifically, in the
New York (1998) toxicology and histology areas. In
9. Dotson, N. A., Galvan, R., Laurence, R. L. and 1991, she began a new assignment
Tirrell, M. Polymerization Process Modeling. in the Polyurethane department, which include the physical
VCH Publishers, Inc. (1996) test laboratory for Polyurethane and Epoxy. In 1995, she
joined the Appliance Group where she is currently a Senior
BIOGRAPHIES Development Chemist in the Polyurethane Products
Research & Development Laboratory
B ernard E. Obi

Bernard E. Obi received his B.S.

degree from SouthWest Missouri
State University in Springfield
Missouri, and an M.S. in physical
chemistry from the University of
Iowa in Iowa City. He went on to
earn an M.S. in chemical
engineering from New Mexico
State University in Las Cruces,
New Mexico in 1985. He joined

226 / Obi, Latham, Brown

 New DMC Catalysts for Manufacturing Polyols

JAMES M. O’CONNOR MARK H. McADON

SynUthane International Inc. The Dow Chemical Company
677 North Mountain Road 1702 Building
Newington, Connecticut 06111 Midland, Michigan 48674

DAVID E. LAYCOCK
Dow Chemical Canada Inc.
1086 Modeland Road
P.O. Box 1012
Sarnia, Ontario N7T 7K7
Canada

ABSTRACT well understood. As an example, several of the ARCO

patents [10] attribute improved catalyst activity to DMC
Double-Metal Cyanide (DMC) compounds are known catalysts being “substantially amorphous”, i.e.,
catalysts for epoxide polymerization. These catalysts are “substantially noncrystalline, lacking a well-defined crystal
highly active and are especially useful in the production of structure, or characterized by the substantial absence of
polyether polyols having low unsaturation. Conventional sharp lines in the XRPD pattern.”
DMC catalysts are prepared by reacting an aqueous Current zinc hexacyanocobaltate-based DMC catalysts
solution of a metal salt and a cyanometallate to form a are prepared by reacting an aqueous solution of zinc salt, a
precipitate of the DMC compound Activity requires use of hexacyanocobaltate, an organic complexing agent (such as
organic complexing agents in the synthesis, such as glyme 1,2 -dimethoxyethane (glyme) or /-butanol), and optionally,
or /-butanol. This paper discusses a new family o f DMC as a secondary complexing agent, a functionalized polymer
catalysts based on bifunctional complexing agents. These such as a poly-(propylene oxide)-based polyol. Essentially
catalysts achieve significant activity from a variety of new all such DMC catalysts produce poly-(propylene oxide)
compositions and crystal structures. These new DMC polyols with low unsaturation (<0.02 meq./g) at almost
catalysts are compared with previous DMC catalysts, “any” temperature, and hence allow the polymerization to
including details of synthetic preparation, elemental reach significantly higher molecular weights in comparison
analysis, x -ray powder diffraction (XRPD) and catalytic with other methods, such as base catalyzed polymerization
activity. withKOH.
This paper discusses a new family of DMC catalysts
INTRODUCTION made with bifunctional polyglycol ether complexing
agents. These complexing agents are based on the formula
Double-metal cyanide catalysts (DMC), discovered RO-(CH2-CHR'-0)x-H, where R is an alkyl group
thirty-eight years ago by researchers at General Tire and containing one to four carbons, R' is H or Me, and where x
Rubber Co. [1-2], are active catalysts for manufacturing takes on values 1, 2, or 3. Some of these ligands are listed
alkylene oxide polyols. DMC technology was revisited in Table 1. These ligands combine the ether functionality
starting in 1983, with improvements made by others - o f the glyme with alcohol functionality, eliminating the
including Shell [3-4], Asahi Glass [5-7], and ARCO [8- need for secondary complexing agents.
11]. The ARCO/Lyondell DMC technology was recently When this work was initiated, based on the previous
purchased and extended by Bayer[12-13]. ARCO work, SynUthane’s plans included evaluating both
Despite these efforts, and despite numerous studies of “amorphous” and “crystalline” catalysts. To that end,
metal cyanide structures [14-15], DMC catalysts, their catalyst controls of both types were prepared based on
reaction mechanisms, and their crystal structures are not procedures from the patent literature. In addition, these

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 227

procedures were applied to catalysts using the new [10]. The XRPD given in these patents [10] actually
polyglycol ether complexing agents. An outside laboratory indicates that the catalysts are crystalline, but not with the
evaluated the XRPD of many new compositions of both familiar cubic structure [15a]. Furthermore, Dow found the
types (new ligands vs. old technology). Surprisingly, it was DMC/glyme catalysts [1-4] to be monoclinic (see Figure
discovered that crystalline catalysts were obtained. Work 1 ).
from Dow then confirmed these outside findings and The new DMC catalysts described herein achieve
further revealed that no amorphous catalysts were significant activity from a variety of new compositions and
prepared. crystal structures. These new catalysts are compared with
On examination of the ARCO patents [10], Dow found previous DMC catalysts, including details of synthetic
that ARCO’s description of the DMC catalyst structure preparation, elemental analysis, XRPD and catalytic
runs counter to XRPD data given in these same patents activity.

Table 1. Selected Polyglycol E ther Complexing Agents, R O -(C H 2 -C H R '-0)x H

Acronym R R* X MW (amu) Chemical Name
MEG Me H 1 76.1 ethylene glycol, monomethyl ether
MDEG Me H 2 120 .1 di-(ethylene glycol), monomethyl ether
MTEG Me H 3 164.2 tri-(ethylene glycol), monomethyl ether
PEG n -Pr H 1 104.1 ethylene glycol, monopropyl ether
BEG n-Bu H 1 118.2 ethylene glycol, monobutyl ether
BDEG n-Bu H 2 162.2 di-(ethylene glycol), monobutyl ether
tBEG t-Bu H 1 118.2 di-(ethylene glycol), mono-t-butyl ether
MPG Me Me 1 90.1 propylene glycol, monomethyl ether
MDPG Me Me 2 148.2 di-(propylene glycol), monomethyl ether
MTPG Me Me 3 206.3 tri-(propylene glycol), monomethyl ether
PPG n -Pr Me 1 118.2 propylene glycol, monopropyl ether
PDPG n-Pr Me 2 176.3 di-(propylene glycol), monopropyl ether
tBPG n-Bu Me 1 132.2 propylene glycol, mono-t-butyl ether

Figure 1. XRPD for DMC/glyme (S60, Shell; see Table 3) and for a partiallyd/dehydrated DMC

BACKGROUND in 1963 by the original inventors, Herold, Milgrom, and

Belner of The General Tire and Rubber Company [1,2].
C atalyst Technology The earliest work seems to concentrate on ferrates, but a
glyme-zinc hexacyanocobaltate complex was identified
The development and evaluation of DMC catalysts for early on in a patent filed in 1965 [2]. Although a broad
the production of polyether polyols has been proceeding range of metals are claimed, zinc cyanocobaltate appears to
for roughly thirty-eight years, beginning with patents filed be the optimum combination for these catalysts, although

228 / O 'Connor, McAdon, haycock

one patent [16] claims a glyme-zinc hexacyanoiridate compositions containing functional polymers and increased
catalyst that provides a lower monol polyol in comparison zinc content.
to a glyme-zinc hexacyanocobaltate catalyst. A SciFinder (Chemical Abstracts) search [17] on die
DMC catalysts were “re-discovered” in 1983 by terms “double metal cyanide” and “polymerization” turned
scientists from Shell [3,4], claiming simpler preparation up 118 patents and ten miscellaneous scientific
methods and improvements in activity, by use of publications. O f the 118 patents, Bayer/ARCO/Lyondell/
cocatalysts such as metal sulfates, sulfuric acid, or Olin head the list with 75, (12+49+9+5, respectively),
hydrochloric acid. These cocatalyst promoters are added in followed by Asahi with 22, General Tire with 13, BASF
addition to the “excess” zinc chloride used in the basic with 7, Shell and East German Academy of Sciences tied,
synthesis, and are reported to result in much more active each with 3, Dow Chemical with 2, OSI Specialties Inc
catalysts with significant rate enhancement. Shell’s interest with one (plus anonymous, with one UK research
in DMC included both polymerization and separations. disclosure). This search no doubt excludes a fair number of
Their studies advanced a cubic structure model where patents, but it does seem to give a reasonably good idea of
molecular sieve pores result from site vacancies, consistent the breadth of the DMC technology.
with their sorption studies related to separating carbon
dioxide from natural gas. In addition, Shell introduced a C atalyst C rystal Structure
new type o f DMC compound, Zn2[Co(CN)6]OH, closely
related to DMC catalysts, but non-cubic. This new DMC The difficulty in characterizing the different structural
compound can be further washed with glyme and HC1 to forms of DMC catalysts was earlier recognized with
produce very active alkoxylation catalysts. catalysts containing glyme as die complexing agent. For
Asahi Glass [5-7] started patenting in 1989, with instance Herold and Livigni [Id] describe DMC/glyme
approximately twenty Japanese patents over a three-year catalysts as “largely amorphous”. On the other hand,
period. Improvements include various catalyst deactivation Kuyper et al. [4] describe DMC/glyme catalysts as
and removal methods [5], mixed-metal-cyanide DMC crystalline by giving “fairly simple line patterns.
preparation [6], and numerous variations on ligand and Depending upon the preparative conditions (particularly
chelant types. Specifically, JP 4,145,123 [7] teaches DMC during the isolation of the solid catalyst), at least three
catalysts coordinated with ferf-butanol, prepared by stirring different, though similar, diffraction patterns are obtained.
an aqueous solution containing zinc chloride, potassium Unfortunately, no assignment of cell parameters was
cyanocobaltate, and die /erf-butanol ligand This catalyst found.” Both of these references foil to provide XRPD
had a substantially improved catalytic life over die data. Kuyper reported that attempts to index the XRPD
DMC/glyme catalyst. were unsuccessful.
ARCO flooded the DMC -polyol patent literature with The structural form of a material can be determined by
broad claims, starting in 1988 [8] and continuing today the nature of the XRPD peaks, including their positions,
through Bayer [12-13]. The current Bayer/ARCO DMC intensities and widths. Widths are determined as full width
catalysts [10-13] favor tert- butanol as the primary at half maximum (FWHM)), and are dependent on the
ligand. AR C O ’s first use of / - butanol came several wavelength (X) of the X -ray source [Cu K a radiation (X =
years after its appearance in the Asahi patent. 1.54 A)]. For perfect crystals, XRPD resolution is limited
A R C O ’s claim ed im provem ents include increased by the diffractometer (perfectly sharp scattering) and
activity, reduced induction periods, plus the capability typically the FWHM < 0.3° 2-theta. For imperfect
to produce polyols with utra-low unsatruation (ultra- (defective) crystals, and also for very small crystals
low monol; unsaturation levels less than 0.007 (domains less than 500 A), the peaks are broadened, with
meq./g). Their teachings on DMC catalyst preparation 0.3 < FWHM < 4° 2-theta. Peak broadening can also be
include variations on mixing (homogenizer, order of due to a variety o f effects, including crystallite size, defects
mixing for zinc and cobalt components, and mixing in the and structural disorder. For liquids and amorphous solids
presence of complexing agents), and variations on ligands (glasses), the XRPD intensity is a continuous, slowly
(replacement o f glyme either by /-butanol alone or by t- varying function with only a few broad maxima, each of
butanol in combination with secondary ligands such as PO which has a FWHM exceeding 4° 2-theta. For these cases,
polyols or other functional polymers). Recently, Bayer [12- the atomic structure exhibits no periodicity or long-range
13] has patented a number of new catalysts based on t- order. For the XRPD pattern for amorphous silica in
butanol and other complexing agents besides the polyols Snyder’s text book [21], the main peak has FWHM >12° 2-
covered by ARCO. theta. Amorphous materials, including amorphous
The ARCO/Bayer DMC catalysts are called either catalysts, will exhibit no more than two or three broad
“substantially amorphous” [10] or “substantially XRPD peaks (“maxima”). In order for an amorphous
crystalline” [11] depending on the characteristics of the structure to exist, the atomic structure would need to
XRPD in comparison to cubic DMC [15]. Bayer [12] (prior exhibit either no periodicity or no long-range order.
to their acquisition of ARCO/Lyondell) extended claims on Some crystalline DMC catalysts, as shown in the
the “subkantially crystalline” DMC to include ARCO patents [10c], have broad peaks (FWHM roughly
2°) at approximately 14 and 18°. In such DMC catalysts,

O'Connor, McAdon, Laycock / 229

these broad peaks are associated with a relatively sharp combination of starting materials, filtration, drying, or by
peak at about 23.6° (see below for a discussion of the U1 simple shelf-aging. XRPD for partially dehydrated DMC
and U1A phases). An amorphous DMC catalyst would is also shown in Figure 1.
have very broad peaks (FWHM >4°) that are not associated In addition to the three DMC crystal phases shown in
with any sharp peaks. These ARCO catalysts may be Figure 1, we define three new DMC crystal phases, U l,
poorly crystalline, but they are not amorphous. U l A, and U2, according to XRPD peaks listed below.
The characterization of the DMC catalyst in US 1. A sharp peak at 23.64 ± 0.32 deg. 20 [d = 3.76 ±
5.712.216 [10c] as “substantially amorphous”, i.e., 0.05 A], with width FWHM < 0.45 deg. 20.
“substantially noncrystalline, lacking a well-defined crystal 2. A broad, asymmetric peak with a maximum at about
structure, cm* characterized by the substantial absence of 14.5 ± 0.6 deg. 20 [d = 6.10 ± 0.27 A], rising sharply at
sharp lines in the XRPD pattern” is thus technically low angle, and falling gradually at high angle, with width
incorrect. Further, the XRPD patterns for three different FWHM =1.0 ± 0.6 deg. 20.
DMC catalysts using /-butanol as complexing agent differ 3. A broad peak with a maximum at about 18.5 ± 0.3
very little. For instance, the XRPD labeled in US deg. 20 [d « 4.79 ± 0.08 A] with width FWHM =1.8 ± 0.8
5.712.216 [10c] as 25% cubic is essentially identical to that deg. 20. This peak can also be asymmetric, in some cases
labeled as “Comparative Example 3 from Asahi” and is rising gradually at low angle and felling sharply at high
very similar to that labeled as substantially crystalline in angle.
US 5,627,122 (Example 1) [11]. The XRPD patterns for 4. A broad peak with a maximum at about 17.0 ± 1.0
these three DMC//-BuOH catalysts are also very similar to deg. 20 [d = 5.21 ± 0.32 A] with width FWHM =1.4 ± 0.6
that shown in WO 99/19063 (Bayer) for a t- deg. 20.
BuOH/polyester DMC catalyst [12]. In fact, all of these 5. A broad peak with a maximum at about 20.1 ± 0.3
DMC catalysts have varying amounts of two crystalline deg. 20 [d = 4.41 ± 0.07 A] with width FWHM =1.2 ± 0.5
phases - cubic and a second phase (either U1 or U2; see deg. 20.
below). The U l and U2 DMC crystalline phases are defined as
Independent of the joint efforts with SynUthane, Dow having a combinations of three peaks (1, 2 & 3) and four
synthesized and took XRPD for several DMC/glyme peaks (1, 2, 4 & 5), respectively, as listed above. Ul
catalysts using both Shell procedures (HC1) [3-4] and contains all three features, and U2 contains all four
General Tire procedures (acid ion exchange) [1-2]. features.
SynUthane also made a DMC/glyme catalyst using a Shell ARCO’s US 5,482,908 [10b] includes a claim based on
procedure, and Dow took the XRPD for that catalyst as essentially three of these features (1, 2, & 3): Claim 11.
well. Surprisingly, all these catalysts had nearly identical “(d-spacing, angstroms): 5.75 (br), 4.82 (br), 3.76” [where
XRPD patterns; i.e., all gave XRPD patterns similar to that br means broad] ARCO’s U.S. 5,470,813 [10b] includes a
shown in Figure 1. In addition, these XRPD patterns are claim based on two of these features (1 and 3): Claim 24
essentially the same as that shown previously for a “(d-spacing, angstroms): 4.82 (br), 3.76 (br).” The XRPD
“conventional DMC catalyst” (see Figure 2 of US. patterns shown in the ARCO patents [10] plus that shown
5.712.216 ) [10c]. Indexing requires that either die sample in the Bayer patent [12] and described in Table 1 of the
is a single crystalline phase, or that overlapping multiphase Bayer patent [12] in feet contain all three features of the
peaks (as shown for other cases in Figure 1) can be U l crystal phase.
separated into peaks for separate crystalline phases. For
selected DMC/glyme catalysts, no substantial impurity DMC//-BuOH/polyester WO 99/19063 [12]
peaks could be identified from multiple samples using 14.5°, d = 6.10 A, FWHM =1.3°
standard techniques (see Cullity, Chapter 9) [21]. As 18.6°, d=4.77 A, FMHM =1.4°
determined by XRPD, the DMC/glyme catalyst of Figure 1 23.7°, d=3.75 A, FWHM =0.4°
does not contains any detectable amounts of the known
cubic DMC (hydrated) or rhombohedral DMC DMC//-BuOH/P4000 US 5,712,216 [10c]
(dehydrated) crystal phases (see below), nor does it contain 15.4°, d=5.75A, FWHM=1.2°
any detectable amount of U l, U2, or U1A. The peaks for 18.5°, d=4.82 A, FMHM =1.7°
this catalyst were indexed [20] to a monoclinic unit cell 23.6°, d=3.76 A, FWHM =0.4°
with cell parameters (a, b, c) = (12.85, 7.61, 9.84 A); P =
107.6 A. Cases that contain the sharp peak (23.6 deg. 20) plus
Hydrated cubic DMC [15] contains no excess Zn nor broad, poorly-resolved features between 13.5 and 22.5 deg.
complexing agent. It has the formula Zhl.5Co(CN)6 *6 20 are called U1A. This U1A phase is also shown in
H20. Partially dehydrated DMC also contains no excess Zn Figure 6 of US 5,712,216 [10c]. U1A is related to U l, and
and consists of a mixture of cubic [15a] and rhombohedral its characteristic XRPD pattern can occur from a merging
[15b] phases. It may easily be made by simple variations of of peaks 2 and 3 due to a variety of factors taken singly or
known procedures for making cubic DMC, such as by in combination, such as a narrowed splitting, increased
asymmetries, and increased FWHM values. Further
examples of U l A plus examples of U2 are given herein.

230 / O ’Connor, McAdon, haycock

EXPERIM ENTAL
Pow der X -R ay D iffraction A nalysis
DM C C atalyst Preparation Studies
Powder x-ray diffraction (XRPD) analysis was carried
There were 7 processes used to prepare the catalysts of out using two methods from different labs. In some cases,
this paper and these are designated in the following samples were analyzed by both methods. In these cases,
manner. The first process (S) was taken from U.S. Patent similar results were obtained.
No. 4,477,589, issued to Shell [5]. This process was only In the first method, experimental conditions were
used to develop comparative activity data and none of the chosen to match those described in Table 1 of US
new SynUthane examples were prepared in this way. The 5,470,813 [10b], except that the counting time was
second process (A) was a procedure taken directly from increased to give higher quality data (reduce noise). Here,
US 5,712,216 (Example 8, ARCO) [10c] that is reported to XRPD was carried out on a Siemens D500 Kristalloflex
produce amorphous catalysts. This gentle mixing process is diffractometer. Samples were analyzed using
carried out in a round bottom flask utilizing a paddle monochromatized C uK al radiation (k= 1.54059 A) at 40
stirrer. In addition to preparing a comparative catalyst from kV and 30 mA. Slit sizes were 1° (for the divergence slit),
f-butanol, catalysts were prepared using this process with 0.05° (for the monochrometer slit) and 0.15° (for the
several of the new complexing agents. The third process detector slit). XRPD data were collected from 10° to 70°
(A2) from another ARCO patent, US 5,482,908 (Example 20 with a step size of 0.02° 20 and a counting time of 3-6
1) [10b] was used to prepare a DMC/f-BuOH/poly-PO seconds per step.
catalyst to compare the activity with the new catalysts of In the second method, XRPD was collected using a
this papa*. In this procedure, we substituted a blender for Siemens D500 Kristalloflex automated powder
mixing instead of the homogenizer. The blender is an diffractometer equipped with a Co X -ray tube source,
Osterizer™ Classic Blender capable of achieving about 17- primary beam monocrometer, and a position sensitive
18,000 rpm with a high shear propeller type agitator. The detector (PSD). The incident beam was collimated using a
fourth process (B) is based on a procedure from WO 1° divergence slit. The tube was operated at 35 kV and 30
99/19063 (Example 3, Bayer) [13] that is reported to mA, and the sample was illuminated with C oK al radiation
produce crystalline catalysts. This process utilizes a high (A, = 1.78897 A). Samples were mounted on zero-
speed dispersator (Premia* Mill Corp.) with a 2 inch background single-crystal quartz sample holders having a
diameter circular dispersion blade for mixing. The cavity depth of 0.5 mm. XRPD data were collected from 5°
dispersator is capable of achieving speeds of up to 15,000 to 90° 20 with a step size of 0.02° 20 at a rate o f 0.5° 20
rpm. Besides producing the catalyst of the example for a per minute. The sensitivity, as measured by the effective
comparative activity run, other catalysts were produced counting time, is roughly 25-50 times greater than that of
utilizing this process with several of the new complexing method one. This increased sensitivity is due primarily to
agents. The remaining three catalyst preparation processes die PSD, which serves as a linear array of simultaneous
(SI, S2 and S3) are hybrids of die above processes. In detectors, thus increasing the counting time for each 20
process SI, the high speed dispersator is used for the Is* step by a factor of 250 (the PSD window is open to about
mixing step and the blender is used for the subsequent 2 5° 20). Data were converted from C oK al to C uK al using
washing steps. In S2, the blender is used for all 3 steps. the Jade software. Computational indexing of the XRPD
And lastly for process S3, the blender is used for all 3 steps pattern was carried out using commercial software [20].
but unlike die S2 process, the complexing agent is present Computational indexing requires that either the sample is a
during the initial precipitation of the DMC. single crystalline phase, a* that overlapping multiphase
peaks can be separated into peaks for separate crystalline
Elem ental A nalysis
phases [20-21]. When converted to C uK al, XRPD data
covered a range of 4.3° to 75.0° 20 with a step size of
Elemental analysis was carried out on several of the
new catalyst compositions. Samples were analyzed in 0.0116° 20.
duplicate or triplicate lots. For most cases, the expected
C atalyst A ctivity Determ inations
relative error for Zn, Co, Cl, C, H, N is ± 3% [as an
example carbon 38.6 ± 1.2 wt %]. Potassium occurs at
Examples 1-37 are catalyst activity runs, as listed in
relatively low levels (generally less than 0.60 ± 0.06 wt Table 3, using catalysts prepared with the new proprietary
%). Oxygen was obtained by subtraction. Formulas were complexing agents as well as runs with the reportedly
assigned assuming a mole fraction of CN/Co = 6.0 active catalysts from the literature. Each of the catalysts
(assumed no additional CN or HCN was present). prepared above was used to polymerize propylene oxide
Elemental analysis for K, Zn, Co, and Cl was obtained by (PO) using 2,4-diethyl-1,5-pentanediol as starter [18]. The
neutron activation analysis, as described in Rigot [19]. feed charge was comprised of 2,4-diethyl-1,5-pentanediol,
Elemental analysis for C, H and N was performed by freshly distilled propylene oxide and DMC catalyst, (either
combustion analysis, on a LECO CHNS932 analyzer. The 50 or 25 PPM based on a final molecular weight of 4,000).
results for the selected cases are given in Table 2.

O ’Connor, McAdon, haycock / 231

The PO/starter mole ratio in the charge was 3 unless initial exotherm was recorded (t-exo). Manual cooling was
otherwise specified. The charge was introduced into a 300 applied to try to maintain the temperature at about 120°C.
ml Parr™ pressure reactor at room temperature. The The maximum temperature achieved ( T-max) and the time
reactor was flushed three times with nitrogen at 30 psi and from t-start to the point when all of the PO was consumed
the reactor was then heated to 120°C. The time to reach were also recorded (t-end).
100°C was used as t-start and the time from t-start to the

Table 2. M olecular Formulas and Elem ental Analysis (wt. %) of Selected DM C Catalysts
Compound K Zn Co Cl N c H O Ligand
DM C/M DEG (B67, process S2)
Zn2.4 C I1.0 (O H )0.8 Co(CN)6 1.4 MDEG •0.8 H 20 ' 0.03 KCI
analysis 0.2% 25.6% 9.6% 6.0% 14.0% 25.8% 3.1% 15.7% 27.6%
formula 0.2% 25.9% 9.7% 6.0% 13.9% 25.8% 3.2% 15.3% 27.8%
difference 0.0% -0.3% -0.1% 0.0% 0.1% 0.0% -0.1% 0.4% -0.2%
DM C/M TEG (A74, process A)
Zn2.1 CIO.55 (O H )0.65 C o(CN)6 -2.6 MTEG 0.2 H 2 0 0.012 KCI
analysis 0.1% 16.4% 7.1% 2.4% 10.8% 35.8% 5.2% 22.3% 51.9%
formula 0.1% 16.9% 7.2% 2.4% 10.3% 35.7% 5.3% 22.1% 52.4%
difference 0.0% -0.5% -0.2% -0.1% 0.4% 0.1% -0.1% 0.2% -0.5%
DMC/BEG (B27, process S1)
Zn2.5 C I1.15 (O H )0.85 Co(CN)6 1.8 BEG •1.1 H 2 0 0.02 KCI
analysis 0.1% 24.5% 8.8% 6.1% 12.6% 30.2% 4.2% 13.5% 31.7%
formula 0.1% 24.5% 8.8% 6.2% 12.6% 30.2% 4.3% 13.3% 31.9%
difference 0.0% 0.0% 0.0% -0.1% 0.1% 0.0% -0.1% 0.2% -0.1%
DM C/M PG (B42, process S2)
Zn2.18 C I0.45 (O H)0.91 Co(CN)6 -1.2 MPG 0.5 H 2 0 0.02 KCI
analysis 0.1% 27.4% 11.4% 3.3% 17.0% 25.5% 2.6% 12.6% 20.5%
formula 0.2% 28.1% 11.6% 3.3% 16.6% 25.6% 2.8% 12.0% 21.3%
difference 0.0% -0.7% -0.2% 0.0% 0.5% 0.0% -0.2% 0.6% -0.8%
DM C/M DPG (B63, process S2)
Zn2.4 CM .15 (O H )0.65 Co(CN)6 1.6 MDPG -1.1 H 2 0 0.03 KCI
analysis 0.2% 22.4% 8.5% 6.0% 12.7% 30.5% 4.1% 15.6% 34.6%
formula 0.2% 23.0% 8.6% 6.1% 12.3% 30.2% 4.2% 15.3% 34.7%
difference 0.0% -0.6% -0.1% -0.1% 0.4% 0.3% -0.1% 0.3% -0.2%
DM C/M TPG (A87, process A)
Zn2.3 C I1.0 (O H )0.6 Co(CN)6 -1.1 MTPG 0 .7 H 2 0 0.03 KCI
analysis 0.2% 22.6% 8.8% 5.4% 13.9% 30.9% 4.0% 14.2% 32.6%
formula 0.2% 23.0% 9.0% 5.6% 12.9% 31.3% 4.0% 14.0% 34.8%
difference 0.0% -0.4% -0.2% -0.2% 1.0% -0.4% 0.0% 0.3% -2.2%
DM C/M TPG (A171, process S2)
Zn2.4 C I0.95 (O H )0.85 C o(CN)6 -1.95 MTPG -1.0 H 2 0 0.05 KCI
analysis 0.2% 18.2% 6.9% 4.2% 10.0% 36.0% 5.3% 19.2% 47.1%
formula 0.2% 18.6% 7.0% 4.2% 10.0% 36.3% 5.5% 18.3% 47.7%
difference 0.0% -0.4% -0.1% 0.0% 0.0% -0.3% -0.2% 0.9% -0.6%

RESULTS AND DISCUSSION seven synthetic processes for preparing the catalysts gives
rise to nearly three hundred possible catalysts.
Table 1 lists thirteen polyglycol ethers that were selected In this work, 25 catalysts were prepared using the new
for preparing catalysts, along with acronyms used to polyglycol ether complexing agents. Table 2 gives
facilitate the identification of the specific complexing approximate molecular formulas for seven of these
agents. Unlike other complexing agents, these polyglycol catalysts, based on elemental analysis. For these seven
ethers are multifunctional, that is, they all contain one cases, Zn/Co mole fractions range from 2.1 to 2.5, and
alcohol moiety and at least one ether linkage to complex ligand content ranges from 20 to 50 wt %.
with the DMC compound. This dual functionality allows
active catalysts to be produced without the need for any C atalyst A ctivity
polyol secondary complexing agents. The general scheme
gives rise to 42 possible complexing agents, resulting from Table 3 lists activity data for these catalysts plus, for
combinations of the seven alkyl groups possible for the comparison, four catalysts that were prepared using
ether with the six different oligomers (die mono, di and tri previous technology. The activity screening procedure is
ethylene and propylene glycol). This in combination with described in an earlier section of this paper (experimental).

232 / O'Connor, McAdon, Laycock

The reactor was charged with catalyst, starter, and was observed is listed as t-exo. In many cases this is easily
propylene oxide. Upon heating, the time required to determined, but for the less active catalysts there was a
reached 100°C was recorded as t-start. Subsequent events more gradual rise in temperature with a commensurate
are recorded relative to t-start. Heating continued until the decrease in pressure. Most of the time, the exotherm could
reactor reached 120°C. The time when the initial exotherm be reasonably controlled by manual

Table 3. Summary o f Catalyst Reactivity Screening Tests (Times Are in Minutes)

Run # C a t# Prep Complexing Agent t start te x o te n d T max (°C)
50 PPM Catalyst (Based on 2000 Eq. W t.) PO / Starter = 3 (2,5 -diethyl-1,5 pentanediol)
1 S60 S glyme 18 2 6 185
2 A51 A f-BuOH 20 14 83 135
3 A107 A2 f -BuO H/PPG 4000 21 13 30 172
4* A93 B f -BuOH/PEG2000 29 12 71 130
5 B171 A MEG 23 18 >120 120
6 D15 S3 MEG 18 2 4 212
7 B67 S2 MDEG 21 3 8 157
8 A74 A MTEG 24 31 103 131
9 B4 S1 MTEG 21 9 33 129
10 B29 S1 PEG 23 7 >120 120
11 B27 S1 BEG 22 3 10 133
12 B61 S2 BEG 21 2 6 134
13 B16 S1 BDEG 23 10 49 123
14 D22 S2 BDEG 21 5 21 135
15 B84 S2 tBEG 20 2 5 184
16* A97 A MPG 30 9 98 127
17 B25 S1 MPG 31 3 13 133
18 B42 S2 MPG 20 4 8 137
19 A78 A MDPG 21 5 30 128
20 A98 B MDPG 21 4 14 129
21 B63 S2 MDPG 18 1 4 196
22 D12 S3 MDPG 24 1 4 196
23 A87 A MTPG 24 3 18 131
24 A86 B MTPG 24 7 67 120
25 A168 S1 MTPG 18 7 47 123
26 A171 S2 MTPG 18 3 7 143
27 D26 S2 PPG 19 3 10 135
28 B40 S1 PDPG 21 12 >120 120
29 B23 S1 tBPG 23 10 >180 120
25 PPM Catalyst (Based on 2000 Eq. W t.) PO / Starter = 3 (2,5 -diethyl -1,5 pentanediol)
30 S60 s glyme 18 6 47 130
31 D15 S3 MEG 18 4 18 137
32 B67 S2 MDEG 20 6 18 130
33 B61 S2 BEG 20 5 26 132
34 B84 S2 tBEG 22 5 21 131
35 B42 S2 MPG 20 6 13 134
36 B63 S2 MDPG 19 2 14 130
37 A171 S2 MTPG 22 6 43 131
* = PO / starter ratio = 5 rather than 3
t start is the tim e to reach 100 C. t exo and t end are measured relative to t start
t exo is the tim e to reach the exotherm
t end is the tim e to reach completion (PO fully reacted)

cooling (120 < T-max < 140 °C, where T-max, as listed in catalyst. Other considerations are obviously just as
Table 2, is the maximum temperature achieved during important, including catalyst lifetime, and properties o f the
reaction). Finally, the time for complete PO consumption polyol produced, such as the monol content, the high
was recorded as t-end. In a few instances, for some of the molecular weight tail, etc.
very active catalysts, efforts to control exotherm failed. In At the 50 PPM catalyst level, catalyst activity can be
most of those cases, the catalyst was re -run at 25 PPM to characterized as low, medium, or high, according to time
get a better determination of activity. required for exotherms and completion (in minutes), as
This test was established as a screening tool and is not follows.
the only criterion involved in defining an acceptable

O ’Connor, McAdon, Laycock / 233

low 7 < exotherm 40 < completion catalysts. For cases where two processes gave catalysts
medium 4 < exotherm < 7 20 < completion < 40 with high activity (e.g., BEG, MPG, MDPG, and MTPG),
high exotherm < 4 completion < 20 the highest activity is obtained with either the S2 or S3
process. Cases that yield low activity with process A, B or
These ranges cover all catalysts except for Run 9 SI are expected to benefit with the all blender processes S2
(catalyst B4), where the activity is low relative to or S3. Examples of this include DMC/MEG and
exotherm, but medium relative to completion. Based on DMC/MTPG (see Figure 2). Later on in this discussion,
completion time, Run 9 is classified as medium. O f the 25 we show that die synthetic process also affects the crystal
new catalysts, nine had low activity, three had medium structure of the catalyst.
activity and 13 had high activity. Four catalysts made using previous technology are
included in Figures 2-3 and Table 3 for comparison. O f
Using these reactivity classifications, catalyst these four, the most active were the glyme catalyst (Shell,
reactivities are plotted in Figure 2. Figure 2 shows that high activity) and the f-BuOH/PPG catalyst (ARCO). The
catalyst activity depends on both the ligand and on the /-BuOH/PEG (Bayer) and the f-BuOH-only catalysts
method of preparation. Furthermore, high activity is almost (ARCO) both exhibited low activity. Using 50 PPM of
always obtained using the S2 and S3 methods. Most, but catalyst, the times of the initial exotherm ranged from 1 to
not all of the more active catalysts are produced when the 14 minutes. The catalysts featured in this paper tend to
blender is used in all steps of the preparation (S2 or S3). give a faster initial exotherm than the catalysts prepared
Figure 3 shows completion times for the more active according to previous technology (the glyme catalyst from
Shell was the only exception.)

Figure 2. Activity o f DMC catalysts at 50 ppm catalyst loading (see Table 1)

Figure 3. Activity of DMC catalysts at 50 ppm and 25 ppm catalyst loading (see Table 1).

234 / O ’Connor, McAdon, Laycock

 Figure 4. XRPD for selected DMC/MDPG and DMC/MTPG catalysts

Eight o f the most active catalysts were run at 25 PPM. final polyol unsaturation, high molecular weight tail, and
Four still exhibited high activity (MEG, MDEG, MPG, and catalyst life, etc. need to be considered along with the high
MDPG), two others exhibited medium activity (BEG and activity. This will be the subject o f future work.
tBEG), and two gave low activity (MTPG and glyme).
Based on the activity screening tests, it appears that C atalyst C rystal Structure
many of the new DMC catalysts ligands are quite active.
However, it is not possible at this time to say which of the Figure 4 shows XRPD for seven of the new DMC
new complexing agents are best. Other criteria such as catalysts, including DMC/MDPG prepared using processes

O ’Connor, McAdon, Laycock / 235

A, B, and S2, and DMC/MTPG prepared using processes catalysts that were either described incorrectly or poorly
A, B, SI and S2. The XRPD peak assignments include understood.
DMC crystal phases U1A, U2, cubic, a new monoclinic
phase, plus Zh5(0H)8C12H20 [22-23] as an impurity, as REFERENCES
indicated in Figure 4. For mixed-phase cases, relative
proportions of the various phases are related to integrated 1. (a) Herold, R. J. and R. A. Livigni. 1972.
peak areas. As an example, the amount of “Hexacyanometallate Salt Complexes for Epoxide
Zn5(0H)8C12*H20, which is believed to be catalytically Polymerization,” Polym. Prepr., Amer. Chem. Soc.,
inert, appears to occur at less than 1 wt % in all cases Div. Polym. Chem., 13(l):545 -550. (b) Herold, R J.
except for MDPG-S2, where it is likely present at less than and R. A. Livigni. 1973. “Hexacyanometalate Salt
5 wt %. The catalyst activities are also indicated on Figure Complexes as Catalysts for Epoxide Polymerizations,”
4. Advan. Chon. Ser., 128:208-29. (c) Herold, R.J.
Four of these catalysts have very similar XRPD 1974. “Amorphous, High Molecular Weight
patterns, with roughly a 75/25 wt % UlA/cubic split: Polypropylene oxide),” Macromolecular Synthesis,
MDPG-A, MDPG -B, MTPG-B, and MTPG-S1; the first 5:9-13. (d) Livigni, R. A., R. J. Herold, O. C. Elmer,
two are medium and high activity; whereas the latter two and S. L. Aggarwal. 1975. “Poly(Propylene Ether)
are low activity. MTPG-S2 has high activity with these Polyols Prepared with a Zinc Hexacyanocobaltate
same two crystal phases in a 95/5 wt % split. MDPG-S2 Complex Catalyst,” ACS Symp. Ser., 6:20-37.
also has high activity, with roughly a 95/5 wt % split 2. Milgrom, J. 1968. “Production of Polyether Diols
between U2 and the Zn5(OH)8C12*H20 impurity. Finally, Using Water as a Telogen,” US 3,404,109, assigned to
MTPG -A gives a high activity catalyst that is a new The General Tire and Rubber Company.
monoclinic phase, with no detectable amounts of other 3. van der Hulst, H., G. A. Pogany, and J. Kuyper. 1984.
crystal phases. Indexing [20] gave a monoclinic crystal “Catalysts for the Polymerization of Epoxides and
with cell parameters (a,b,c) = (12.96, 7.50, 14.87 A); Process for the Preparation of Such Catalysts,” US
P= 107.5°. The structural phase differences between these 4,477,589, assigned to Shell Oil Company.
two DMC/MTPG catalysts (A vs. S2) is believed to result 4. Kuyper, J., and G. Boxhoom. 1987.
from differences in the catalyst preparation and not directly “Hexacyanometallate Salts Used as Alkene-Oxide
from differences in ligand content (elemental analysis is Polymerization Catalysts and Molecular Sieves.” J.
listed in Table 2). Catalysis 105:163-74.
XRPD was carried out on a number of additional 5. (a) Takeyasu, H., T. Watabe, T Doi, Takao. 1990.
catalysts listed in Table 3. This data is not included here; “Polymerization Catalyst Deactivation and Removal
however, all samples that were examined were crystalline. in Polyoxyalkylene Manufacture,” JP 2,265,921
As an example, two of the high activity catalysts prepared (Asahi Glass Co., Ltd., Japan). See CAN 114:123338.
using the S2 method, MPG and BEG, had a U2 crystal (b) Watabe, T., H. Takeyasu, T. Doi, and N. Kunii.
phase similar to that shown in Figure 4 for MDPG-S2. 1990. “Polymerization Catalyst Removal from
Many had patterns similar to those already discussed, but Polyoxyalkylene Polyols,” Eur. Pat. Appl. EP
some appeared to contain new crystal phases. 383,333 (Asahi Glass Co., Ltd., Japan). See CAN
114:7776
CONCLUSIONS 6. Morimoto, T. and N. Yoshida. 1991. “Manufacture
of Polyether-Polyols by Using Heteronuclear Cyano
1. A new family of very high activity DMC catalysts Complex Catalysts,” JP 3,000,733 (Asahi Glass Co.,
based on polyglycol ether complexing agents has been Ltd., Japan). See CAN 115:93171
discovered. (A patent application has been filed.) 7. Yamada, K., H. Takeyasu. 1992. “Preparation of
2. These catalysts are highly active and don’t require Polyether Compounds by Using Double Metal
polyether or other polymeric modifiers to achieve the high Cyanide Complex Catalysts,” JP 4,145,123 (Asahi
activity. Glass Co.). See CAN 117:192560.
3. The activity of these catalysts is not defined only by the 8. S. D. Harper. 1988. “Preparation of Filterable
specific complexing agent or the crystalline structure but Double Metal Cyanide Complex Catalysts for
rather by some combination of the two. Propylene Oxide Polymerization,” EP 283148
4. Very active catalysts with different crystalline forms (ARCO Chemical Co., USA). See: CAN 110:39529
can be produced with the same complexing agent. 9. Hinney, H. R. and D. S. Wardius. 1992. “Process for
5. Reaction conditions during catalyst preparation can Preparing Metal Cyanide Complex Catalyst,” US
affect both activity and crystal structure. 5,158,922 (ARCO Chemical Technology, L.P., USA).
6. Based on our work, and contrary to the existing patent See CAN 118:60285
literature [10], it doesn’t appear that any amorphous DMC 10. ARCO “substantially amorphous:” (a) Le-Khac, B.
catalysts have ever been produced. The catalysts reported 1995. “Double Metal Cyanide Complex Catalysts,”
as “amorphous” [la, 10] are most likely, crystalline US 5.470,813 (ARCO Chemical Technology, L.P.,
USA), (b) Le-Khac, B. 1996. “Highly Active

236 / O'Connor, McAdon, haycock

 Double Metal Cyanide Catalysts,” US 5,482,908 are in good agreement with that reported for
(ARCO Chemical Technology, L.P., USA), (c) Le- simonkolleite: (a,c) = (6.334, 23.58 A) and (6.34,
Khac, B., P. T. Bowman, and H. R. Hinney. 1998. 23.64 A). The assignment is confirmed by matching
“Highly Active Double Metal Cyanide Complex file intensity pattern. See reference 23:
Catalysts,” US 5,712,216 (ARCO Chemical 23. (a) Hesse, F. R. G., K. Schmetzer, G. Schnorrer-
Technology, L.P., USA). Koehler, and O. Medenbach. 1985. “Wulfingite,
11. ARCO “substantially crystalline:” Le-Khac, B., H. R. epsilon-Zn(OH)2, and Simonkolleite,
Hinney, and P. T. Bowman. 1997. “Highly Active Zn5(OH)8C12 □H 20, Two New Minerals from
Double Metal Cyanide Complex Catalysts,” US Richelsdorf,” Neues Jahrb. Mineral., Monatsh. 4:145-
5,627,122 (ARCO Chemical Technology, L.P., USA). 54. (b) Allmann, R. 1968. “Refinement of the
12. Bayer “substantially crystalline:” Hofinann, J., P. Structure of Zinc Hydroxide Chloride,
Gupta, R. Kumpf, P. Ooms, and W. Schaefer. 1999. Zn5(OH)8C12* 1H20,” Z. Kristallogr. 126(5-6):417-
“Improved Double-Metal Cyanide Catalysts for the 26.
Manufacture of Polyether Polyols,” WO 99/19063
(Bayer A.-G., Germany). See CAN 130:312232.
BIOGRAPHIES
13. Ooms, P., J. Hofinann, P. Gupta, and L. Groenendaal.
2000. “Bimetal Cyanide-Based Catalysts Used for
Jam es M. O ’Connor
Preparing Polyether Polyols,” US 6,204,357 (Bayer
James M. O’Connor is one of the cofounders of
A.-G., Germany). SynUthane International Inc. (Incorporated September
14. Dunbar, K. R. and R. A. Heintz. 1997. “Chemistry of
1998). Prior to SynUthane, Jim managed research and
Transition Metal Cyanide Compounds: Modem
development in the polyurethane and chemical industry for
Perspectives,” Prog. Inorg. Chem. 45:283-391.
nearly 30 years. Jim’s experience and research background
15. (a) Mullica, D. F., W. O. Milligan, G. W. Beall, and covers all segments of the polyurethane industry from
W. L. Reeves. 1978. “Crystal Structure of Trizinc
polyols to isocyanates; TPU’s to coatings as well as
Hexacyanocobaltate(III) Dodecahydrate,” Acta surfactants, polymers and other specialty chemicals. He
Crystallogr. B34(12):3558-61. (b) Siebert, H. and W.
holds a B.S. Chemistry from Fairfield University and a
Jentsch. 1981. “Rhombohedrally Crystallizing Zinc
Ph.D. in Organic Chemistry from Purdue University. He
Hexacyanometalates(III) Zn3[M(CN)6]2,” Z.
started his career at Goodyear Tire and Rubber company,
Naturforsch., B: Anorg. Chem., Org. Chem.
then to NDM Corp., Olin Corporation, Olin Hunt
36B(1): 123-4.
(Electronics), and ARCO Chemical Company. He has
16 Behrendt, G.; H. U. Schimpfle, G. Wagner, and H.
authored ova* 60 U.S. patents and publications.
Becker. 1981. “High-Molecular-Weight Polyether
Alcohols.” Ger. (East) DD 148,957 (East German
David E. Laycock
Academy of Sciences; see: CAN 96:53198). David Laycock is a Research Associate with Dow
17. SciFinder Version 2000. See
Chemical Canada Inc. in Sarnia, Canada. He has been with
http://www.cas.org/SCIFINDER/scicover2.html. The
Dow since 1980 engaged in research projects in oxide
search was carreid out on 18 May 2001. polymers, epoxy coatings, foams and RIM technology,
18. James M.O’Connor and Donald L. Lickei, Polyols
principally in the areas of catalysis and material science.
Derived from Double Metal Cyanide Catalysts, Utech
He holds a B.Sc. and M.Sc. in organic chemistry from
2000 Proceedings).
Carleton University (Ottawa) and a Ph.D. in
19. Rigot, W. L., M. E. Buchmann, T. J. Quinn, J. D.
organometallic chemistry from Queen’s University at
Romick, and S. O. Yusuf. 2000. “The Role of
Kingston.
Neutron Activation Analysis at The Dow Chemical
Company,” J. Radioanal. Nucl. Chem. 243(1):205-
M ark H. M cAdon
208. Mark McAdon works on a variety o f problems
20. MDI Jade software for pattern analysis and phase
dealing with catalysts and polymers. He has been with
dentification, see: http://www.materialsdata.com/
Dow Corporate R&D in Midland, Michigan since 1988.
products.htm
Mark has a B.A. in chemistry from The State University of
21. (a) Snyder and Jenkins “Introduction to X -ray Powder
New York (at Binghamton) and a Ph.D. in chemical
Diffractometry”, Wiley-Interscience, 1996, p 25. (b)
physics from The California Institute of Technology.
Cullity, B. D. and S. R. Stock. 2001. Elements of X -
ray Diffraction, 3rd Ed., (Prentice Hall, Upper Saddle
River, New Jersey)].
22. The phase assignment was made using the database
from the International Centre for Diffraction Data
(ICDD; see Cullity, Chapter 9). The impurity can be
indexed to a hexagonal unit cell with approximate
parameters (a,c) = (6.325,23.60 A). These parameters

O'Connor, McAdon, Laycock / 237

Nano- and Micro-Fillers for Polyurethane Foams: Effect on
Density and Mechanical Properties

B. KRISHNAMURTHI, S . BHARADWAJ -SOMASKANDAN a n d R SHUTOV

Tennessee Technological University
Department o f Chemical Engineering
Center for Manufacturing Research
Cookeville, Tennessee 38505

other hand, increasing the filler content might be effective

ABSTRACT
to enhance the modulus and strength, but also increases
the density of foam. There are some limitations in
A comparison analysis have been undertaken to study
polyurethane foam processing using high percentage of
the effect of different commercial fillers on density and
filler content: agglomeration and/or settling, large surface
mechanical properties of two different types of area of filler, complete wetting of particles sometimes not
commercial polyether-based polyurethane (PUR) foams: possible, high viscosity of polyol/iso components, etc.
low-density flexible foams and high-density rigid foams. Several concepts have been developed to find the
Two types of fillers have been used: micro-fillers based compromise, such as [1-3]:
on wood flour (wood waste of the furniture industry) -Foams with non-uniformal density (structural or integral
particles of size 70-250 microns and nano-fillers based on foams),
mineral type which exists in form of clusters of average - Foams with microcellular structure,
size 8 to 10 microns; the thickness of individual clay flake -Reinforcement of foams with fiber or particulate fillers.
particle is 0.3-0.5 nanometer. The relationship between This study deals with the latter concept. The goal of our
the filler size and content, density and various mechanical research is to present a comprehensive study and critical
properties has been studied. It was found that micro- and analysis of the modern technical developments in the area
nano-fillers provide various effects depending on of reinforcement of PUR foams with different size of
flexibility/rigidity and low/high-density nature of PUR fillers having also different chemical nature. There are
foams. For low -density flexible PUR foams, the density many discrepancies and deficit of experimental data in
and mechanical properties increased for both types of understanding the mechanism of the reinforcement effect
fillers. However, for high-density rigid PUR foams these in foamed polymers [4-6].
effects is just the opposite: the density and compression Taking into account these circumstances, we have
strengths decreased for both types of fillers. However, decided to make a comprehensive analysis of the filler
nano-fillers decrease less in compression strength when size effect on the mechanical and chemical properties of
compared to micro-fillers. In general, the relative PUR foams using the commercial PUR systems. The
mechanical property (property/density ratio) is higher for main idea of our research was to screen out a very broad
nano-size filler than micro-size filler. The comparison of range of size of the fillers from nano-size (10 ~9 microns)
the reinforcement effects of micro- and nano-fillers has to macro-size (10 "3 microns). The presented data deal
been discussed. with nano-and micro-fillers only. More data including
macro-fillers will be presented elsewhere.
INTRODUCTION
M ATERIALS AND TESTING
It is well known that flexible and rigid polyurethane
(PUR) foams are widely used in many applications. One C hem icals
of the fundamental problems of any type of polymer
foam, including PUR foams, is how to find an optimal For low-density flexible PUR foam: polyol mixture
compromise between the low -density, high strength and (water, amine catalyst and silicone surfactant),
low cost. Filling of PUR foams with inert and inexpensive diphenylmethane diisocyanate (MDI) solution
fillers do not much improve the mechanical properties, manufactured by Insta-Foam Products, Inc, USA; Ratio
but may be effective way to decrease the cost. On the A/B: 55/45, Density 0.020 to 0.022 gm/cc. For high-

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 239

density rigid PUR foam: system was provided by Demilec RESULTS AND DISCUSSIONS
(USA) LLC, and named SR 5925. Component A was a
modified polymeric diphenylmethane diisocyanate and Flexible Foams
component B was a formulated polyol system that
employed water as the sole blowing agent. When the two Figure 1 demonstrates that increasing in filler content
components were reacted together at 50 parts isocyanate increases the density. For micro-sized fillers, the density
to 50 parts polyol without any filler we got 218 kg/m3 increases by nearly two times at 60 percent of filler
foam. content for wood flour of size 70 microns. For wood flour
of size 250 microns, there is not much change in density
Micro - fillers with increase in filler content. For micro-sized fillers, we
can conclude that smaller the size of the filler higher the
density obtained. Nano-sized filler shows an increase in
Micro-filler wood flour: wood flour manufactured by
American Wood Fibers Corp. Wood flour was sieved and density with increase in filler content as well. It reaches a
two average sizes of 70 and 250 Jim are used as filler. maximum density at 10 percent of filler content with an
increase in density by nearly 1.5 times of that of unfilled
foam. For nano-sized filler 5 percent and 15 percent of
Nano -fillers filler, content has nearly the same density as 10 percent of
filler content.
Mineral clay flake type which exists in form of clusters One can see from Figure 2, that both micro-sized and
of average size 8 to 10 microns, supplied by Nanocor, Inc. nano-sized filler show an increase in modulus of elasticity
Thickness of individual particle is 0.3-0.5 nanometer. with filler content. For micro-sized wood flour of size 250
microns, modulus of elasticity has a maximum at 60
Tests percent of filler content and increases by nearly 50
percent to that of unfilled foam. Wood flour of size 70
Apparent density: an electronic densitometer ED -120T microns has a maximum at 60 percent of filler content and
was used. Sample size large enough to be sunk by the increases by nearly 45 percent to that of unfilled foam.
cage of the equipment was used. For each composition 5 The size of the filler does not influence the modulus of
samples were tested; the average of the densities were elasticity much for micro-sized fillers.
reported. For nano-sized filler at 10 percent and 15 percent of
Compression force deflection test and modulus of filler content has nearly the same modulus of elasticity
elasticity for flexible foams: a test specimen in the size of with nearly 50 percent increase to that of unfilled foam. It
3x3x3 inches was used. An Instron Table Model TM is interesting to note that nano-sized filler at 10 to 15
instruments, having 100-kg capacity with a 10.2-cm percent of filler content has nearly the same modulus of
indentor and a perforated plate was used. The foam was elasticity as that of micro -sized filler 40 percent filler
preflexed twice to 75 percent at a speed of 25.4 cm/min content.
and allowed to rest for 6 minutes. The foam was then Figure 3, shows that wood flour of size 250 microns
compressed to fifty percent of its original thickness at 5.1 provides more compressive strength than wood flour of
cm/min. The indentor was then stopped and a force was size 70 microns at the same filler contents. At 60 filler
recorded after sixty seconds. Modulus of elasticity was content percent for 250 microns size there is an increase
calculated for the PUR foams from the graph for 50 in compressive strength by nearly 2.5 times while for 70
percent compression force deflection test [7]. microns there is a decrease by 10 percent. For micro-sized
Compression test for rigid PUR foams: compressive filler, the lesser the size of the filler the lower the
properties of rigid foams both unfilled and filled were mechanical strength at all filler content. This might be due
tested according to ASTM D 1621-94. Test specimens of to the alignment of the filler in the cell struts, walls and is
cross section area 25.8 cm2 and height 2.54 cm were used also dependent on the property of the material used for
for compression test. Load was applied to the specimen reinforcement. Large sized reinforcement decreases the
with the crosshead movement of 2.5 mm/min for each size of the cell thus requiring more force to compress
25.4 mm of specimen thickness. Stress was calculated at them. For nano-sized filler at 10 and 15 percent of filler
10 percent deformation of the sample. For every sample content the compression force deflection value does not
three specimens were used for testing. Compressive change much. It shows an increase in strength by nearly 2
strength was calculated by dividing the load by the initial times to that of unfilled foam at 10 and 15 percent of filler
horizontal cross-sectional area of the specimen. Density content.
of foams varies from filler to filler so compression
Rigid Foams
strength values were “normalized” before the comparative
study [8].
Figures 4 and 5, demonstrate that with increasing the
filler content the density of rigid foam decreases. In the
case of wood flour a 20 percent content decreased the

240 / Krishnamurthi, Bharadwaj-Somaskandan, Shutov

density by 3.5 times, whereas 80 percent content filler at much lower filler content offers the same modulus
decreased the density by 4 times. It seems that increasing of elasticity as that of micro-sized filler at a higher filler
micro-filler content from 20 percent to 80 percent does content.
not change much the density. In the case of nano-filler the The size of the micro-sized filler affects the
density decreased by 2.5 times at 5 percent filler content, compression force deflection values for flexible PUR
whereas 10 percent filler content decreased the density by foams: the larger the size of the filler the higher is the
about 3.5 times. compression force deflection values while lower the size
Probably, this effect may be explained by the fact that of the filler decreases the compression force deflection
the moisture content in wood flour is about 10-15 percent values.
and in nano-filler it is about 3 percent. When fillers are Nano-sized fillers exhibit very positive effect on
used, the mixing time of the two components of the rigid relative compressive strength for high-density rigid PUR
PUR system and the filler increases. This means that more foams. Relative compressive strength decreases up to 1.5
amount of ambient moisture absorbed by the system times when nano-filler content is only 5percent, further
during the foaming. The ambient moisture in addition to increasing of nano-filler up to 10 percent does not change
the moisture present in the filler increases the foaming much the relative compressive strength. High-density
activity. Additional water reacts with the isocyanate to rigid PUR foams with nano-size clay particles as filler has
liberate more carbon dioxide gas. Figures 4 and 5 very low-density and high relative compressive strength.
demonstrate that in the case of nano-filler even at a low Micro-filler exhibit significant effect on relative
loading level of 10 percent the density reduction is almost compressive strength for high-density rigid PUR foams:
that achieved by 80 percent micro-filler. when the micro-filler content increases from 0 to 20
From figures 6 and 7, it can be noted that relative percent for the particle size greater than 212 microns,
compression strength values decrease with increase in relative compressive strength decreases up to 3 times,
filler content. This confirms that properties of rigid PUR further increasing of filler from 20-80 percent decreases
foam largely depends on the density of the foam. The the relative compressive strength up to 10 times.
mechanism of reinforcement in the case of rigid PUR ACKNOW LEDGEMENT
foam filled with nano-filler is proposed based on the
homogeneous dispersal at the microscopic level of the Authors acknowledge the following persons and
nano-filler, which results in improved and unexpected companies who donated chemicals for the research:
properties of very thin PUR films and struts inside the Mr. Dave Lall, Demilec (USA), LLC;
cellular structure of the foam. Dr. R. Brown, Insta-Foam Products, Inc;
Mr. Mike Pisczor, American Wood Fibers Corp.;
CONCLUSIONS Mr. Tie Lan, Nanocor Inc.
REFERENCES
Nano-size fillers exhibit very different effects on the
density for both low-density flexible PUR foams and 1. F.Shutov, 1991, Blowing Agents, in Polymeric
high-density rigid PUR foams: for flexible foams nano- Foams, Hanser Publishers.
filler increases the density up to 50 percent at the filler 2. F.Shutov, 1991, Cellular Structure and Properties of
content 10-15 percent; for rigid foams it was noticed just Foamed Polymers, in Polymer Foams, Hanser
an opposite effect, density decreases up to 3 times when Publishers.
the nano-filler content is 5 percent only, further increasing 3. F.Shutov, 1986, Integral/Structural Polymer Foams,
of nano-filler up to 10 percent does not change much the Springer-Verlag.
density. 4. Woods George, 1990, The ICI Polyurethane Book,
Micro -size fillers exhibit different effects for PUR John, Wiley & Sons, Chinchester, England.
foams: for flexible light weight foams the density does 5. L.J.Gibson and M.F.Ashby, 1988, Cellular Solids:
not change significantly up to 60 percent of filler content Structure and Properties, Pergamon Press, Oxford,
for the particles of size 250 microns; for smaller particles England.
about 70 microns the density is same to about 30 percent 6. Gunter Oertel, 1993, Polyurethane Handbook,
filler content and density increases twice at 60 percent. Hanser/Gardner Publications, Cincinnati, USA.
For high-density rigid foams, the density decreases up to 7. ASTM D 3574-95, “Standard Test Method for
five times when the content of micro-filler grows from 0 Flexible Cellular Materials-Slab, Bonded and Molded
to 20 percent; further increasing of filler from 20-80 Urethane Foams,” 1999 Annual Book of ASTM
percent does not change the density. Standards. American Society for Testing and
The modulus of elasticity for flexible PUR foam shows Materials, West Conshohocken, PA, pp. 352-369.
that nano-sized filler with 15 percent of filler content as 8. ASTM D 1621-94, “Standard Test Method for
favorable. The modulus of elasticity increases by nearly Compressive Properties of Rigid Cellular Plastics,”
50 percent of that of unfilled foam. Micro-size filler 1999 Annual Book o f ASTM Standards. American
(wood flour at 70 and 250 microns) shows similar effect Society for Testing and Materials, West
to that of nano-sized filler for modulus of elasticity. Both Conshohocken, PA, pp. 347-353.
the size fractions at 40 percent of filler content are
advantageous. It is interesting to note that nano-sized Krishnamurthi, Bharadwaj-Somaskandan, Shutov / 241
 Figure 1. Relationship between the apparent density of PUR flexible foam and different filler content

Figure 2. Relationship between the modulus of elasticity of PUR flexible foam and different filler content

Figure 3. Relationship between compression force deflection of PUR flexible foam and different filler content

242 / Krishnamurthi, Bharadwaj-Somaskandan, Shutov

 Figure 4 Relationship between the apparent density of PUR rigid foam for micro-filler and diiferent filler content

Figure 5 Relationship between the apparent density of PUR rigid foam for nano -filler and diiferent filler content

Figure 6 Relationship of micro-filler on relative compressive strength of high-density PUR rigid foam and diiferent filler content

Krishnamurthi, Bharadwaj-Somaskandan, Shutov / 243

Figure 7 Relationship of nano-filler on relative compressive strength of high-density PUR rigid foam and diiferent filler content

BIOGRAPHIES

Fyodor Shutov

PICTURE l.doc

Dr. Fyodor (Ted) Shutov is a professor of Chemical Sathyam Bharadw aj - Som askandan
Engineering Department and Center for Manufacturing
Research and Technology Utilization at Tennessee PICTURE 3.doc
Technological University (since 1995). Formerly, he was a
professor and director of AMOCO Center of Excellence in Sathyam Bharadwaj-Somaskandan received the degree
Polymer Science and Engineering at Illinois Institute of of Bachelor of Science and Technology in Chemical
Technology, Chicago (1991-1995). Engineering in July 1999 from the University of Madras
Before coming to the USA, he was a professor at the and a Master’s degree in Chemical Engineering at
Department of Polymer Processing and Material Science at Tennessee Technological University on May, 2001. His
Mendeleev University of Chemical Engineering, Moscow, research area focuses on polyurethane composites.
Russia. He specializes in the processing, material science
and application of foamed and filled polymer materials and
composites as well as in plastics recycling. He holds 20
patents and has published 16 books and more than 350
papers in more than 20 countries. He was the Editor-in-
Chief of the international peer-review journal “Polymer
Recycling,” England (1997-1999). He was awarded a gold
medal by the Cellular Plastics Committee (FSK) of
Germany for his “outstanding contribution in theory and
application of foamed plastics.”

B alachandhar K rishnam urthi

PICTURE 2.doc

Balachandhar Krishnamurthi received a Bachelor’s

degree in Chemical Engineering from Burdwan
University in 1996 and a Master’s degree in Chemical
Engineering at Tennessee Technological University on
May, 2001. His research area focuses on polyurethane
composites.

244 / Krishnamurthi, Bharadwaj-Somaskandan, Shutov

 SESSION G
Automotive Seating Comfort and Durability

TECHNICAL SESSION ORGANIZERS

Jim Tobias, Chair

AIR PRODUCTS & CHEMICALS, INC.

Jim Lambach
BAYER CORPORATION

Allan James
DOW CHEMICAL CANADA INC.

TECHNICAL SESSION MODERATOR

Hamdy Khalil
WOODBRIDGE FOAM CORPORATION

TECHNICAL SESSION VICE - MODERATOR

Jim Lambach
BAYER CORPORATION
Benchmarking of Polyurethane Technologies for Automotive
Seat Cushions

MARK A. KOSHUTE a n d BRIAN L. NEAL

MICHAEL BLASZKIEWICZ Bayer Corporation
Bayer Corporation South Charleston Technical Center
100 Bayer Road P.O. Box 38007
Pittsburgh, Pennsylvania 15205 South Charleston, West Virginia 25303

ABSTRACT construction. The specific attributes required o f the PU

foam depends on the total car concept and seat
Polyurethane foam has been the material of choice for construction. The other elements include the support
automotive seat cushions for over thirty years. Worldwide, construction and cover. All components have to be in
this material is broken down into four main technologies: harmony with the entire car concept.
TDI-HR, TDI/MDI-HR, MDI-HR, and TDI-Hot Cure The PU foams used in automotive seat cushioning can be
foam. There has been much discussion recently as to classified into three main chemistries or technologies.
which technology is best for seat comfort and durability. They are Hot Cure (HC), TDI-HR (including TDI/MDI)
These technologies have been compared in past and MDI-HR foam. Different regions of the world have
publications, but in most cases, comparisons were not different philosophies in seat design; therefore, the
made at the same hardness and density. This type of performance requirements that need to be met are diverse,
comparison often results in misleading conclusions. as well. Thus, the level to which the three main
In this study, foams of different technologies were technologies are used depends on existing manufacturing
systematically prepared at equivalent density and hardness facilities, performance requirements, and market
targets. The hardness/density grades were chosen based on conditions. Figure 1 illustrates the usage breakdown for
typical industry targets. These various foam grades were the three major technologies in the three major market
subject to a battery of durability, comfort, and physical areas, NAFTA, Europe and Japan.
property tests. Results were analyzed and correlations With the globalization of automotive OEM’s as well as
developed. complete seat and foam manufacturers, the versatility in
The results showed that no one technology meets all producing these technologies has increased to the point
criteria over all foam grades. Moreover, within certain
density/hardness grades, the differences in performance
between technologies were small. In addition, in many
cases the results were more dependent on foam density and
hardness than on the technologies that were used. Each
technology had its strengths and therefore the selection of
appropriate technology should be made based on the
specific criteria that are trying to be met.

INTRODUCTION

Polyurethane flexible foam has been the material of

choice for automotive seating for nearly four decades. Its
success is due to the soft initial feel yet firm support that it
provides for the occupants. However, the PU foam Figure 1. Breakdown of world-wide PU usage by region
cushion is just one of several elements of the complete seat and technology.

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 247

where the question arises, “which technology is best for the of the technologies and in doing so, develop a baseline for
various performance requirements?” Many publications in future work.
the past have compared the different technologies, but in Foam density was varied for all foams by water level
most cases, if not all, the comparisons were made at changes. Hardness in the Hot Cure foams was achieved
different densities and/or different foam hardnesses. through index changes and polymer polyol additions.
Therein lies the purpose of this paper. We wanted to Hardness in the TDI and TDI/MDI foams was varied
systematically benchmark the different technologies at primarily by polymer polyol content, and in the MDI
equivalent density and hardness and answer the question: foams through index changes only. A constant final part
“Is one single technology superior or more optimum than thickness of four inches was used to develop a base line for
any o f the others from a performance stand point?” In future studies. Foam thickness variations will be
addition, we wanted to establish a baseline methodology investigated in follow -up work. Test blocks produced at
for evaluating new PU raw materials and technologies. the same foam grade (density and hardness targets) were
We chose to break down the three main technologies into produced at equivalent part weight. The density and
four categories. In addition to the Hot Cure technology, hardness targets for these test foams are summarized in
TDI-HR was broken down into straight TD-80 and Table 1.
TDI/MDI, or TM-20, as it is widely known. The MDI Foams were not made for all technologies in every grade.
technology used in this study was High Monomeric For example, Hot Cure foam technology is not suitable for
Polymeric (HMP) MDI. high density and high hardness grades.

EXPERIM ENTAL Testing Procedures

Test Sample P reparation The test methods used in this study are summarized in
Table 2. The Ball Rebound Test was run according to
For each technology, foam test blocks measuring 15x15 ASTM D 3574. A 16-mm diameter, 16.3-g steel ball was
x 4 inches were prepared in water-jacketed aluminum dropped from a height of 500 mm and the percent rebound
molds. All test blocks, with the exception of the Hot Cure was recorded.
foam technology, were molded using state-of-the-art high Hysteresis Loss was determined using a Japanese
pressure Hennecke foam machines. The Hot Cure samples automotive manufacturer test method. The load-deflection
were produced using the hand mix technique employing a curve was obtained using a cross-head speed of 50
drill-press mixer. The mold temperature and mold release mm/min and a maximum deflection of 75%. The area
used were dependent on the technology. All foams were within this curve is the hysteresis loss.
hand-crushed on de-mold to ensure good dimensional The Transmissivity Test was performed according to
stability. Subsequently, the foams were aged a minimum Japanese Automobile Standard, JASO B 407. A 46-kg
o f seven days at ambient conditions and for 24 hours at 72 “Tekken”-type pressing plate was employed. An Instron
°F and 50% relative humidity before testing commenced. 8500 servo-hydraulic system was used to provide a 5-mm
total vertical amplitude sine wave excitation. The input
Foam Chem istry and output vibration data were collected and analyzed
using two accelerometers and a Bruel and Kjaer 3550
Typical foam formulations were used for each of the four multi-channel analyzer.
technologies studied: Hot Cure, TDI (TD-80), TDI/MDI The 50% Humid Aged Compression Set Test was
(TM -20), and MDI. The formulations used were generic in performed according to ASTM D 3574. The samples were
nature and were not necessarily optimized for any humid aged in an autoclave at 105 °C for 3 hours, and
particular physical property. The goal was to identify subsequently dried in a dry-air oven at 100 °C for 3 hours.
general trends, as well as areas for improvement, for each The samples were compressed 50% and placed in a 70 °C

Table 1. Foam Density and Hardness Targets (or Grades)

____________________ Technology_________
Density Hot Cure TDI TDI/MDI MDI

32 kg/m3 250 N 250 N 250 N 250 N

250 N 250 N 250 N 250 N

40 kg/m3
400 N 400 N 400 N 400 N
250 N 250 N 250 N
56 kg/m3
400 N 400 N 400 N

248 / Koshute, Blaszkiewicz, Neal

 Table 2. Summary of Test Methods in the case of full foam seating. Initial seating comfort is
Method
generally characterized as the soft initial “show room” feel.
Test
Several tests have become important measures of this static
Ball Rebound ASTM D 3574
comfort. These tests include Ball Rebound and Hysteresis
Hysteresis Loss Japanese Standard Loss [2] described in the Experimental section. Results of
Transmissivity JASO 407B these tests are shown in Tables 3 and 4, respectively. (In
50% Humid Aged C. S. ASTM D 3574 these tables, the average of all HR technologies appears in
Wet Compression Set Japanese Standard the last column, and the overall average of a given
Constant Force Pounding ASTM D 3574 technology appears in the last row). As expected, all three
HR foam technologies yield much higher ball rebound
results than the Hot Cure foam. Among the HR foam
oven for 22 hours. The final thicknesses of the samples types, both TDI and TDI/MDI foams yielded slightly
were measured and the compression set as a percentage of higher values, on average, compared to MDI. However,
the original deflection was calculated. increases in ball rebound values can be attained more
The Wet Compression Set Test was performed according effectively by increasing the foam density than by
to a Japanese automotive manufacturer test method. The choosing one HR technology over another. This density
samples were compressed to 50% deflection and placed in effect occurs because as foam density increases, the
a 50 °C / 95% relative humidity oven for 22 hours. The polymer matrix must become less rigid and more elastic to
final thicknesses o f the samples were measured and the maintain the same IFD.
compression set as a percentage of the original thickness Hysteresis loss results in Table 4 show a similar but
was calculated. inverse relationship to the ball rebound results. For both
The Constant Force Pounding (CFP) Dynamic Fatigue tests, best results were recorded with higher density, more
Test was performed according to ASTM D 3574. A force elastic foam. The highest hysteresis losses, regardless of
of 750 N was applied at a rate of 70 cycles per minute for technology, were recorded for the mid-density, hard foams
80,000 cycles. The thickness and 40% IFD hardness were and the low density soft foams. Figure 2 illustrates the
measured one hour after the dynamic pounding had ended, reasonably good linear inverse correlation of ball rebound
and the load loss and thickness loss were calculated. and hysteresis for the HR foams and confirms the
relationship reported in earlier work [1,2].
RESULTS AND DISCUSSION
Transmissivity
Static Com fort
In recent years vibration transmissivity testing has evolved
Although the flexible foam is only one element in the to be an important test for assessing dynamic seating
total comfort spectrum, it can play a major role, especially comfort. The different technologies were tested in our

Table 3. Ball Rebound, %

Density (kg/m3) IFD50R (N) Hot Cure TDI TDI/MDI MDI HR Avg
32 250 47 60 61 55 59
40 250 50 63 60 61 61
40 400 45 56 60 59 58
56 250 68 64 60 63
56 400 65 65 60 63
Technology Avg 47 62 62 59

Table 4. Hysteresis Loss, %

Density (kg/m3) IFD50R (N) Hot Cure TDI TDI/MDI MDI HR Avg
32 250 29.0 27.4 25.8 33.0 28.7
40 250 24.3 21.2 25.8 27.7 24.9
40 400 31.0 28.3 26.4 31.0 28.6
56 250 16.7 14.3 21.1 17.4
56 400 18.2 18.1 23.0 19.8
Technology Avg 28.1 22.4 22.1 27.2

Koshute, Blaszkiewicz, Neal / 249

 with the natural frequency of both internal and external
human body parts causing possible discomfort. It is
desirable to have an amplitude reading below 1.0
(attenuating condition) at 6 Hz for improved comfort.
Table 7 lists these results for the different technologies at
various foam grades. Although some small differences in
technology were measured, the combination of density and
foam hardness had a greater effect than the differences
between technologies. The TDI/MDI technology had a
slightly lower average amplitude than the other foam
chemistries. The data shows that foams produced at high
density and hardness will attenuate at 6 Hz.
The final piece of information that can be gleaned from
the transmissivity test is the frequency at which the foam
starts attenuating or damping vibrations. These results are
Figure 2. Ball Rebound vs. Hysteresis Loss for HR foams. compiled in Table 8. Again, lower frequencies are
desirable for optimum comfort. As with the amplitude at 6
Hz, high density and high hardness foams correspond to
Pittsburgh testing laboratory according to the JASO lower attenuating frequencies. Comparing technologies,
procedure discussed earlier. The results of this testing are TDI/MDI foams attenuated at slightly lower frequencies on
listed in Tables 5 through 8. average than the other technologies. This transmissivity
Table 5 lists the peak transmissivity or amplitude at data should be taken in context knowing that the results are
resonance. The results show that foam hardness plays a influenced by the type of test plate, type of excitation, and
major role in this property. At a given density, as hardness sample thickness used [3].
increases the amplitude at resonance increases. This
relationship held true for all technologies. TDI gave the Fatigue Testing
lowest peak transmissivity of all technologies.
The type of PU technology had very little influence on In order to maintain seating comfort, the occupants must
the resonance frequency, shown in Table 6. On the other be kept as close to the original seating position as possible
hand, increases in foam density and hardness both tended under both static and dynamic fatigue conditions. The
to shift the resonance frequency to lower values. most common static fatigue-type tests are compression set
Another key parameter that can be determined from tests that measure percent set after being held under
transmissivity testing is the amplitude at 6 Hz. This constant deflection for a specified amount of time at
frequency is considered important because it coincides elevated temperature. A more severe variation

Table 5. Peak Transmissivity

Density (kg/m3) IFD50R (N) Hot Cure TDI TDI/MDI MDI HR Avg
32 250 2.2 2.5 3.3 2.7 2.8
40 250 2.7 2.0 3.0 2.2 2.4
40 400 3.4 2.4 3.9 3.1 3.1
56 250 2.0 2.9 2.0 2.3
56 400 2.7 3.6 3.3 3.2
Technology Avg 2.8 2.3 3.3 2.7

Table 6. Resonance Frequency (Hz)

Density (kg/m3) IFD50R (N) Hot Cure TDI TDI/MDI MDI HR Avg
32 250 4.6 4.7 4.4 5.0 4.7
40 250 4.3 4.5 4.3 4.6 4.5
40 400 3.9 4.1 4.1 4.1 4.1
56 250 3.9 3.8 4.4 4.0
56 400 3.6 3.6 3.6 3.6
Technology Avg 4.3 4.2 4.0 4.3

250 / Koshute, Blaszkiewicz, Neal

 Table 7. A m p litu d e a t 6 H z
Density (kg/m3) IFD50R (N) Hot Cure TDI TDI/MDI MDI HR Avg
32 250 1.5 1.6 1.4 1.7 1.6
40 250 1.2 1.3 1.2 1.4 1.3
40 400 1.0 1.2 1.2 1.1 1.2
56 250 1.1 0.9 1.3 1.1
56 400 0.8 0.7 0.8 0.8
Technology Avg 1.2 1.2 1.1 1.3

Table 8. A tte n u a tin g F re q u e n cy (Hz)

Density (kg/m3) IFD50R (N) Hot Cure TDI TDI/MDI MDI HR Avg
32 250 7.8 7.5 6.9 8.1 7.5
40 250 6.8 7.7 6.7 7.5 7.3
40 400 6.2 6.9 6.6 6.5 6.7
56 250 6.7 6.0 7.6 6.8
56 400 5.8 5.6 5.5 5.6
Technology Avg 6.9 6.9 6.4 7.0

of this test is the Humid Aged Compression Set (HACS). compression set testing. Figure 4 shows that the foam
This test consists of an initial high temperature and density influences these results with higher density giving
humidity aging followed by a drying step before the desired lower compression set values. The effect of
conducting the compression set test. Another variation is density was greater than the effect of different HR foam
the Wet Compression Set Test. In this variation, the technologies. In addition, all foam technologies gave
compression set test is run at moderately high temperature comparable results at high density.
and high humidity. The results of the 50% HACS Test and The other class of fatigue testing is known as dynamic
the Wet Compression Set Test are shown in Figure 3. The fatigue. It can be argued that these types are more
average compression set value for each technology foam representative of real life use. Dynamic fatigue tests can
grade is plotted. be conducted two different ways: at a constant pounding
The Hot Cure foam compression set results were force or at constant deflection.
substantially lower than the HR foam results. In fact, the We chose to run the constant force pounding (CFP) test
Hot Cure foams gave the lowest values in both types of according to ASTM D 3574. The description o f this test
compression set testing. This behavior is a well-known was discussed in an earlier section. The different foam
positive characteristic of Hot Cure foam. Among the HR technologies are plotted at three different densities in
foam types, MDI foam performed the best in both types of Figures 5, 6, and 7. Figure 5 shows that at the lowest

F ig u r e 3 . 50% H A C S a n d W e t C om pression S e t re su lts F ig u re 4. E ffe c t o f fo a m d e n s ity on c o m p re ssio n s e t

a v e ra g e d o v e r a ll fo a m grades. V alues a ve ra g e d o v e r a ll te chnologies.

Koshute, Blaszkiewicz, Neal / 251

Figure 5. Load Loss results at 32 kg/m3 using ASTM Figure 7. Load Loss results at 56 kg/m3 using ASTM
Constant Force Pounding Test Constant Force Pounding Test

density and hardness, all foam technologies perform the of fatigue performance. However, Figure 8 shows that the
same in that they all give CFP Load Loss results of CFP test does not correlate very well with hysteresis loss
approximately 25%. as the sole variable.
The middle density load loss results (Figure 6) show that If hardness is also included in the model, the correlation
harder foams tend to perform better than softer foams. improves dramatically. A plot of predicted vs. actual CFP
This is a well-known artifact of the constant force Load Loss using this improved model is shown in Figure 9
pounding test. That is, harder foams are not strained as and is described by the equation:
much as softer foams and thus undergo less stress
softening. Comparing different technologies, TDI and CFP Load Loss = 11.95 - 0.07 x Hardness (N)
TDI/MDI foams tend to give lower load loss than Hot Cure + 1.04 x Hysteresis Loss(%) * '
or MDI foam. The MDI used in this study was a higher
hardness type. Optimization of MDI isomer content and Attempts were also made to correlate the CFP Load Loss
polyol resin would give results similar to the TDI and results with the transmissivity data. Resonance Frequency
TDI/MDI results. was the one parameter found to correlate well with CFP
At the high density (Figure 7) there are some subtle load loss (Figure 10) provided that the HR foam data and
differences between technologies, but all performed well Hot Foam data are treated as separate populations. This
regardless of foam hardness. An interesting follow-up means that the less time consuming transmissivity testing
study would be to repeat this load loss comparison after can be run as at least a screening test to predict constant
performing the fatigue-conditioning step at more severe force pounding fatigue results.
conditions (high temperature and humidity) to determine if The correlation coefficient can be improved even further
the same trends occur. (0.75 to 0.86) if the variables of foam density and hardness
Seat designers and foam producers are continually are included. The improved model is shown in Figure 11
searching for short-term tests that can correlate to and and is described by the equation:
approximate these longer, more time consuming dynamic
fatigue tests, such as CFP. Hysteresis loss is one short­ CFP Load Loss = -14.20 - 0.25 x Density (kg/m3) - 0.006 x
term test that is often looked upon as a possible predictor Hardness(N) + 10.59 x Resonance Freq. (Hz) (2)

Figure 6. Load Loss results at 40 kg/m3 using ASTM

Constant Force Pounding Test Figure 8. Load Loss vs Hysteresis Loss for HR foam.

252 / Koshute, Blaszkiewicz, Neal

Figure 9. Actual and predicted Load Loss using the model Figure 11. Actual and predicted Load Loss using the
with hardness and hysteresis loss. model with density, hardness, and resonance frequency.

This model shows that increasing the density and • Evaluating new PU raw materials and technologies
hardness and keeping the transmissivity resonance using this methodology
frequency low will give better fatigue data. Additional
data should be collected to confirm these models. • Validating the models for long-term dynamic fatigue
data.
FUTURE W ORK
SUMMARY
Using this data as a baseline, future work should include:
A methodology has been developed to systematically
• Evaluating these foams in the constant force pounding evaluate different PU foam technologies at constant
test under higher temperature and high humidity density and hardness. In this study, typical foam systems
conditions. for each technology were used to establish a baseline for
future raw material and technology evaluations. Data from
• Reducing the foam thickness and reformulating each this preliminary work provided some interesting results
technology to get the same density and hardness and insight. No one technology was superior in all
comfort, durability, or physical property tests. However,
each technology had one or more strengths. In several
cases, foam density and/or hardness had more o f an effect
than did the choice of PU technology. Moreover, with
each of the four technologies, foam formulations can be
optimized in most cases to meet the particular OEM
specifications.
The choice of technology should be based on the type of
seat parts being produced, the specifications that have to be
met, and regional economics. A wide variety o f different
PU foam technologies are available to the OEM design
engineers. Thus enabling the challenges provided from the
complex profile of desired seat cushion performance and
whole seat assemblies to be met. Having established a
powerful tool for profiling foam chemistries in conjunction
with one of the most balanced product portfolios within the
industry, Bayer is in the pole position for customer support
for the foams o f a new millenium.

Figure 10. Load Loss as a function of resonance

frequency.

Koshute, Blaszkiewicz, Neal / 253

ACKNOWLEDGEMENTS Michael Blaszkiewicz

The authors would like to acknowledge and thank Ken Dr. Michael Blaszkiewicz is a
Sumner and Ron Zibert for their technical support with Senior Associate Scientist in
sample generation and data compilation. We would also Materials Characterization of
like to thank Walter Heckla, John Liddle, Lon Grant and Bayer Corporation’s
Don Glass for their support in foam testing. Without their Polyurethanes Division. He is
dedicated work, this paper would not have been possible. responsible for developing test
Thanks also to Joachim Werner and Sven Meyer-Ahrens methods that simulate material
for their contributions and technical discussions. performance in real-world
applications and for investigating
REFERENCES structure-property correlations for
polyurethane materials. Michael
1. Hilyard, N.C. 1994. ’’Hysteresis Loss in Flexible received a Ph.D. from The Pennsylvania State University
Polyurethane Foams,” in Low Density in 1992 and spent one year as a visiting scientist at
CellularPlastics: Physical Basis o f Behavior, N.C. Siemens in Munich, Germany, investigating the creep
Hilyard and A. Cunnigham, eds. Chapman and Hall, properties of piezoelectric actuators. Prior to joining
Chapter 8. Bayer, Michael was involved in failure analysis, product
design, and computer modeling of materials at the
2. Blair, G. R., R. So, A. Milivojevich, and J. D. van Westinghouse Science and Technology Center.
Heumen. 1998. “Automotive Seating Comfort:
Investigating the Polyurethane Foam Contribution-
Phase 1,” SAE International Congress and Exposition Brian L. Neal
980656.
Brian Neal received a Ph.D. in
3. Neal, B.L., 2000. ’’Differences in Dynamic Chemical Engineering from the
Performance of Molded Polyurethane Foam as a University of Delaware in 1997.
Function of Pad Thickness and Supported Load,” In 1996, he joined ARCO
Proceedings o f the API Polyurethanes Conference, Chemical Company, which
subsequently was acquired by
BIOGRAPHIES Lyondell Chemical Company. He
joined Bayer Corporation’s
M ark A. Koshute Polyurethanes Division in 2000,
following Bayer’s acquisition of
Mark A. Koshute is a Senior Lyondell’s global polyols business. At Bayer, he is a
Development Scientist in the Principal Research Engineer at the South Charleston
Polyurethane Division of Bayer Technical Center in West Virginia, working to develop
Corporation. He joined Bayer new polyols and systems for HR molded foam
Corporation (then Mobay applications.
Chemical Corporation) after
receiving his M.S. degree in
Polymer Science from the
University of Akron/Institute of
Polymer Science. He has 22 years
of experience developing new polyurethane raw materials
and foam systems for various automotive interior
applications including seating, headrests and headliner
applications.

254 / Koshute, Blaszkiewicz, Neal

New Comfort MDI Molded Foam for Automotive Seating— A
Study of Foam Properties for Riding Comfort Performance

K. SAIKI, S . MURAKAMI, M. HAYASHI, S . EGAWA a n d T. FUKAMI

Nippon Polyurethane Industry Co., Ltd.
Central Research Laboratory
440 Akiba-cho, Totsuka-ku
Yokohama 245-0052
Japan

ABSTRACT characteristics has become a more important factor. It is

well known that measuring of both resonance frequency
Riding comfort performance of automotive seating has and vibration transmissibility by means of a vibration test
become more and more important, and polyurethane described in the Japanese Automotive Standards
flexible foam should have excellent vibration properties Organization (JASO) B407 is typical evaluation method.
and high durability. As more large-scale test, S. E. A. T is know for a
Recently, many studies on vibration properties of substantial evaluation method[l, 2\. But the device for
polyurethane foam have been reported. In general, it was the test is too large-scale and too expensive, and further
thought that to have low peak transmissibility and to keep more, to collect the vibration data is very laborious work.
low hysteresis loss were contrary factors. However we Since it was thought that low vibration transmissibility,
have already developed new comfort polyurethane foam especially at 6 Hz related to a good driving comfort, the
based on MDI technology, which shows not only low materials which gave such vibration character have been
peak transmissibility but also low hysteresis loss, besides used. High BR foam which gives high transmissibility at
good fatigue properties. resonance frequency has been used as a well-known
In this study, a chain of MDI based foams including technique to reduce vibration transmissibility at around
new comfort foams are evaluated by means of vibration 6Hz. But as such foam amplifies the vibration at
test and compression test, and the effect of the physical resonance frequency, great vibration amplification at
and chemical structure of MDI foam on the vibration resonance frequency has become serious problem in
properties is discussed in order to clarify the mechanism. Japan. Actually, reduction of vibration transmissibility
Especially the compatibility of low peak transmissibility at 6Hz shows adverse effect at lower frequency of 1-4Hz.
and low hysteresis loss in new comfort MDI foam is Nippon polyurethane Industry Co., Ltd has already
focused on in view of visco-elastic properties. reported MDI based PUF technology which gives
relatively low peak transmissibility with maintaining low
INTRODUCTION vibration transmissibility at around 6Hz [3,4]. We have
continued the investigation on comfort performance of
Polyurethane flexible oams (PUF) have been widely seat cushions to satisfy both driving and sitting comfort,
used in numerous applications. In Japan. 57% by weight and successfully developed new comfort MDI based
of the total PUF production is used for automotive foam w hic. exhibits good dynamic properties (low
application. The weight of 10kg of polyurethanes is used vibration transmissibility) and static properties (low
for one vehicle and the weight of seat pads accounts for hysteresis loss and low stress relaxation property).
about 90% of them. Weight reduction and durability of In this study, the effect of the chemical and physical
seat cushion and seat back have been main targets. The structure of MDI based foam on both dynamic and static
improvement of comfort properties for seat cushion has properties was evaluated in view of visco-elastic
recently attracted more attention than the development of properties to obtain the compatibility of low peak
further weight reduction. Seat cushion for vehicle plays a transmissibility and low hysteresis loss. In this paper,
role of eliminating a shock and vibration during driving. firstly a brief review of PUF rheology is presented, and
Therefore it should be designed to absorb the shock and new comfort MDI based foam is discussed, which has
vibration as much as possible. Although the evaluation of been developed by considering elasticity and viscosity.
comfort performance of seat cushion has been commonly The riding comfort performance of new MDI comfort
carried out by objective evaluation and the static test such foam through actual human felling test is also evaluated.
as the ball rebound value (BR) and hysteresis loss, the
evaluation of driving comfort by measuring the vibration

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 255

 EXPERIMENTAL Basic foam physical properties were determined in
accordance with JIS K-6400. Dynamic properties such as
Test foams were prepared by using dispensing machine vibration characteristics were investigated according to
and two type of molds which were actual front seat mold JASO test method B 407. As vibration testing equipment,
and square block mold. The molding conditions are Ito Seiki, model C-601 VF was used (Figure 1).
summarized in Table 1, and these are typical conditions to
produce MDI based HR foams in Japan. The raw RESULTS AND DISCUSSION
materials and basic formulation employed in this study
are listed in Table 2 and 3, respectively. Isocyanates Rheological Review
employed in this study were MDI based pre-polymer type.
The functionality of pre-polymer was controlled by In order to develop the foam that has good
changing the ratio of monomeric MDI and polymeric characteristic to give both good driving comfort and
MDI. As polyols, polyether polyols and polymer polyols sitting comfort, it was necessary to understand the
were used. The tertiary amine catalysts and silicone mechanism and the influence on a dynamic
surfactants commonly used for HR molded foams were characteristics and static characteristics. Vibration
employed. transmissibility and resonance frequency were mainly

Table 1. P rocess conditions in h igh - pressure m achine foam ing

Stream temperature (°C)

Isocyanate 25
Polyol 25
Mold temperature (°C) 55-65
Demold time (min.) 5
Machine CANNON
Output rate (g/sec.) 280
Output pressure (Mpa)
Isocyanate 13-15
Polyol 13-15

Table 2. R aw m aterials

Isocyanates functionality NCO%

MDI based prepolymer 2.14-2.16 28.0

Polyols functionality* OHV Molecular Weight

Polyether polyols 2.2-2.8 21-34 5,000 - 8,000
Polymer polyols 2.2-2.8 21-28 5,000 - 8,000

*functionalities were measured byNMR

Table 3. From ulation u se d fo r evaluations

Formulation pbw

Polyether polyols 45-70

Polymer polyols 30-45
Closs linker 1 -5
Water 3 .0- 4. 0
Amine catalysts 0.5 - 0.7
Silicone surfactans 0.5 - 1.5

Isocyanate index 8 0 - 110

256 / Saiki, Murakami, Hayashi, Egawa, Fukami

 studied as the representative characteristics of driving
comfort. As sitting comfort, stress relaxation and
hysteresis loss have been generally discussed. A
spring/dash pot model is usually used to understand the
mechanism of these characteristics as shown in Figure 2.
The elasticity is explained by the Hook’s law using spring,
and viscosity of resin is described by the Newton’s law.
The vibration characteristics is explained by using Voigt
Model as illustrated in Figure 3a, which is the
combination of a spring with spring constant Evc in N
m^and a dash-pot of apparent viscosity /ivc in N s m'1.
The stress relaxation as sitting comfort factor is modeled
as Maxwell Model illustrated in Figure 3b. Resonance
Figure 1. Vibration test equipment
frequency is mathematically derived from Equations (1),
(2) and (3) [5], with fr being the resonance frequency in
Hz, Evc the dynamic spring constant in Nm'1, g being the
gravity constant in m s'2, and m the applied load in N. At
this resonance frequency, the transmissibility ar is the
ratio between transmitted and applied displacements,
and is calculated by using Equation (4). Hence, according
to Equation (3), resonance frequency depends on the
spring constant Evv which may show elasticity of the
foam. The vibration transmissibility will change under the
influence of both of spring and dash-pot.
The stress relaxation is represent as Equation (6),which
is derived from Equation (5), and the property of the
foam will be depend on the ratio of viscosity and
Figure 2 . Spring m o d e l a n d D ash - pot m oled
elasticity. From this equation, with higher elasticity or
lower viscosity, stress relaxation is getting small.
The mechanism of hysteresis loss is explained by using
the model described as Figure 4 and 5 [3]. Hysteresis loss
is thought like a creep in a short period, therefore, loss
property must be influenced by three factors, i.e.
instantaneous elastic deformation, delayed elastic
deformation and plastic deformation. According to these
equations, the higher the elasticity of the foam, the lower
the hysteresis loss.
From the rheological and mathematical studies as
described above, it is expected that resonance frequency,
stress relaxation and hysteresis loss become small when
the ratio of elastic ch tracteristic to viscosity characteristic
becomes high. However, on the other hand, adverse effect
that the peak amplification becomes high appears by Figure 3a. Voigt M odel Figure 3b. M axw ell M odel

(1 )
( 2 )

(5 )
(3 )
( 6 )

(4 )

Saiki, Murakami, Hayashi, Egawa, Fukami / 257

 (7)

instantaneaous ( 8)
elastic deformation

delayed (9 )
elastic deformation

plastic deformation ( 10)

Elasticity and Viscosity of Foam

Some factors which affect the elasticity and viscosity

of the foam are listed in Table 4. It is well known that
both static and dynamic properties are affected by foam
density and foam hardness. This relationship is shown in
Figure 6 and 7. Although both static characteristics and
resonance frequency became lower value with an increase
of the foam density, under the same foam hardness, the
vibration transmissibility showed different tendency
against the change of foam density. One is an increase of
vibration transmissibility. and the other is decrease of that
Under the condition of same density, it was confirmed
F ig u re 4. Mechanism model for hysteresis loss
that the influence of foam hardness on vibration
magnification became high with an increase of foam
hardness. In contrast, the static characteristics were
getting worse with an increase of foam hardness. There
was the hardness point at which resonance frequency
showed the lowest value. In fact, foam density and
hardness for seat cushion have been decided by the
design of the vehicle. Therefore, the following
evaluations were carried out by using constant foam
density and foam ardness.
The typical contribution of isocyanate functionality to
foam properties is summarized in Table 5 and Figure 8.
Isocyanate functionality influenced on the overall static

Table 4. F a cto r affecting elasticity a n d viscosity

Figure 5. Illustration of stress- strain curve

Density
Hardness
simply increasing elasticity of the foam. Because Isocyanate type (functionality)
transmissibility depends on both elasticity and viscosity Polyol type (functionality, molecule weight)
properties. Therefore, to enhancing viscosity property in Copolymer solid level
cell structure within not spoiling the elasticity and to Cell structure
increase elastic property in resin matrix are key points to
obtain both good driving and sitting comfort properties.

258 / Saiki, Murakami, Hayashi, Egawa, Fukami

 Figure 6. Relation between each properties and foam density

Figure 7. Relation between each properties and foam hardness

characteristics and dynamic characteristics of the foam. The comparison results summarized in Table 6 and
The higher the functionality of isocyanate, the better the Figure 9 indicate a correlation between polyol
static characteristics and the resonance frequency. As functionality or molecule weight and foam properties.
expected, the vibration magnification became higher with Since functionality of initiator is not necessary the
an increase of isocyanate functionality. same as that of obtained polyether polyol, the value

Saiki, Murakami, Hayashi, Egawa, Fukami / 259

 Table 5. Typical contribution of isocyanate functionality to foam properties

Formulations Formulation A Formulation B Formulation C

Isocyanate functionality 2.13 2.14 2.16 2.17 2.14 2.16 2.17 2.16 2.17 2.18

Hysteresis loss 22.8 22.0 21.5 21.3 24.4 23.4 22.8 26.2 24.8 24.1
Stress relaxation 11.9 12.8 10.5 11.9 11.9 11.3 11.2 12.7 12.1 11.0
Ball rebound 65 66 67 68 63 64 64 63 63 64
Resonance frequency 3.59 3.55 3.55 3.56 3.73 3.69 3.63 3.63 3.64 3.61
Vibration magnification 2.75 2.84 2.82 3.13 2.68 2.66 2.76 2.77 2.70 2.56
Transmissibility at 10Hz 0.31 0.28 0.31 0.26 0.33 0.32 0.30 0.33 0.31 0.33

*Hardness of all foams were adjusted to225N at 25%ILD

Figure 8. Typical contribution o f isocyanate functionality to foam properties

determined by using NMR was used in this study.[6, 7] functionality of polyol was higher than 2.6 or the
As seen in Figure 9, both functionality and molecule molecule weight was higher than 7000. Although
weight of polyol influenced on comfort characteristics. vibration magnification became low with an increased
Higher functionality and higher molecule weight which POP in the ratio of PPG and POP. resonance frequency
meant higher elasticity of the foam were in connection and stress relaxation became worse. With changing the
with lower resonance frequency and higher vibration ratio of PPG and POP, there was not a remarkable relation
magnification. However, a significant difference on in hysteresis loss (Figure 10). It seemed that other factor
comfort properties was not observed, when the such as air flow might affect on this property.

260 / Saiki, Murakami, Hayashi, Egawa, Fukami

 Table 6a. Typical correlation between polyol functionality

Formulations Formulation D1 Formulation E1 Formulation F1 Formulation G1

Functionality 2.41 2.49 2.50 2.40 2.70 2.80 2.66 2.71 2.78 2.48 2.51 2.54 2.60
Hysteresis loss 26.0 23.1 23.1 26.0 24.5 24.9 18.5 18.5 18.3 20.1 20.4 20.1 18.2
Stress relaxation 13.0 11.6 11.7 13.0 12.0 12.4 9.0 8.8 8.7 9.6 9.8 9.8 8.7
Resonance frequency 3.71 3.49 3.45 3.71 3.64 3.64 3.50 3.51 3.53 3.63 3.61 3.68 3.59
Vibration magnification 2.97 2.89 2.79 2.97 2.71 2.78 2.19 2.16 2.16 1.79 1.96 1.81 2.13
Transmissibility at 10Hz 0.30 0.29 0.29 0.30 0.33 0.32 0.33 0.33 0.33 0.44 0.39 0.42 0.35

F ig u re 9. Typical correlation betw een p olyol functionality and foam properties

Table 6b. Typical correlation betw een p olyol m olecule w eight a n d

foam p roperties

Formulations Formulation M

Molecule weight 6000 6300 6600 7000 7500

Hysteresis loss 26.0 23.1 23.1 21.7 18.5

Stress relaxation 13.0 11.7 11.6 10.7 8.8
Resonance frequency 3.71 3.45 3.49 3.57 3.51
Vibration magnification 2.49 2.55 2.54 2.79 2.56
Transmissibility at 10Hz 0.30 0.29 0.29 0.28 0.26

Saiki, Murakami, Hayashi, Egawa, Fukami / 261

 Figure 9b. Typical correlation between polyol molecule weight and foam properties

Figure 10. E ffect o f copolym er solid level on foam properties

262 / Saiki, Murakami, Hayashi, Egawa, Fukami

 Cell Control Technique The importance of cell control technique is how to adjust
cell size and to eliminate cell membrane. When the cell
The cell structure is also important factor to study the adjustment was done by using the strong silicone
characteristics of PUF. Without changing the composition surfactant, the cell membrane was formed and the foam
of the resin, the easiest way to controls the cell structure became tightness. This cell membrane influences on the
is the choice of silicone surfactant. The comparison of the deterioration of static properties.
effect of silicone surfactants is shown in Figure 11. As Since cell structure is also influenced by the reactivity
shown in this Figure, a strong silicone surfactant can be and viscosity of the resin, some samples were prepared
used to decrease vibration magnification without by adjusted reactivity and viscosity of resin as shown in
deteriorating the resonance frequency. However, on the Figure 12. Although the vibration magnification
other hand, it was confirmed that static characteristics decreased largely, significant changes on resonance
were deteriorated by using the strong silicone surfactant. frequency, hysteresis loss and stress relaxation were not

Figure 11a. E ffe ct o f silicone su rfa cta n t types Figure 12a. E ffe ct o f cell co n tro l on static
on static properties properties; Form ulation K a n d L w ere
included cell co n tro l technique

Figure 11b. E ffe ct o f silicone su rfa cta n t types Figure 12b. E ffe ct o f cell co n tro l on dynam ic
on dynam ic p roperties p roperties; Form ulation K a n d L w ere included
cell co n tro l technique

Saiki, Murakami, Hayashi, Egawa, Fukami / 263

 Table 7. Typical contribution of cell control! on hysteresis toss and air flow rate

Air Flow Rate, cc/cm2/sec. (%) Hysteresis Loss, % (A%)

Speed 8.0 11.6 15.3 18.7 22.5 Slow Quick

Cell controlled foams

Foam A 7.9 10.2 12.9 15.3 18.0 18.97 28.45
(99.1) (87.4) (84.1) (81.7) (79.9) 4.60

Foam B 7.5 9.8 12.7 15.8 18.2 23.85 22.91

(93.9) (84.3) (83.3) (84.3) (81.1) 3.94

None cell controlled foam

Foam D 7.8 11.5 15.1 18.6 21.6 26.36 27.69
(98.4) (99.1) (98.4) (99.4) (96.2) 1.34

observed. This mechanism can be explained as shown in

Figure 13. The transformation of El became small by the
influence of dash-pot V 3, when the foam was
compressed in a high speed as the vibration test. However,
when the foam was compressed slowly, hysteresis loss
and stress relaxation became small value by influence of
both El and E2 due to the transform of 7? 3. The
measurement of foam air flow under different conditions
was carried out to understand this mechanism. The
hysteresis loss test by changing of testing speed was also
carried out. These results are shown in Table 7. The
cell-controlled foam showed almost the same air flow as
uncontrolled foam, under the low flow rate test condition.
In contrast under the high flow rate condition, the air flow
of the controlled foam was lower than that of the
uncontrolled foam.
The hysteresis loss also showed similar tendency as
above. When the stress was removed slowly, hysteresis
loss of both foams were almost same and enough low.
But on the other hand, when the stress was removed
quickly, the cell controlled foam showed larger hysteresis Figure 13. Mechanism
loss than the uncontrolled foam model for cell-controlled
In summary, the ce I control technique is good way to foam
obtain the higher viscosity characteristic without the
deterioration of elastic ity of the foam.

Table 8. C om parisons o f the physical properties o f the developed foam s a n d conventional H R foam

Pysical Test Definition Units Foam A Foam B Foam C Foam D

Overall density kg/m3 55.6 55.6 55.6 55.6

25%ILD N/314cm2 220 221 229 230
Ball rebound % 66 66 70 72
Hysteresis loss % 17.9 22.7 17.8 19.3
Stress relaxation % 8.7 10.8 8.5 10
Resonance frequency Hz 3.47 3.62 3.36 3.39
Vibration magnification - 1.82 2.57 2.82 4.34
Transmissibility at 10Hz - 0.33 0.31 0.28 0.25

264 / Saiki, Murakami, Hayashi, Egawa, Fukami

 Table 9. Summarize of objective evaluations

Standard Foam Foam A Foam B

Actual vibration condition

Low frequency region B B+ B+
High frequency region B B+ B

Static condition
Cushioning feeling B B B-
Bottming feeling B B B-

*Foam Dwas used as standard foam

*A : Excellent, B: Standard, C: Poor

Physical Properties of developed Foam CONCLUSION

We have developed new comfort foam which satisfied In order to develop new flexible polyurethane foam
desirable visco-elastic properties through the above suitable for automotive seat cushion, static and dynamic
studies. The comparison of the physical properties properties of flexible polyurethane foam were studied in
between new comfort MDI based foam and other foams this paper. Especially, the effect of resin component and
are shown in Table 8. Foam A is newly developed new cell structure on hysteresis loss and vibration
MDI based comfort foam in this study. Foam B and Foam characteristics was examined. From the results, it was
C are also MDI based foams having low vibration found that to enhance resin elasticity led to lower
transmissibility with good durability, which were already hysteresis loss and to increase viscosity by controlling
reported in previous paper. Foam D is conventional HR cell structure led to the improvement of vibration
type foam. Foam A shows a lower vibration characteristic absorbance.
than Foam B, and expressed lower hysteresis loss and Through these fundamental study, we developed new
smaller stress relaxation than Foam C and Foam D. The comfort MDI molded foam which showed the
vibration characteristic figures of these foams are shown compatibility of low peak transmissibility and low
in Figure 14. Foam A showed not only low resonance hysteresis loss. Good comfort property of new foam was
frequency, but also low vibration magnification. In also demonstrated by means of actual human feeling test.
addition, Foam A showed low vibration transmissibility
in the high frequency region as compared with Foam B.
REFERENCES
Evaluation of Driving Comfort and Sitting Comfort 1. Leenslag, J. W., E. Huygens, and A. Tan. 1997.
“Recent Advances in the Development and
The studies described above were carried out on the Characterisation of Automotive Comfort Seating
evaluation of visco-elastic properties, such as resonance Foams”; Proceeding of the Polyurethane World
frequency, vibration magnification, hysteresis loss and Congress 1997, pp. 436-446
stress-relaxation property. However, since visco-elastic 2. Croco, G. L., M. R. Kinkelaa, and B. L. Neal. 1997.
properties described above are not necessary consistent “Using S.E.A.T. Values to Improve Automotive
with the evaluation of actual driving and sitting. Seating Foam Dynamics and Seat Comfort”;
Therefore as Foam D being standard, the objective Proceeding of the Polyurethane World Congress
feeling evaluation was also carried out. The comparison 1997, pp. 428-435
results are summarized in Table 9. The objective feeling 3. Kasuga, K., M. Hayashi, K. Saiki, and T. Satou.
evaluation by plural people was carried out under static 1998. “Low Density MDI-based HR Seat Cushion
and driving conditions. As shown in the Table 8, Foam A Having Excellent Durability”; Proceeding of the
and B exhibited superiority on the driving comfort in the Polyurethanes EXPO ‘98, pp. 407-416
low frequency region. Also, Foam A showed slightly 4. Murakami, S., K. Saiki, M. Hayashi, T. Satou, and T.
better comfort in the high frequency range than Foam B. Fukami. 2000. “A Newly Developed MDI-Based
Regarding the sitting comfort, almost of the same result Polyurethane Flexible Foam for Automotive Seat
was obtained between Foam A and standard foam. Cushion Having Both Superior Static and Dynamic
Properties”; Proceeding of the Polyurethane
Conference 2000, pp. 281-286

Saiki, Murakami, Hayashi, Egawa, Fukami / 265

 5. Japanese Automobile Standard Organization, Masahiro Hayashi
B-407-82, “Test Code of Seating Comfort for
Automobile Seats.” Masahiro Hayashi joined
6. Carr, R. H., J. Hernalsteen, and J. Devos. 1994. Nippon Polyurethane Industry
“Determination of Functionality and Functionality Co., Ltd. in 1991 after receiving
Distribution of Polyether Polyols by Quantitative B. S. degree from Kanazawa
13C NMR Spectroscopy”; Journal of Applied University. He has engaged in
Polymer Science, Vol. 52, pp. 1015-1022 the R&D of rigid polyurethane
7. Toyota, Y., N. Hasegawa, H. Wada, A. Horie, S. foam for five years. He is
Hatano, T. Sasaki, and S. Oki. 1997. “NEW currently a technical chief of
GENERATION PREMINOL™: Novel Polyether flexible polyurethane foam
Polyol for High Performance HR Foams”; group, and works for R&D of
Proceeding of the Polyurethane World Congress new MDI flexible polyurethane foams at Central
1997, pp. 421-427 Research Laboratory in Yokohama.

BIOGRAPHIES Shinichi Egawa

Kazunori Saiki Shinichi Egawa joined

Nippon Polyurethane
Kazunori Saiki joined Nippon Industry Co., Ltd. in 2000
Polyurethane Industry Co., after receiving B.S. degree in
Ltd. in 1987 after Chemical Science and
receiving B. S. degree in Technology and M.S. degree
Industrial Chemistry from in Material Physics and
Yamaguchi University. He Chemistry from Kyushu
has worked on developments University in 1998 and
of RIM, integral skin foams 2000,respectively. He has
and flexible foam. He is engaged in the R&D of
currently a manager and flexible polyurethane foam
responsible for flexible foam and he currently works for R&D of new MDI flexible
R&D at Central Research polyurethane foams at Central Research Laboratory in
Laboratory in Yokohama. He Yokohama.
completed a training program on isocyanate-based
polymer foams at Polymer Technologies, Inc., a
subsidiary of the University of Detroit Mercy, in 1992. Takao Fukami

Shunsuke M urakami Takao Fukami joined Nippon

Polyurethane Industry Co.,
Shunsuke Murakami received Ltd. in 1972 after receiving a
his M.S. degree in organic B.S. degree in Industrial
synthesis chemistry from Chemistry from Kyoto
Sibaura Institute of Institute of Technology in
Technology in 1994. He 1972. After development
joined Nippon Polyurethane work on TPU, elastomers,
Industry Co., Ltd. in 1994. He coatings and adhesives, he
has engaged in the has been engaged in the
development of polyurethane development of PU foam for
coating and adhesives for 5 past seven years. He is
years. He is currently a technical chief of MDI flexible currently Assistant to Research Director and rigid foam
polyurethane foam group at Central Research Laboratory R&D manager at Central Research Laboratory in
in Yokohama. Yokohama.

266 / Saiki, Murakami, Hayashi, Egawa, Fukami

A Study of the Influence of Polymer and Cell Structure on
Polyurethane Foam Properties for Better Riding Comfort

MASAHIRO ISOBE, KAZUTO USAKA, HIROYUKI UTSUMI a n d KAZUHIKO OHKUBO

Mitsui Takeda Chemicals, Inc.
R&D Center, Urethane Materials Laboratory
580-32 Nagaura, Sodegaura-city
Chiba 299-0265
Japan

ABSTRACT In terms of seat back, low -density is strongly desired. But

the demand for seat cushion is divided into low density and
Control technology of some comfort properties of comfort. Low-density seat cushion has been developed for
polyurethane seat cushion has been getting important. rear cushion and compact car, and comfortable seat cushion
has been developed for front cushion and luxury car.
• control of ball rebound resiliency suited to each car We consider that durability is also one of the important
concept properties relating to comfort, because durability can be
• control of transmission of the vibration from an described as performance of maintaining certain riding and
automotive body to a passenger seating comfort for a long time. That is one of the reason
• good durability for keeping riding comfort for a long why high durability is strongly desired.
time A few years before, higher resiliency was desired for
spring elimination and higher durability in Japan. Spring
We consider that it is important to control ball rebound elimination means flexible polyurethane foam is substituted
resiliency and transmission of the vibration individually. In for seat spring. As it had been considered that higher
this study, we compared some properties of foams with resiliency contributed better riding comfort, some
wide range of polymer and cell structure in order to discuss researches of high resilience flexible foam were reported [1].
the effect of these structures on characteristics mentioned But there is no report that higher resilience lead to better
above. Ball rebound resiliency is influenced by both riding comfort by human sensitive test. Flexible
polymer and cell structure, but polymer stmcture is the polyurethane foam with higher resiliency, consequently, is
more influenced factor. Transmissibility of lower frequency, not accepted widely, and it is used for limited market.
2-4 Hz, is affected by the cell structure strongly. The Recently, demand for comfort has been getting
coefficient of viscosity of polyurethane foam can be important, and many papers were reported, such as
adjusted by control of the viscous resistance of air evaluation of distribution of occupant-to-seat interface [2] ,
flowability through cells appropriately. This resistance is vibration characteristics [3], and human sensitive tests for
defined with the cell window size. On the other hand, comfort of seat cushion [4]. As mentioned above, it is
transmissibility of higher frequency, such as 6 Hz, is reported that flexible foam with lower ball rebound (60-
influenced by polymer structure. Both polymer and cell 70%) leads to better seating comfort by the measurement of
structure affect durability, but it was found that the distribution of occupant-to-seat interface [2]. Riding
contribution of polymer structure is bigger. comfort affected by the properties of flexible foam was
These polymer and cell structure can be controlled by the studied by M. Kubo et al [4]. As mentioned above, the
selection of raw materials and formulation. Low air interest for vibration characteristics and comfort has been
flowability generally causes tight cell at demolding, but getting bigger.
regulating cell open mechanism using appropriate raw In addition, there is the method in JASO concerned with
material can avoid the problem. vibration characteristics, and we consider the vibration
characteristics correlates with comfort strongly.
INTRODUCTION We have developed polyurethane raw materials and
formulation technology to lower vibration transmissibility,
Demand for flexible polyurethane foam for automotive because we assume the reduction of the vibration
seat pad, in Japan, is low -density and comfort. Needless to transmissibility provides better riding comfort. But we
say, low -density contributes cost saving and lighten weight should know it is impossible to solve this problem only by
of automotive, and comfort contributes passenger’s the flexible foam, it depends on concept of automotive,
satisfaction. suspension, seat spring and so on.

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 267

 It is valuable to develop the elemental technology of consider this value is the best parameter to indicate
controlling vibration characteristics to satisfy the needs viscoelasticity of polymer matrix.
from customers. • Cell structure: effecting to air flowability.
There are two types of the demands for vibration
characteristics. First, reducing transmissibility at resonance At last, we introduce new flexible foam system with
frequency. This type is adopted in automobiles that absorb better vibration characteristics by controlling polymer
vibration in higher frequency by another countermeasures. elastic property and cell structure. The system will provide
Second, reducing transmissibility at higher frequency. better riding comfort.
These types are accepted in automobiles that absorb
vibration in lower frequency (2-4Hz). Needless to say, low
transmissibility in wide frequency is much preferred. But it EXPERIMENTAL
has not been developed the ideal flexible foam that absorbs
vibration wider frequency. Preparation of Foam Samples
We mentioned above about contribution to vibration
characteristics for riding comfort. In addition to these Molded foam samples used in this study were prepared in
characteristics, there are other important physical properties our laboratory using either a general handmixer or a high
mentioned below which affect comfort of seat cushion. pressure dispenser (MEG MQ -18 and MKE20-3B). Other
conditions are summarized as follows:
• durability: for keeping comfort constantly. Static
durability (diy/wet compression set) and dynamic • Test mold; 400 x 400x 100mm, an aluminum alloy
durability, such as the method A of fatigue by mold
constant-load pounding test of JIS K6400, are used • temperature of raw materials; 25C
for the index of durability. • mold Temperature; 65C
• stress relaxation: for keeping driver’s eye point • cure; for Sminites
constantly for long driving. As stress relaxation is
large, thickness of foam changes from time to time Foam samples of various polymer elastic property and
and eye point gets below during driving. From cell structure were made in the hand pouring.
above reason, small stress relaxation is desired. The basic formulation is shown as follows:

In this report, we focused on polymer elastic property Polyol/Polymerpolyol 100

and cell structure individually.• Cross Linker 1.0- 3.0
H 20 3.0
• Polymer elastic property: indicating polymer itself Amine Catalyst 0.5
without influence of cell structure. Flexible foam Silicone Surfactant 1. 0 - 2.0
shows viscoelastic behavior, but it’s difficult to
measure the viscoelasticity of polymer matrix. Isocyanate Cosmonate™ TM -20
Therefore we used the foam resiliency at higher air NCO Index 95-100
flowability (named as latent resiliency) as an index (TM-20: TDI-80/polymericMDI=80/20)
of polymer elastic property in this study. We
We controlled polymer elastic property by using three
Table 7. Physical Properties of foam samples on this study (Hand pour test).

Sample A1 A2 A3 B1 B2 B3 C1 C2 C3
Controlled Properties
Latent resiliency (%) 76/High 76/High 76/High 70/Mid. 70/Mid. 70/Mid. 64/Low 64/Low 64/Low
Air flowability Hiah Mid. Low Hiqh Mid. Low Hiqh Mid. Low
Tight cell value (N) 350 710 840 340 580 720 340 430 670
Density over-all (kg/m3) 54.2 54.2 54.1 54.1 54.3 54.2 54.3 54.3 54.1
25%ILD (N/314cm2) 207 202 204 204 200 201 203 200 202
Ball Rebound (%) 76 74 71 70 67 65 64 62 61
Wet set (%) 9.1 10.2 12.3 12.4 16.0 17.8 17.5 19.9 21.3
Air Flowability (cc/cm2/s) 87 42 18 61 36 27 89 51 25
Stress Relaxation (%) 8.9 8.8 8.6 11.3 12.0 11.3 13.2 13.4 13.7
Vibration Characteristics
Resonance Frequency
(Hz) 3.1 3.2 3.3 3.2 3.3 3.4 3.5 3.6 4.0
Transmissibility at
Peak (A/A0) 5.0 3.2 2.2 4.7 3.0 2.5 4.5 3.1 2.2
Transmissibility at
6Hz 0.5 0.6 0.7 0.6 0.7 0.8 0.7 0.8 1.0

268 / Isobe, Usaka, Utsumi, Ohkubo

types of polyol (Table 1), the high (sample A series), middle
(sample B series) and low (sample C series) resiliency.

F igure 1. C ru sh in g ro lle r sensor. Figure 6. Correlation among latent resiliency,

air flowability and transmissibility at

Figure 3. Correlation among latent resiliency, Figure 4. Correlation among latent resiliency, air
air flowability and wet set. flowabilitv and stress relaxation.

Figure 5. Correlation among latent resiliency, Figure 6. Correlation among latent resiliency,
air flowability and transmissibility at peak. air flowability and transmissibility at 6Hz.

Air flowability was controlled by the selection of silicone property. The foam was made at 200N/314cm2 of hardness.
surfactant. Generally, the cell structure of flexible foam is opened,
We used the latent as the index of the polymer elastic and it leads air flowability high. In this study, we controlled

Isobe, Usaka, Utsumi, Ohkubo 1 269

the cell structure from opened cell to closed cell by the Foam sample (thickness of 100mm) was crushed
selection of silicone surfactant. mechanically by the crushing roller, and the load during
crushing was measured. In the case of the value over 800N,
M easurement of Foam Properties it is difficult to demold and sometimes physical splitting
apart happens. So it is preferable that tight cell value is
General measurement of physical properties was done in under 700N.
according with JIS K-6400.
Vibration characteristics were measured in accordance DISCUSSION
with JASO B-407. In this measurement, 490N of load was
applied using “tekken” plate. In this paper, we measured vibration characteristics, wet
Stress relaxation was measured by the way mentioned compression set and stress relaxation as indexes of comfort
below. After preliminary compression to 75% of thickness on purpose to discuss the correlation among polymer elastic
foam sample, 196N/314cm2 of load was applied in load- property, cell structure, and riding comfort.
limit deflection, and deflection fixed. After 5minites, the Latent resiliency and air flowability were used as indexes
percent of decreased load was measured. of polymer elastic property and cell structure respectively.

“Tight Cell Value” M easurement As index of comfort:

• vibration characteristics for riding comfort
Flexible foam is an aggregation of small cell, and it can • wet compression set for durability
be divided into opened and closed cell. As a lot of closed • stress relaxation for maintaining eye point
cells are remained, it causes tight foam at demolding. In
order to discuss the intensity of tight foam, we measured As index of polymer elastic property and cell structure:
load to crush flexible foam immediately after demolding by • latent resilience
crushing roller (Figure 1). We call this intensity “tight cell • air flowability
value”.
The conditions of crushing roller are mentioned below. Physical Properties and Vibration Characteristics

• Flexible foam sample was compressed to 50% of its Table 1 shows physical properties of flexible foam
original thickness. samples prepared in this study, and Figure 2-6 indicate the
• Diameter of roller was 140mm. correlation among latent resiliency, air flowability and the
• Rotation speed was 5rpm. indexes of comfort.
• Roller with load cell sensor measured the load Ball rebound is influenced by both polymer elastic
during crushing. property and cell structure, but we can recognize influence
of polymer elastic property is clearly (Figure 2).

Figure 7. Vibration characteristics of A2 Figure 8. Vibration characteristics of A1,

and B1. A2 andA3.

270 / Isobe, Usaka, Utsumi, Ohkubo

 Table 2. P h ysica l P ro p e rtie s o f fo a m sam ples on this s tu d y
(H a n d p o u r test)

Sample D1 D2 D3 D3’
Controlled Properties
Latent resiliency (%) 65/Low 65/Low 65/Low 65/Low
Air flowability Hiah Mid. Low Low
Tight cell value (N) 480 820 1200 680
Density over-all (kg/m3) 54.3 54.3 54.1 54.1
25%ILD (N/314cm2) 201 204 203 202
Ball Rebound (%) 65 63 61 61
Wet set (%) 12.7 15.7 21.3 19.1
Air Flowability (cc/cm2/s) 79 38 21 20
Stress Relaxation (%) 12.8 12.9 13.1 12.8
Vibration Characteristics
Resonance Frequency
(Hz) 3.4 3.6 4.0 4.0
Transmissibility at
Peak (A/AO) 4.5 3.2 2.2 2.0
Transmissibility at
6Hz 0.7 0.8 0.9 0.9

F igure 9. O bservation o f ce ll structures, h ig h e r vs. lo w e r a ir flo w a b ility .

Wet compression set (Wet set), as an index of durability, sample with higher ball rebound like a sample A1 shows
is influenced by both latent resiliency and air flowability higher transmissibility at resonance. But then we can reduce
shown in Figure 3. Flexible foam with higher latent transmissibility at resonance by reducing air flowability (A2
resiliency and higher air flowability shows better durability. and A3 from A1 in Figure 7).
Figure 4 shows that stress relaxation is influenced only In the case of flexible foam with 70% of ball rebound
by latent resiliency, and no influence of air flowability was (Bl), both resonance and 6Hz transmissibility are decreased
found. by reducing air flowability and by using higher resilience
In vibration characteristics, Figure 5 shows polyol (A2 from B l in Figure 8).
transmissibility at resonance frequency (peak) is strongly Stress relaxation is influenced only by polymer elastic
affected by air flowability. So it is efficient to reduce air property, and no influence of air flowability was found. It
flowability f o r . lowering transmissibility at resonance indicates higher polymer elastic property is the only way to
frequency. lower stress relaxation. In a viewpoint of keeping eye point
On the other hand, transmissibility at 6Hz (higher constantly, higher polymer elastic property is better.
frequency) is influenced by both latent resiliency and air Furthermore higher resilience foam shows better wet
flowability shown in Figure 6. compression set as the index for durability. It is one of the
From these results, we assumed the indexes of flexible reasons that flexible foam with higher resiliency is
foam for comfort could be mainly controlled by polymer preferred.
elastic property and cell structure.
So we can easily design properties, such as vibration
characteristics, for the demand from our customers, by
selecting of polyol, additives and so on. For example, foam

Isobe, Usaka, Utsumi, Ohkubo / 271

 L o w e r a ir flo w a b ility foam o f cu rre n t
Figure 10. Correlation among latent resiliency,
form u la tio n; quite a fe w cell w in d o w s are
air flowabilitv and tioht cell value. re m a in e d (D3).

Correlation Between A ir Flowability and Cell Structure;

Observation with Microscopy

In order to understand correlation between cell structure

and air flowability we observed cell structure using optical
microscopy.
In comparison with sample D1 and D3, cell size of D1
using low -efficient surfactant that promotes open cell, is
larger (Figure 9). And D3 includes more closed cell
windows in comparison with D l. It can be said that air
flowability is determined by cell size and remained cell
windows.

The M ethod of Improving Durability L o w e r a ir flo w a b ility foam o f new

d e ve lo p e d form ulation; m a n y cell w indow s
Durability of flexible foam is mainly determined by are o p e n e d ra th e r than cu rre n t form ulation
polymer elastic property and cell structure mentioned above. (D 3 ’).
In general, decreasing polymer elastic property and air
flowability causes worse durability as shown in Figure 3. In
terms of cell structure, we assume flexible foam consisted
of fine cell structure is more affected by humidity because
cell fraction become thinner and it’s surface area was wider.
However other factors also affect durability. We reported
durability of foam was provided by making phase
separation clearer [1,5]. In these papers, it is written that
durability of flexible foam is improved by controlling the
phase-separated morphology.
Wet compression set of C l -3 is not good. Especially,
sample with fine cell (C3) shows worse durability. In order
to improve durability of these samples, we used “New
Polyol 2” (Dl-3) [6]. Durability of Dl-3 is better than Cl-3 S h ru n k cell stru ctu re , sh rin ka g e
by partially using “New Polyol 2” with same resiliency h a p p e n e d n o n - m e c h a n ic a l cru sh in g fo r
(Table 2). the tig h t cell fo a m (D3).

Cell Structure and Tight Cell Value: F igure 11. O bservation o f ce ll structures.
An approach to B etter Moldability
demold happen sometimes. In this study, foam samples with
In general, fine cell structure controlled by silicone high-efficient surfactant show higher tight cell value such as
surfactant cause tight cell at demolding. In this case, A3, B3, and D3. We consider these systems as out of
physical splitting apart at crushing and/or shrinkage after productive line. Table 1, 2 and Figure 10 show the tight cell

272 / Isobe, Usaka, Utsumi, Ohkubo

 Table 3. Hypothesis of cell opening mechanism.

Silicone surfactant, Cell profile at blowing Cell structure Air flowability

Formulation
Low-efficiency, Cell-opened Coarse cell High
Current
High-efficiency, Cell-closed Fine cell Low
Current
High-efficiency, Cell-opened Fine cell Low
New developed
value measured with crushing roller sensor. In the case of cell windows are already opened at demolding. Table 3
using low-efficient surfactant, the values were 350-450N, indicates the comparison new developed formulation with
and sample with high-efficient surfactant showed over current formulation about considerable cell opening
800N. mechanism.
We developed new formulation to reduce the tight cell In order to confirm the hypothesis mentioned above, we
value, and it became possible to satisfy both moldability observed cell structure of D3 and D3’ by microscopy before
and the index of comfort. The tight cell value and physical mechanical crushing (Figure 11). No difference in cell size
properties of D3 and developed new formulation, D3’ are between sample D3 and D3’ was observed, but more
shown in Table 2. This new developed formulation can opened cell windows could be observed in sample D3’
reduce only the tight cell value without influence on foam Many cells of D3 with no mechanical crushing were
properties and curing. Furthermore, air flowability can not shrunk shown in Figure 11, but ones of D3’ were not
change, in spite of reducing the tight cell value. shrunk.
The mechanism of reducing the tight cell value can be After mechanical crushing, D3 and D3’ indicates same
explained below. air flowability, and same cell structure. New formulation
Generally, cell windows open by copolymer particle and did not appear to have an effect on the cell morphology of
the discreteness of hard segment. At the same time of cell crushed foam.
windows open, coalescence happens, and cell size becomes
larger. Foam Preparation by Machine Foaming Test
Flexible foam with low -efficient surfactant consisted of
large and opened cell windows. On the other hand, flexible Properties of D3’ prepared by machine foaming test are
foam with high-efficient surfactant includes a lot of fine cell shown in Table 4. In this system, we use “New Polyol 2” to
and closed cell windows, and these cause tight cell. improve durability, high-efficient surfactant to make cell
In comparison with these foams, as new developed structure fine, and new developed formulation to reduce
formulation technology is used, the foam consists of many tight cell. Flexible foam of D3’ with low resiliency
fine cells and opened cell windows. In other word, many indicates better durability, and lower transmissibility at low

Table 4. Physical Properties of foam samples on this study

(Machine pour test)

Sample number D3’ Modified A3

Controlled Properties
Latent resiliency (%) 65/Low 76/High
Air flowabilitv ___ Low_____________ Low
Tight cell value (N) 700 500 480 720 ; 600 540 500
Density over-all (kg/m3) 54.3 54.3 54.2 58.6 i 52.1 52.3 52.0
25%ILD (N/314cm2) 202 223 248 247 ■ 198 226 246
Ball Rebound (%) 62 62 62 62 . 73 73 73
Wet set (%) 14.7 14.5 14.1 13.8 ; 10.6 10.6 11.2
Air Flowability (cc/cm2/s) 20 22 26 17 i 23 24 22
Stress Relaxation (%) 12.0 12.2 12.1 11.8 ! 9.2 9.8 9.7
Vibration Characteristics
Resonance Frequency
(Hz) 3.8 3.7 3.7 4.0 j 3.3 3.4 3.5
Transmissibility at
Peak (A/AO) 2.3 2.6 2.8 2.1 ; 2.6 3.0 3.3
Transmissibility at
6Hz 0.9 0.8 0.8 1.0 ! 0.6 0.7 0.7
◄— —► ◄----------- — ►
Index changes Index changes

Isobe, Usaka, Utsumi, Ohkubo / 273

frequency. In comparison with current system, new BIOGRAPHIES
developed system is superior in terms of durability, tight
cell value, and transmissibility at resonance frequency. M asahiro Isobe
Flexible foam with higher resiliency prepared by
machine foaming was also evaluated (Table 4). This system, Masahiro Isobe received a B.S. and
modified A3, shows higher durability. This system showed M.S. degree in Molecular Science
lower tight cell value because of developed formulation, and Technology from Kyushu
and lower transmissibility in wider frequency (3-10Hz). University in 1992 and 1994,
respectively. He joined Mitsui
CONCLUSIONS Chemicals, Inc. as a research chemist
in 1994 has been engaged in
We studied the influences of polymer elastic property and development of flexible polyurethane
cell structure to the foam properties, which are concerned foam, and he is currently researcher
with riding comfort. In this paper, we focused on vibration of flexible polyurethane foam at
characteristics, wet compression set, and stress relaxation as Mitsui Takeda Chemicals, Inc.
indexes of polymer elastic property and cell structure.
Kazuto Usaka
• Vibration characteristics, durability were influenced by
both polymer elastic property and cell structure. Stress Kazuto Usaka received a B.S. and
relaxation was influenced only by polymer elastic M.S. degree in Chemistry from
property. Science University of Tokyo in 1994
• In terms of vibration characteristics, transmissibility at and 1996, respectively. He joined
higher frequency (6Hz) could be reduced by higher Mitsui Chemicals, Inc. as a research
polymer elastic property, and transmissibility at chemist in 1996, and engaged in
resonance frequency was decreased by fine cell development of polyurethane
structure. elastmers for four years, and he is
• Higher polymer elastic property and fine cell structure currently researcher for development
reduce vibration transmissibility in the wide of flexible polyurethane foam at
frequency. Mitsui Takeda Chemicals, Inc.
• The countermeasure of tight foam caused by high
efficient silicone surfactant has been done by our new Hiroyuki Utsumi
formulation technology (controlling of cell opening
mechanism). This new technology did not affect to
Hiroyuki Utsumi received a B.S.
foam properties and fine cell structure.
degree in Industrial Chemistry from
• The problem of lower durability was dissolved by
Science University of Tokyo in 1989.
substituting “New Polyol 2” partially for conventional
He joined Mitsui Chemicals, Inc. as a
polyol.
research chemist in 1989, engaged in
the polyetherpolyol research for three
PREFERENCES
years, and he is currently a chief
chemist and engaged in the
1. Utsumi H., Obata M., Nishioka T., Ueno K., and Sakai S.,
development of flexible polyurethane
Proceedings of the 35th Annual Polyurethane Technical/
foam at Mitsui Takeda Chemicals,
Marketing Conference, 489 (1994)
Inc.
2. Isobe M., Hiraide T., Utsumi H., Ohkubo K., and Sakai
5., Proceedings of the Polyurethanes EXPO ’98, 491
Kazuhiko Ohkubo
(1998)
3. Toyota Y., Wada H., Sasaki T., Takahashi K., Takeyasu
H., and Min K., Proceedings of the Polyurethanes Kazuhiko Ohkubo received a B.S.
EXPO ’99, 389 (1999) and M.S. degree in Synthetic Organic
4. Kubo M., Terauchi F., Aoki H., Isobe M., and Ohkubo Chemistry from Chiba University in
K., Proceedings of the 35th United Kingdom Group 1982 and 1984, respectively. He
Meeting on Human Responses to Vibration (2000) joined Mitsui Chemicals, Inc. as a
5. Hiraide T., Utsumi H., Obata M., Ohkubo K., Kita M., research chemist in 1984, engaged in
and Sakai S., Proceedings of the Polyurethanes the polyetherpolyol research for two
EXPO ’96, 487 (1996) years, and he is currently a chief
6. Utsumi H., Isobe M., Hiraide T., Ohkubo K., and Sakai chemist and the manager of flexible
5., Proceedings of the Polyurethanes Conference 2000, polyurethane foam team of Mitsui
355 (2000) Takeda Chemicals, Inc.

274 / Isobe, Usaka, Utsumi, Ohkubo

Alternative Methods for Durability Specification

ANDREW Le FEVER a n d JAM ES McEVOY

Johnson Controls, Inc.
Automotive Systems Group
49200 Halyard Dr.
Plymouth, Michigan 48170

ABSTRACT measure will better predict long-term mechanical

durability than will the density test.
In the automotive seating industry, convention holds
that the density of a polyurethane (PU) foam pad has a INTRODUCTION
strong relationship to its overall durability. As foam
density is increased in a given seat system, durability Automobile manufacturers require that seats undergo
performance with time is thought to improve. Thereby, design verification testing that includes a cyclical fatigue
seat systems suppliers attempt to control long-term test called Jounce. The Jounce test uses a lever arm with
complete seat comfort and appearance by constraining an attached metal buttform applied to the surface of a
foam density. This density to performance trend is cushion or back. The buttform has the dimensions of the
generally true; however, recent studies have shown that SAE 3D H-point machine for a 50th percentile male [2].
foams of the same density may have different A load is added to the lever arm above the buttform, and,
performance based on chemistry and processing [1]. This for the tests discussed here, is 200 lbs. The platform on
leads to a problem for seat manufacturers that receive which the seat sits is oscillated for a given number of
foam from several different production facilities with cycles, 100,000 in this paper, and a string potentiometer
different line sizes, catalyst amounts, cure times, and is used to measure the sag of the system with time.
other fundamental processes. Also, as advances in PU To quantify the foam component of the seat system, a
formulation are made, or as process conditions change Urethane Foam Dynamic Fatigue, or UFDF, test is often
with time, foam durability may fluctuate at a constant used. The machine setup for this test is described in [3].
density. The test is essentially very similar to Jounce, but more
These discoveries led to the postulate that although precisely controlled and uses a standard circular flat
foam density is a large factor in determining mechanical metal plate indenter of 200mm. The only notable
durability, density alone does not quantify other factors differences between Jounce and UFDF are that the Jounce
that affect foam durability. The hysteresis loss test may test is taken on a complete seat while UFDF is run on
better determine foam durability. This test measures the foam alone, and the Jounce starts at a standard load,
percent of energy not recovered during a compression while the UFDF starts at a standard compression.
cycle of the foam. In fact, the word hysteresis means One fact that will be reiterated throughout this paper is
mechanical lost energy. McEvoy has previously this: there is a strong relationship between density and
presented the hysteresis loss test method [1]. This hysteresis loss. In fact, density is a prime factor affecting
measure suggests that a lower number is better for foam the hysteresis loss of a pad of given geometry. While
durability, in that, the foam itself absorbs less energy at a acknowledging the relationship between density and foam
lower hysteresis loss. It will be shown that the hysteresis durability, this paper’s emphasis is to determine whether
loss test is a composite measure of several factors hysteresis loss, as compared to foam density, will more
affecting durability including density and foam precisely define foam durability.
formulation. Therefore, the belief is that this composite

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 275

 Table 1: Jounce Study Comparison
Study Sample Size Hysteresis Density Hysteresis Pad Weight H ysteresis
Loss Variance Correlation to Correlation to Correlation to
Variance (StDev/M ean) Total Jounce Total Jounce Pad Weight
(StDev/ Mean) Sag (r2) Sag (r2) <r2)
1 8 S e a ts 18.72% 8.85% 0 .37 0 .3 8 0 .8 5

2 12 S e a ts 34.67% 37.71% 0.74 0.81 0 .8 8

JOUNCE STUDIES CONTROLLING THE HYSTERESIS LOSS

Initially, density and hysteresis loss were studied in In order to accurately show the difference between
relation to the performance of a complete seat. Two hysteresis loss and density, it is necessary to know how to
different sets of foam were poured and tested for change hysteresis loss at a given density. To this end a
hysteresis loss and pad weight. (Note for this paper, pad study was created to understand the effects of formulation
weight and density are used interchangeably as they have on the hysteresis loss number. In this designed
a very strong correlation for these formulations, r2 =0.99). experiment, the factors that we used were water, solids,
Table 1 shows that in both studies, the foam pads’ pack, and index. The water level is given in parts per
hysteresis loss and density gave similar correlation to hundred of polyol. The term “solids” refers to the
total Jounce sag. This is due to the strong relationship percent of acrylonitrile and styrene. Pack is the amount
between density and hysteresis loss in these two of extra material beyond the minimum fill, and index is
experiments (r2of 0.85 and 0.88). The second study was the stoichiometric ratio of TDI to polyol blend. The
improved over the first, in that, the inaccuracy of the parameters and levels for this experiment are given in
Jounce test as a foam measure was overcome. This was Table 2.
accomplished by increasing the sample size and The experimental design was a Taguchi L27 and
increasing the variance of hysteresis loss and density, and required that we manufacture 27 foams of various
so the correlation drawn to seat sag is improved for the compositions. This type of designed experiment is
second study (Table 1). Again, in the second study, explained in [4]. Anova analysis was used to determine
hysteresis loss and density give a similar correlation to the hysteresis response and the results are shown in
total sag from the Jounce test. From these studies it is Table 3 and Figure 1. Table 3 gives the percent
apparent that both density and hysteresis loss can give a contribution of each factor. From the response graph
good correlation to seat durability performance, and also shown it is obvious that water is the largest contributor to
that they typically will agree with each other. hysteresis value. However, there appears to be a strong
The Jounce test, however, is not an accurate interaction when solids are held constant and water is
experimental test to measure foam as a component in varied, seen in BxA. This implies that there is an
order to distinguish between density and hysteresis loss, optimum water level for a given solids content.
due to the noise from the other seat components. The
inaccuracy of the Jounce test can be seen in study 1, Table 3: Contribution to H ysteresis Response
where a lower foam property variance resulted in a poor Contribution
Variables
correlation to total sag (Table 1). For this reason the
Dynamic Fatigue test is used, and the UFDF number is Water 56.38%
taken to be a good approximation of the mechanical Solids 0.17%
durability [3]. Water Held Constant 0.58%
Solids Varied
Table 2: Experimental Parameters Solids Held Constant 34.76%
Level 3 Water Varied
Variables Level 1 Level 2
Pack 0.86%
Water 1.6 2.3 3.0
Index 2.82%
Solids 10 15 20
Error and 4.43%
Pack 5% 10% 15% M iscellaneous Factors
index 95 100 105 Total of all Factors 100%

276 / LeFever, McEvoy

 based on the hypothesis that softer foam would perform
better. We have seen from the previous DOE that
reducing the solid’s content improves the hysteresis loss
number, which makes this a plausible hypothesis.
However, the seat is constrained by the occupant’s seated
position in vehicle, called H-point, and so the firmness of
the foam cannot be radically altered [2]. After several
optimizations, a formulation was found that passed the
test. In this case the solution was to use the lowest levels
of water and co-polymer polyol. This solution agreed
with the parameter contributions from the DOE and led
the development of the next block study.

DYNAMIC FATIGUE CORRELATION

To prove which property (hysteresis loss or density) is

Figure 1 better at predicting durability, 100mm thick foam test
blocks were poured. The experiment involved 5 levels of
From this experiment it was determined that the lowest
density, each with a high and low hysteresis loss at each
hysteresis can be obtained as follows:
density, for a total of 10 formulations (Figure 2).
1. Select the approximate hysteresis loss desired and
For a given density, the higher hysteresis loss
then determine water level. formulation is achieved using high levels of water, co ­
2. Select solids needed to meet foam firmness. polymer, pack and index. The lower hysteresis loss for a
3. Optimize with index and pack
given density is the converse. It has the lowest possible
water, minimum fill, less co-polymer and lower index.
CASE STUDY Formulations A and A’ are an example, and in this case
the higher hysteresis formulation is 300% harder than the
The seat system for this study is a full foam cushion
lower hysteresis formulation. In this manner, the
supported by a metal pan. The seat test method is a heat
typically high correlation between density and hysteresis
and humidity vibration test, used to determine the settling
loss is shattered (r2= 0.35). It is also interesting to note
of the seat system with time. Since the foam is acting as
that formulations A and B are separated by approximately
the suspension, the foam’s performance is critical to
28 kg/m3 in density and over 30kgf in firmness, measured
meeting the maximum sag specification.
at 25% compression, and yet they have the same
The seat system is placed into a heat and humidity
hysteresis loss (Figure 2, Figure 6).
controlled chamber at 30°C and 80% relative humidity.
In this manner, we were able to ensure that the
A 68-kilogram buttform is applied to the seat, allowed to
distinction between hysteresis loss and density could be
settle, and then vibrated for a period of 500,000 cycles at
made. That is, for a variable such as durability to have a
0.7 G. In order to pass the test, the seat system must sag
high correlation to one of the independent variables, it
less than 12mm at the completion of 500,000 cycles.
would necessarily have a poor correlation to the other. A
Initial samples failed the test; therefore, an attempt was
situation could exist where one of the dependent factors
made to optimize the foam’s durability performance at
would have an average correlation to both hysteresis loss
the specified 52 kg/m3density.
and density. In this case we would assume that the factor
Several experiments were performed varying weight,
is dependent on both variables. Also, it is possible that
firmness, and hysteresis loss. Table 4 summarizes the
the factor could show a poor correlation to both variables.
results. Trials 1 through 4 were run with lower firmness

Table 4: H ysteresis Loss and Performance Optimization

Sample Density IFD H ysteresis Dynamic Weight
(Kg/m3) (Newtons) L oss Result (grams)
1 52 380 28% 17mm 1250
2 52 378 30% 19mm 1250
3 52 360 25% 16mm 1250
4 52 355 24% 15 mm 1250
Solution 52 380 21% 10mm 1250

LeFever, McEvoy / 277

 Figure 2 Figure 3

The experimental setup involves extreme conditions,

and so will not necessarily show strong correlation for
either factor to durability, but should show a definitive
difference between the two factors. It was necessary to
take these boundary case examples due to the fact that the
hysteresis loss number is largely dependent on the
density. Again, this can be seen in Figure 2, in that, as
density is increased, both the high and low hysteresis loss
formulations have improved hysteresis loss.
The pads were subjected to a UFDF test at 8Hz to
50,000 cycles. The test initialized at 45% to 55%
compression, measured loads at those compressions, and
cycled between the constant loads measured.
Figure 4
After the test was completed, measurements of percent
height loss, percent load loss, and creep were recorded.
These values are compiled in a composite measure of Then linear correlations were calculated from the
durability, called the UFDF number, with a lower number UFDF and components to hysteresis and density. The
being better. hysteresis loss gave a much higher correlation to UFDF
number than density, as can be see in Figures 3 and 4.
UFDF= 5xHeight Loss + 1.5xLoad Loss + 4 Creep (1) Figure 4 shows pad weight versus UFDF number and the
graph appears similar to Figure 1, which compares pad
The UFDF number, as mentioned before, will be taken weight with hysteresis. This appearance similarity is
to be an accurate approximation of the long-term indicative of the relationship of hysteresis to UFDF that is
mechanical durability of the foam. shown in Figure 3. Due to test time constraints, three

Figure 5 Figure 6

278 / LeFever, McEvoy

samples were tested at A, A’ and B and only one sample BIOGRAPHIES
at each other level. Formulations A and B are labeled on
Figures 2 and 3 to show that despite the difference in Andrew J* LeFever
density between those two formulations, they perform
similarly on dynamic fatigue. Figure 6 shows a summary Andrew LeFever is a graduate of The
of the results for formulations A and B. These University of Michigan, Ann Arbor.
formulations gave foams that are remarkably different in He has a Bachelor of Mechanical
firmness and density, and yet have the nearly the same Engineering, and has worked in the
hysteresis loss and UFDF number. foam industry for one year. He
Hysteresis loss and density were then compared with currently works for Johnson Controls,
every component of the UFDF number. Figure 5 Inc. as a Foam Engineer.
summarizes the results. In every case, the hysteresis loss [email protected]
number was a better predictor of each facet of mechanical
durability as measured by UFDF. Janies T. McEvoy

CONCLUSIONS James McEvoy has worked for

Johnson Controls, Inc. for 15 years in
Depending on the experimental setup, both hysteresis various roles, focusing on
and density can give a good correlation to complete seat polyurethane foam and manufacture.
durability performance. Clearly, this is due to the fact He has a Bachelor of Science in
that hysteresis loss and density are strongly related. Chemistry from the University of
Hysteresis loss, however, incorporates density and several Wisconsin. His entire 23 years have
other factors that affect durability to provide a better been spent in either polyurethane foam
overall measurement of foam long-term mechanical manufacture or seating development.
durability. This is evidenced in the case study discussed, [email protected]
where a complete seat was improved to pass cyclic
durability testing at a given density and hardness.
Furthermore, it is proved by the Dynamic Fatigue ACKNOWLEDGEMENTS
correlation study discussed. The case study shows that
hysteresis loss can be used to optimize durability a The authors would like to thank Bob Whitman of the
constant hardness and density, while the UFDF study Bayer Corporation for his knowledge, experience, and
shows that hysteresis will predict durability regardless of assistance in data collection. They would also thank
firmness and density. In closing, although density plays a Dawn Belding, Paul Blackburn, John McNeill, Patricia
large role in determining the performance of polyurethane Schultz, and Mike Kolich at Johnson Controls, Inc. who
foam, clearly hysteresis loss is a more accurate helped make this possible. Also our thanks to the entire
measurement of that durability, and a necessary step Johnson Controls Urethane Technology Center.
toward changing the concept of foam to an engineered
product.

REFERENCES

1. McEvoy, J. T. 2001. “Advances in

Polyurethane Foam Property Control,” SAE
Technical Paper Series. 2001-01-0325
2. SAE. 1995. Surface Vehicle Standard. J826.
3. Cavendar, K. D. 1990. “New Dynamic Flex
Durability Test,” SPI 33rdAnnual Technical/
Marketing Conference ”, pp. 282 -288.
4. Peace, G. S. 1993. Taguchi Methods. Reading,
MA: Addison-Wesley Publishing Co.

LeFever, McEvoy / 279

 Accelerated Aging and Durability Testing of
Polyurethane Foams

JAMES T. MCEVOY a n d RYOKO YAMASAKI

Johnson Controls, Inc.
49200 Haylard Drive
Plymouth, Michigan 48170

ABSTRACT foam is dependent on the ap p licatio n o f reliab le laboratory

test p ro ced u res that can b e d irectly rela ted to in - use
T oday, purchasers o f v eh icles are expecting th e ir perform ance.
investm ents to pro v id e high levels o f perfo rm an ce for
longer p eriods o f tim e. O ne o f th e few features o f th e D urability is an im portant m aterial p ro perty
vehicle that directly affects h o w the occupant feels about co n sid eratio n for p olyurethane foam in autom otive seating
his or h er vehicle is the seat system . O ne o f th e m ajo r applications. In fact, du rab ility is u su ally discussed in
com ponents o f this com posite system is th e m o ld ed term s o f short - and long - term durability. S hort - and long ­
polyurethane foam th at su p p o rts the vehicle occupant. te rm du rab ility are natu rally related; th e ir classification
V alidation testing is em ployed to ensure th at the foam w ill stem s from th e ir relationship to th e in - use situation. S hort ­
provide optim um perform ance fo r the life o f th e vehicle. te rm du rab ility deals w ith th e changes in fo am perform ance
T here are m any test p ro ced u res to qualify th e d u rab ility o v er a single driving session and is rela ted to optim um
perform ance o f polyurethane foam in the laboratory. A perfo rm an ce o f th e fo am ’s functions o f support and
greater challenge is to relate lab o rato ry d ata to in-vehicle v ib ratio n attenuation [1]. L ong-term durability is
perform ance in a tim ely fashion. T he goal is to p ro d u ce co n cern ed w ith changes in p erfo rm an ce th a t are influenced
an accelerated aging te st p ro ced u re that w ill aid th e b y th e daily cycles o f use, an d rest and is related to ability
seating engineer in th eir effort to m eet th e cu sto m e r’s o f th e seat to m aintain a sp ecified level o f perform ance
long - term durability and p erform ance expectations. o v er th e life o f th e v ehicle [2, 3]. T his p a p e r is concerned
w ith long - term durability perform ance.
T his p ap e r w ill explore tw o aspects o f durability testin g
for polyurethane foam qualification. First, th e m echanical A utom otive cu sto m ers ’ ex p ectatio n s h av e changed in
breakdow n o f th e foam is affected b y tem p eratu re - reg ard to v ehicle lifetim e. It is not unusual for auto
induced accelerated aging. T hen, the resu lts o f th e m an ufacturers to advertise 100,000 m ile claim s for tune -
accelerated aging tem p eratu re study are used to gen erate ups. T he polyurethane foam m ust p erfo rm to th is standard
com pression testing data w ith lab o rato ry p ro d u ce d fo am as w ell and th e foam qu alificatio n p ro ced u res n eed to
sam ples to com pare w ith com pression te st results reflect th e m ore stringent requirem ents. It is unreasonable
conducted w ith vehicle foam that had been in service for to su b ject foam to te st track testin g for 100,000 m iles.
ten to tw elve years. T h erefore, fatigue tests have been developed th at correlate
w ith long - term foam p erform ance [2]. T h ese tests are used
INTRODUCTION to p red ict the ability o f th e foam to w ithstand norm al use
co n d itio n s over the lifetim e o f a vehicle. But, it w ould be
T he autom otive industry continually strives to red u ce o f in terest to b etter sim ulate th e effects o f foam aging.
m anufacturing costs and im prove th e q u ality o f its
products. A lthough there are a num ber o f m eans o f A ccelerated aging tests are designed to m im ic the effects
accom plishing these goals, th e evolution o f te ch n olo g y o f tim e an d /o r exposure on a m aterial o f interest in a tim ely
offers an attractive long - term solution. T he im provem ent fashion. O f p articu lar interest is th e use o f tem perature to
o f m aterial properties in this case, m olded p o lyurethane condition the foam in a m an n er sim ilar to the effects o f

,
POLYURETHANES EXPO 2001 SEPTEMBER 30 - OCTOBER 3, 2001 281
tim e. T he choice o f tem perature has been a subject o f B ayer C orporation p rovided raw m aterials from that era
debate because chem ical degradation that could not take that are still in use today. Johnson C ontrols provided the
p lace in the in - use application can occur at inappropriately form ulation to produce foam blocks com plem entary to the
high aging tem perature [4]. truck bench seat foam. T hese foam blocks w ere m ade at a
co re density o f 3 8 - 4 0 kg /m 3 and JIS 50% IFD o f 280 N.
T his p ap er will first exam ine the effect o f accelerated
aging tem peratures in a standard com pression set test to Aging Tem perature Determ ination Test
determ ine an appropriate value. The chosen tem perature
will be em ployed to age foam block sam ples p roduced T h e first question to be answ ered w as w hat tim e -
w ith the sam e m aterials and procedure as foam cushions tem p eratu re com bination w ould p roduce a change in the
pulled from vehicles that had b een in service for a num ber m echanical p roperties o f the foam w ithout affecting
o f years. B oth the original and new ly m ade and aged chem ical degradation. T he test m ethod that w as used w as a
foam sam ples will b e tested in a repeated com pression test com pression set test w here the sam ple size w as 50 x 50 x
to determ ine i f the aging produces results that reasonably 25 mm. T he sam ples co ntained skin and w ere placed into
sim ulate the perform ance o f the vehicle cushions. th e com pression set fixture w ith the skin down. T he
sam ples w ere then com pressed to 33% o f the total height o f
EXPERIMENTAL 25 m m and w ere p laced in an oven for a period o f 24 hours
at th e tem peratures o f interest. All rep o rted values
Test Foam Preparation represent th e average o f three sam ples calculated as Ct
values as th ey reco v er o v er the course o f 96 hours.
T est foam s w ere p repared using state-of-the-art high
pressure m etering m achines. F oam b lo ck s m easured Repeated Compression Test Method
approxim ately 400 x 400 x 100 m m and w ere p rep ared in
a cast alum inum w ater - jacketed m old set to m aintain a A standard repeated com pression test procedure w as
m old tem perature o f 70 ° C. All foam s w ere allow ed to em ployed to com pare th e foam b lo c k s ’ perform ance versus
develop final pro p erties for tw o w eeks. U nless otherw ise th e tru ck seat cushions. Sam ples w ithout skin, sized 100 x
stated, the foam sam ples w ere allow ed to condition for 24 100 x 50 m m , w ere cut from both sets o f foam s. A n initial
hours in a tem perature an d hum idity controlled m easurem ent o f th e sam p les ’ thickness, 25% and 50%
environm ent o f 21 ° C and 50% R.H ., respectively, p rio r to constant force deflection (C F D ), w as m ade for later
testing. com parison. T he sam ple w as p laced into the repeated
com pression device and com pressed from 100% to 50% o f
T o test the effect o f accelerated aging tem perature, its original thickness 80,000 tim es at a rate o f 60 cycles a
foam blocks w ere produced at a density o f 48 kg/m 3 and m inute (1 H ertz). A fter the com pression cycle w as
JIS 50% IFD o f 350 N using a contem porary cushion com plete, th e test specim en w as rem oved from the tester
form ulation. and allow ed to “ re st ” fo r 30 m inutes. The final thickness
w as m easured and com bined w ith th e initial thickness value
T he repeated com pression test w as used to com pare th e to calculate the height loss. A fter an additional 24 hours,
foam perform ance o f new ly m anufactured m aterial w ith the C F D values w ere then com pared to the initial values.
m aterial that h ad been used in the field for a num ber o f
years and, m ore im portantly, for a num ber o f m iles. T he RESULTS AND DISCUSSION
type o f vehicle chosen w as the 1989 and 1990 G M full -
size truck. T ruck seats w ere obtained from these m odel Aging Tem perature
year trucks found in junkyards. T he m ileage from the
selected trucks w as recorded w hen the foam cushion T able 1 b elo w contains the results o f the com pression set
sam ples w ere obtained. tests as a function o f th e tem perature conditions and the
tim e o f recovery.

Table 1. Compression Set Tests Results

Temperature (°C) 0.5 Hour 1 Hour 4 Hours 24 Hours 48 Hours 96 Hours
70 8.1% 4.5% 4.0% 3.5% 3.3 % 3.3%
90 11.4% 7.7% 6.5% 5.5% 5.1% 5.1%
100 14.4% 11.6% 9.5% 8.7% 8.3% 8.1%
110 33.4% 33.1% 33.2% 33.2% 33.1% 33.0%
120 32.7% 32.8% 32.6% 32.5% 32.6% 32.7%
150 32.9% 32.7% 32.6% 32.8% 32.8% 32.7%

282 / McEvoy, Yamasaki

 T hese results reveal a num ber o f in terestin g facts. First
and forem ost is the fact that the results o f tests conducted
at 110 ° C and above show total perm an en t deform ation.
T his is indicative o f chem ical deg rad atio n and is
consistent w ith the results o f p rev io u s studies [4]. W hat is
equally interesting is that som ew here b etw een 24 and 48
hours, the values ob tain ed at 100 ° C an d b elo w reach a
p lateau o f recovery, w ith som e sm all residual
deform ation. A gain, 100 ° C has been d eterm ined to be
the m axim um tem p eratu re for accelerated ag in g to avoid
chem ical decom position o f th e foam .

R e p e a te d C o m p re ssio n T e s t

T able 2 belo w contains the resu lts o f th e repeated

com pression set tests as a function o f th e foam sam ple
conditions. From th e sa m p le s ’ results, one can see som e Figure 2. Repeated compression test load loss as a
trends for both the height loss and load lo ss show n in function of vehicle mileage.
F igures 1 and 2 below .

L in ear regression o f the rep eated com pression set results

Table 2. Repeated Compression Test Results yields the follow ing equations. T he height loss as a
Height Loss Load Loss function o f th e vehicle m ileage m ay b e ex p ressed as
Sample Condition (%) (%)
A New Foam - 0.68 12.18 Height Loss = 1.82 x 10 ~ 2 x Mileage + 1 .0 1 (1)
0 Miles
B New Foam - 3.42 17.11
0 Miles & w here th e height loss is ex p ressed in p erc en t and the
Aged 24 Hrs m ileage is ex p ressed as thousands o f m iles.
@ 100 °C
C 99K Miles 2.9 16.21 T h e load loss as a function o f the vehicle m ileag e m ay be
D 101.3K Miles 3.2 16.5
ex p ressed as
E 97K Miles 3.0 16.0
F 198.6K Miles 4.3 21.3
Load Loss = 0.46 x 10 ~ 2 x Mileage + 1 1 .9 (2)

w here the load loss is expressed in p ercen t an d the m ileage

is ex p ressed as thousands o f m iles.

T h e heat - aged sam ple results can be substituted into both

equations to find the in - use m ileage th at co rresp o n ds to
heat aging at 100 ° C for 24 hours. T h ese results are
rep o rted in T ab le 3 below . T h e approxim ate average is
120,000 m iles for this com bination o f accelerated aging
conditions.

Table 3. Calculated Values

Property Measured 24 Hours @ 100 °C
Height Loss 132K Miles
Load Loss 113K Miles

Figure 1. Repeated compression test height loss as a SUM M ARY

function of vehicle mileage.
T h e conclusion from this d ata is that 24 ho u rs at 100 ° C
is v ery close to 100,000 m iles o f aged usage. H ow ever,
additional data will be required for confirm ation due to the

McEvoy, Yamasaki / 283

lim ited data and sam ple size o f this effort. T he pro p o sal 4. B rasington, R. and H. De R oeck. 2000. “ A ccelerated
going forw ard is to p erform com plete seat testin g with T esting for D urability P erform ance o f A utom otive
new and aged (100 ° C for 24 hours) foam sam ples to see S eating F oam , " API P olyurethanes 2000 C onference.
the im pact o f several com plete seat tests such as jo u n c e
and squirm , sliding entry, and egress tests. T his p ap er B IO G R A P H IE S
suggests that foam aging at 100 ° C for 24 ho u rs m ay
represent a potential standard for aging conditions. J a m e s T . M e E voy

ACKNOW LEDGM ENTS Jim has w orked for Johnson C ontrols

for 15 years in vario u s roles, m ost
W e w ould like to thank th e m em bers o f B ayer recen tly focusing on polyurethane
C orporation and Joh n so n C ontrols F oam B usiness U nit foam developm ent. H e receiv ed his
for all their support and hard w ork that m ade this p ap e r B ach elo r o f S cience in C hem istry
possible; B rian N eal, B ob W hitm an, D aw n B elding, from the U niversity o f W isconsin. H is
K untal T hakurta, D iane Jones, T racy T uthill, C hris entire 23 - y ear career has been spent in
M organ, B lane N eveaux. eith er polyurethane foam m anufacture
o r seating developm ent. Jim ’s e m ail
REFERENCES address is Jam es.T .M cE v o y @ JC I.com

1. N eal, B. L. and J. L. L am bach. 1999. “ Physical Ryoko Yamasaki

P roperty R esponse o f P olyurethane F oam U n d er
D riving C on d itio n s, ” SA E International C ongress and R yoko receiv ed dual M .S. in C hem ical
E xposition 9 9 M - 1 01. E ngineering an d M aterials Science &
2. C avender, K. D. and M . R. K inkelaar. 1996. “ R eal E ngineering from M ichigan State
T im e D ynam ic C om fort and P erform ance F acto rs o f U niversity in 2000. She jo in e d
P olyurethane F oam in A utom otive S eating, ” SA E Johnson C ontrols in 2 000 and has
International C o ngress and E xposition 960509. b een w orking on polyurethane foam
3. W ilson, A. an d B lair, G. R., “ P olyurethane dev elo p m en t follow ing the com pletion
A utom otive C ushioning: In C ar D urability vs. F oam o f E ngineering D evelopm ent P rogram .
P roperties, ” P ro ceed in g s o f the SPI 35rd A nnual
T echnical/M arketing C onference for P olyurethanes,
pp 4 7 8 - 4 8 8 , 1994.

284 / McEvoy, Yamasaki

Measurement of High Resiliency Moulded Foam Properties as a
Function of Climatic Conditions

G. R. BLAIR, A. R U SS, D. E. BRADT a n d R. J . BAILEY

Woodbridge Foam Corporation
8214 Kipling Avenue
Woodbridge, Ontario L4L 2A4
Canada of PUF cushioning to changes in test conditions. It should be
noted that foe effects of changes in climatic conditions on foam
A B ST R A C T hardness ofcured foam are reversible (3). The presence of water
acts as a plasticizer allowing foe polymer chains additional
mobility and hence facilitating chain slippage (2).
High resiliency moulded foam is normally tested under
standard laboratory conditions, e.g. 23 ± 2°C, 50 ± 5% RH but PUF cushioning must provide comfort and durability support
other conditions may be specified- It is well known that critical but also to act as an almost ‘'neutral -responding” medium to any
foam properties such as hardness and resiliency are influenced by climatic changes. These changes may have considerable influence
environmental changes. This may affect foam performance on seat functionality. As seat complexity increases this is of
especially in automotive seating applications. Changes in seating increasing importance. In addition to providing excellent
support, durability and comfort may occur and critical safety support, a seat must also perform as specified, if a seat heater is
components such as occupant sensors may be affected. provided, (for tactile comfort) and/or an occupant seat sensor for
airbag deployment for impact safety. If a particular PUF type
We have measured property changes that result from different does not maintain a prescribed range of say hardness (IFD)
climatic conditions from -30 to +95°C and 0 to 95% RH. A consistency over a reasonable range of climatic conditions to
number of foam chemistries both experimental and in worldwide allow consistent airbag sensor operation in emergencies, then a
production have been thoroughly examined for property different PUF type may be necessary.
responses after conditioning. Cushion hardness, hysteresis loss
and thickness have been shown to change after environmental
Vibrations transmitted from foe road surface and converted to
conditioning. Resiliency as measured by ball rebound also energy as heat may be transferred to foe cushion and thence to foe
responds to climatic changes. We have analyzed our extensive seat occupant As foe foam rises in temperature, IFD decreases
data for temperature or relative humidity responses and for the
and this may result in a perceived loss of seat comfort especially
combined response to both of these variables. Also we are able to during an extended journey. After cooling and recovery, foe
demonstrate combined responses of properties that contribute to
foam will regain hardness and foe perceived comfort will be
seating comfort as a function of climatic conditions. Therefore,
regained (6). Also a seat may become uncomfortable if it cools
we are able to demonstrate to automotive seating down below 23°C and hardness will increase. By adding heat via
designcrs/engineers the temp/RH ranges where an important
a seat heater, original hardness will be regained and foe occupant
cushioning parameter such as seating hardness (and hence H-
should experience regained comfort Maintaining seat hardness
point) is relatively consistent Thus, important safety components
by heating or cooling (say via a seat blower) may be the best way
such as occupant seating sensors should remain functional over a
to achieve comfort on an extendedjourney.
relatively wide range of climatic conditions. Seating of various
constructions, e.g. metal pan, flexible suspension, sinespring
- has indicated relatively little change
Our earlier work (10 12)
wire, have been tested to ascertain foam responses specific to
in PUF cushion performance with long term seat useage.
these constructions. Full details of foam seat performance after
environmental conditioning will be discussed. However, all our performance/foam measurements were made
under standard test conditions, 23 ± 2°C, 50 ± 5% RH in foe
1. INTRODUCTION laboratory. The data described in this paper seeks to illustrate foe
relative consistency of PUF seating over a wide range of
Moulded temperatures and relative humidities that will include most
climates worldwide. Our data should answer enquiry’s from
OEM or Tier 1 seating engineers that foe properties of PUF
Automotive polyurethane foam (PUF) parts are manufactured cushioning will allow critical seating parameters to be maintained
to meet specifications at standard test conditions, usually 23 ± or safety equipment activated to allow consistent occupant
2°C and 50 ± 5% relative humidity. Other conditions may be protection over a wide range of climatic conditions.
specified, e.g. 20°C, 65%RH, but there is a general trend to
change to the conditions outlined above. Some previous work by major raw material suppliers has
usually looked at specific foam chemistries. In general, these
Foam properties will respond to changes or variations in earlier studies have indicated ’linear” relationships between foam
-
climatic conditions, i.e. temperature/relative humidity (1 9). hardness and climatic conditions. In most cases, foe
Since PUF foam is one of foe major components in seating recommendation is that testing be done in well controlled -
sunDori durability and comfort, we have measured the response

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 285

laboratories using the standard conditions mentioned earlier. conditions. Previously we have shown that these properties are
Further recommendations include testing o f complete seats using correlated (14,15).
various foam chemistries and densities (13).
Also seat cushions have been subjected to vibrational
W e have tested several PUF chemistries including production transmissivity tests to ascertain their comfort responses.
and experimental formulations. Our testing methodology was Cushions have additionally been tested on their support pans to
developed using standard test blocks prior to testing actual two different ball rebound tests to determine impact energy
production cushions. A range o f seating types representative o f absorption.
several major seating constructions from a number o f the w orld 's
leading OEMs have been evaluated. Both block and cushion foams were subjected to dynamic
mechanical analyses (DMA) and standard physical property
2. Test Methodology testing.

All foam blocks or cushions were conditioned for more than 3. Moulded Foam Manufacture
one week in our laboratory at 23 ± 2°C, 50 ± 5% RH prior to
testing. Test blocks m ade using two different PUF chemistries were
made:
Initially samples were tested under these conditions using an I. TDI- based, (1001) foam manufactured using SAN
Instron model 4201. Hysteresis loss, thickness and IFD at 50% polymer polyol on a production line with a core density
deflection were recorded. Only one deflection was selected, as o f 50 kg/m3.
we wanted to test the foam as rapidly as possible after II. MDI -based foam at 85 and 951 was poured in our
conditioning. Samples were then conditioned for a minimum o f Corporate Research and Development facility in
24 hours in an environmental chamber manufactured by Hotpack Woodbridge, Ontario, Canada using high pressure
Corporation (175 Series Incubator). Each sample was then impingement mixing and is representative o f our
removed from the chamber and tested as rapidly as possible as production MDI chemistry, core density 50 kg/m3.
noted above. For temperatures Iowa* than 23°C, samples were III. Cushions for various OEMs were manufactured in
preconditioned in a standard refrigerator or in a Revco Freezer several production plants to meet custom er 's
(model ULT2050). requirements. These cushions were selected to represent
a number o f seating constructions, e.g. metal pan, sine ­
Production cushions w ere supported on their seat pans as used spring and w ire - held support In addition, these
in each vehicle and the foam plus pan preconditioned prior to cushions were chosen as they represented a relatively
testing. In some cases, an extra stability support had to be used to wide range o f cushion thicknesses (76 to 93 m m o f
allow testing to be performed as the seat pans were not stable foam). A composite cushion was also selected as it
enough when supported on the Instron test plate. consisted o f a plastic pan supporting a PUF cushion
(thickness 31 -55 mm) and an overlayer cushion o f
After hardness testing over a wide range o f test conditions, ball rubberized coconut fibres 18-21 mm in thickness. All o f
rebound (resiliency) o f core foam samples was tested to the these cushions were made using TDI/SAN -filled polyol
ASTM D3574 - 95 standard and repeated after 24 hour chemistry.
conditioning as outlined above. This was done to allow us to
compare hysteresis losses against resiliency at known test

These cushions were manufactured to the following specifications:

D ensity S p ecificatio n H a rd n e ss S p ecific a tio n

C u sh io n N u m b e r kg/m 3 N (D eflectio n ) o r S u p p o rt S y stem

m m (F o rc e)

1 >52 2 1 6 - 2 5 5 (25% ) M etal Pan

2 60 ± 5 1 8 1 - 2 1 1 (25% ) M etal Pan

3 >40 243 - 297 (40% ) Sinesprings

4 67 ± 3 2 2 - 2 8 m m (3 0 0 N) Plastic Pan

5 47 ± 2 333 - 392 (50% ) M etal Pan

All o f the cushions we have tested comply with these OEM specifications.

286 / Blair, Russ, Bradt, Bailey

 TA B LE 1 N o rm alized B lock H ysteresis Loss

i Temperature °C
Relative Humidity,
-30 10 23 40 50 70 90
%

A 150
5 B 145
C 145
A 117
20 B 101 99 96 93 90
C 105 102 99 96 94

A -
30 B 101 94
C 103 97

A 120 100 107 107 104 106

50 B 106 100 95 93 94
C 110 100 97 95 95

A 109 103 104 108 111

70 B 99 93 97 93

C 100 94 97 94

A 112 105 108 112

90 B 94 92 88
C 95 92 88

A=SAN/TDI
B=MDI-85I
C=MDI-95I
TABLE 2 Normalized Block Thickness @ 4.5N
| Temperature _5c ----------
Relative Humidity,
-30 10 23 40 50 70 90
%
A 99
5 B 99
C 98
A 99
20 B 99 100 100 100 100
C 99 99 99 100 100
A
30 B 100 100
C 100 100
A 100 100 99 99 100 101
50 B 100 101 101 101
C 100 101 101 102
A 100 100 100 100 100
70 B 100 101 101 102
C 101 101 102 102
A 99 100 100 101
90 B 101 101 101
C 101 102 102
A=SAN/TDI
B=MDI-86I

Blair, Russ, Bradt, Bailey / 287

 TA B LE 3 N o rm alized B lock IF D @ 50% D eflection

I Temperature ~ *c-------------
Relative Humidity, % -30 10 23 40 50 70 90

A 129 110

B 129 110
o
C 130 115

A 108
B 108 104 99 98 94
20
C 106 102 97 95 91

A -
B 105 98
30
C 104 95

A 106 100 90 88 87 86
B 99 100 95 93 89
50
C 100 100 93 91 86

A 92 83 82 80 80
B 92 91 90 85
70
C 91 89 89 82

A 86 74 74 75
B 84 81 78
90
C 83 78 74

A = SAN/TDI
B = MDI 85I
C = MDI 95I

FIGURE 1

288 / Blair, Russ, Bradt, Bailey

 Normalized HystarasI* Loss vs %RH 9 2 i *C

FIGURE 2

FIGURE 3

FIGURE 4

Blair, Russ, Bradt, Bailey / 289

4.1.2 Ball Rebound

Core specimens (100 x 100 x 50 mm) were tested using a Cushion No. 2, which exhibited the largest changes in
Time -Tech Inc. model TT502 Ball Rebound Tester constructed to hysteresis loss has the smallest changes in displacement with
the ASTM D3574 -95 test protocol, i.e. ball drop height o f 500 either temperature or relative humidity increase. Cushion no. 3,
mm. Values were obtained in triplicate on different areas o f the which it may be recalled, sits on sinusoidal springs, appears to be
specimen and means calculated. The ball rebound values at displaced midway between the lowest and highest cushion
standard conditions were considered to be the base values and the displacements. The composite cushion (PU & fibre) shows as
results for the other test conditions normalized to these values m uch displacement as some o f the all-foam cushions as RH is
(Table 4). After conditioning at low temperatures, most o f these increased.
foams had lower resiliency. At 50% RH, ball rebound values
increase with increasing conditioning temperature, Figure 5. The As expected, at very low temperature (-30°C, 5% RH), cushion
TDI- based foam shows a decrease between the 50 and 70°C displacements decrease considerably, i.e. to between 60 and 83%
conditions. o f the base displacements.

Specimens held at 23 ° C over a wide range o f relative

humidities, Figure 6, exhibit quite different ball rebound
responses. TDI- base foam decreases in resiliency from 120 at
20% RH to 100 at 50% RH, then increases again to 110 at 70 - 90%
RH, MDI - based foams increase in resiliency from 95 at 20% RH,
to 100 at 50% and then decrease again to about 94 at 70 - 90% RH

4.2 Production Cushions

To obtain the hysteresis loss values for each cushion/seat pan

combination, an applied force of7 0 0 N was used before unloading,
Figure 7. All five seat loading/unloading traces are shown for the
standard conditions, 23°C, 50% RH. The three cushions
supported on metal pans behave quite similarly whereas the
cushion supported on sinusoidal springs (No. 3) exhibits a much
greater hysteresis loss. The polyurethane/coconut fibre
combination has a very narrow loss envelope and the loading
stress builds up later than the all-foam cushions. The hysteresis
loss values for each cushion type at 23 °C, 50% RH are the norm
or base values, see Table 5. The actual thickness values recorded
after each conditioning treatment are recorded in the table. The
polyurethane/coconut fibre combination exhibits the greatest
changes in thickness (5 to 6%) at very low and at high
temperatures/high humidities. In m ost cases the polyurethane
cushions only change in thickness by 1-2%.

At 50% RH, all cushions except no. 2 show decreases in

hysteresis loss with increasing test temperature, Figure 8.
However, above 50°C, there is little change in hysteresis loss for
the other four cushions.

Cushions tested at 23 ° C and various relative humidities, Figure

9, exhibit increases in hysteresis loss as RH is increased.
Cushions 1 and 3 only show a 1-2% increase in hysteresis loss,
cushions 4 and 5 increases o f 7 -10% and cushion no. 2 almost a
15% increase. At - 30 ° C, large increases (21 to 88%), in
hysteresis loss were recorded, Table 5.

Immediately following the hysteresis loss determination, these

seats were compressed to a maximum force o f 500N and the
resultant displacement o f the cushion after 60 ± 3 seconds
determined. The calculated values are found in Table 5. In
Figures 10 and 11 it can be seen that cushion displacements
increase both with temperature increase (RH held constant) or
with increasing relative humidity (temperature held at 23°C),
respectively.

290 / Blair, Russ, Bradt, Bailey

 TABLE 4 Normalized Ball Rebound

| T e m p eratu re °c
Relative
-3 0 10 23 40 50 70 90
Humidity, %

A 48 110

5 B 47 110

C 59 115

A 110

10 B

A 114 120 108 130 138

20 B 90 95

C 96 94

A 86 100 122 126 113

50 B 90 100 104 104 103

C 96 100 106 108 10 7

A 110 120 124 122

70 B 93 103 110

C 96 106 111

A 110 100 104 112 110

90 B 93 97 98
C 94 107 10 7

Ball Rebound results at 2 3 ° C , 5 0 % R H are considered to be at 1 0 0 all others relative to these.

A = S A N /T D I
B =M D I 85I
C =M D I 95I

FIGURE 5

Blair, Russ, Bradt, Bailey / 291

 FIGURE 6

FIGURE 7

292 / Blair, Russ, Bradt, Bailey

 TABLE 5 Seat Performance Under Loading/Unfoading Forces

A
-30 23 50 70
H Seat Initial Normal Displace­ Initial | Normal Displace Initial Normal Displace Initial Thickness, mm Normal Displace­
# Thickness, Hysteresis ment Thickness, Hysteresis -ment Thickness, Hysteresis -ment Hysteresis ment
mm Loss @ 500N mm ! Loss @ SOON ! mm Loss (5) SOON Loss (d) SOON
1 96.9 187.7 71.4
2 84.6 171.7 82.6
3 75.3 121.0 69.4
4 59.9 151.8 71.7
5 85.9 151.8 60.0

1 93.2 100 100 92.7 95.9 107.8 93.1 96.4 112.9

2 85.8 100 100 85.6 102.7 101.5 87.8 104.6 105.9
3 76.2 100 100 74.6 93.0 102.7 75.1 91.9 109.0
4 63.6 100 100 61.9 92.7 103.3 61.2 93.1 106.7
5 89.1 100 100 87.6 96.9 106.3 87.6 97.3 110.5

1 92.8 97.7 109.0 92.1 98.6 113.5 93.1 100.5 118.6

2 85.6 105.0 100.7 85.9 107.8 105.4 87.3 107.8 110.4
3 76.3 98.6 107.9 74.6 93.3 107.7 75.4 95.8 110.8
4 63.5 99.1 105.8 60.4 97.3 105.5 60.7 97.2 112.5
5 87.3 102.3 109.8 86.4 96.9 111.2 87.2 98.4 113.6

1 93.7 102.7 120.1 92.8 103.6 123.3 93.5 110.0 130.2

2 I 86.1 114.6 110.4 86.2 I 117.4 I 109.6 88.2 121.5 116.1
3 76.8 101.1 115.7 75.5 98.9 115.5 72.3 88.0 113.3
4 63.5 109.6 115.5 63.6 107.8 119.5 60.5 117.0 120.1
5 88.2 107.0 120.1 87.2 101.6 | 120.8 85.4 104.3 124.1

Blair, Russ, Bradt, Bailey / 293

 FIGURE 8

FIGURE 9

294 / Blair, Russ, Bradt, Bailey

FIGURE 10

FIGURE 11

Blair, Russ, Bradt, Bailey / 295

43 Vibrational Transmissivity
support also has a relatively low natural frequency (4.12 Hz).
Previously, we along with others (14-17) have reported on the Cushions 1, 2 and 3 all have transmissivities near 4.8-4.9 but the
vibrational transmissivity o f polyurethane cushions. Our testing other deadpan seat, No. 5 has a much higher A/Ao value o f 6.46.
was done at 23 ± 2°C, 50 ± 5% RH. In the trials reported here, we
have expanded the testing envelope by measuring vibrational The composite seat has the worst transmissivity values (i.e.
characteristics o f cushions alone and combined with their frames. high resonance frequency 5.56 Hz and high A/Ao value 6.41).
In addition, w e have tested cushions plus frames over a range o f
temperatures (23 to 70°C) and relative humidities (50 to 80%) to Cushions removed from their frames and retested (see Fig. 14)
determine how the vibrational characteristics responded to these do not exhibit drastic changes in resonance frequencies when
changes in climatic conditions. For one seat, cushion number 4, tested alone, Table 6B. The greatest change is actually an increase
we have additionally tested the frame plus the polyurethane o f 0.60 Hz for cushion No.3. Only minor
cushion only, i.e. the coconut topper pad was removed and the changes were found for the other seats. Similarly, small changes
response o f the PU cushion m easured This was done to separate (increase, decrease or zero) were found for transmissivity values
the vibrational responses o f the PU cushion from the fibre p a d (A/Ao).

The vibrational responses o f the five seats are tabulated in As an example, a deadpan seat No. 2 does not exhibit large
Table 6A, where the resonance or natural frequency, transmissivity changes with changes in climatic conditions, Fig.
transm issivity at resonance (A/Ao), transmissivity at the important 15. Increasing temperature and RH makes the foam more resilient
6 Hz frequency and the frequency where A/Ao (the output divided and may account for the shift in resonance frequency to a lower
by the input amplitude equals unity) for four climatic conditions. value but this is accompanied by an increase in transmissivity.
In addition, in Table 6B, the same characteristics for the cushions Previously (14 - 16) we have observed that the more resilient the
tested alone are recorded. Note that for cushion number 4, the PU foam, the lower the natural frequency but transmissivity usually
cushion plus the coconut fibre topper pad was measured. In Table increases in magnitude.
6C, results for cushion number 4 are recorded for the frame + PU
+ fibre, frame + PU only, PU + coconut fibre, PU foam only. This The composite seat, No. 4 also shows the same changes but the
allows us to examine the relative transmissivity contributions from magnitudes accompanying temperature changes are greater, Fig.
each component o f this composite seat. 16.

Since we have measured transmissivity at 23, 50 and 70 °C all at In Fig. 17, we note that the natural frequency values decrease
50% RH, we can plot the vibrational characteristics versus seat with increased cushion thickness but we did not find any
conditioning temperature. The natural frequency for all seats correlation between transmissivity values and cushion density.
decreases w ith increasing temperature (Fig. 12A). The all- PU
foam seats on their frames have much lower resonance frequencies 4.4 Rebound
over the temperature range tested than the composite seat. The
lowest natural frequency found was for cushion No. 2 (deadpan We have also measured the resiliency values for these seats
construction). using a modified ASTM D3574 -95 Test H ball rebound method.
Cushions were tested for resiliency by positioning the rebound
I f the test conditions are changed from the laboratory standard equipment on the cushion IFD area suiface, i.e. we m easured the
(23 ± 2°C, 50 ± 5% RH) to tropical conditions such as called for in cushion (skin + core) resiliency with the support frame in place.
BMW QV 52009 part 1 specification (40°C, 80% RH), then the The BR values are found in Table 7. W e measured the BR values
natural frequency decreases even more than by increasing for cushion No. 4 as a composite (foam + fibre) and foam alone.
temperature only, Fig. 12A. When examining these BR values, it must be borne in mind that
normally resiliency is measured on core foam, not on cushion skin
The peak transmissivity A/Ao, Fig. 12B for cushions 1, 2 and 3 or on full cushions.
shows almost no change with temperature increase, cushion No. 5
exhibits decreasing transmissivity but the composite cushion, No. In Fig. 18,seat natural frequency is plotted against the ball
4 gives an increase in A/Ao as temperature is increased from 23 to rebound values. Natural frequencies decrease with increasing BR
70 °C while the RH is held at 50%. At 40°C, 80% RH, all PU percentage. Removal o f the fibre topper pad from cushion No. 4
seating either exhibits a decrease in transmissivity or no change drastically increases BR value (24 to 54%) and as noted
whereas the composite cushion again has a greater A/Ao value previously, natural frequency also decreases (5.56 to 4.35 Hz).
than the all- foam seats, Figure 12B.
4.5 F ord Ball R ebound
The transmissivity values at 6 Hz, Fig.l2C , all show decreases
with increasing temperature. The foam cushions do not exhibit as The Ford Motor Company specifications ESB -M2D221/A/B/C
much decrease as the composite cushion, but the latter has much calls for a more aggressive ball rebound measurement. In their
higher A/Ao values at 6 Hz over the entire range than the foam - test a much heavier steel ball (535 gms, 50.8 mm diameter) is
only seats. dropped from a height o f 1270 mm onto a sample with a thickness
The actual vibrational transmissivity curves for the five seats o f 100 mm. Contrast this with the ASTM D3574 - 95 standard ball
(frame + cushions) at 23°C, 50% RH are shown in Fig. 13. The rebound which calls for a 1 gram ball to be dropped from a height
full foam constructions (also known as deadpan) have resonance o f only 500 mm on to a core foam sample o f 50 mm in thickness.
freauencies o f 4.15 Hz or lower but the seat with sinusoidal soring The rebound height for the Ford method in millimetres is recorded.

296 / Blair, Russ, Bradt, Bailey

W e have used this equipment but measured the resiliency o f the seating foam presumably in the form o f heat dissipated to the
five cushions in their frames by dropping the ball on the IFD area surrounding air and/or to the seat frame.
from various heights between 1270 and 760 mm.
A plot o f the BR results (calculated as percent rebound from the During road travel, vibrations from the road surface are
initial drop height) versus drop height is shown in Fig. 19. transmitted via the wheels and suspension to the floor pan and then
to the seat structure. These transmitted vibrations if not absorbed
Note that the resiliency values increase with decreasing drop or reduced by the seat structure, i.e. mostly by the moulded
height for all seats, i.e. the amount o f energy absorbed by the seat polyurethane foam components, cover and seating construction,
decreases as the drop height is decreased. e.g. spring - type suspension, will be transmitted to the seat
occupant. These vibrations m ust be attenuated as much as
The ball rebound values increase with seat thickness, Fig. 20 possible by the seat construction and the energy dissipated as heat.
irrespective o f seat construction, i.e. it does not appear to matter if Since both vibrational and im pact energy are dissipated as heat,
the cushion sits on a deadpan, wire field or sinusoidal springs. there may be a correlation between vibrational seat characteristics
and impact absorption.
Ball Rebound measures the amount o f energy returned to the
ball after im pact In Japanese terms, resiliency is ta m e d Impact In Fig. 21, we show the relationship between seat natural
Resistance. frequency and the amount o f absorbed ball rebound energy. The
relationship is reasonably linear with cushions that absorb lower
The converse o f energy returned to the impacting ball is energy amounts o f energy having the lowest natural frequencies.
absorbed during impact. This energy m ust be absorbed by the

Blair, Russ, Bradt, Bailey / 297

298 / Blair, Russ, Bradt, Bailey
Blair, Russ, Bradt, Bailey / 299
300 / Blair, Russ, Bradt, Bailey
Blair, Russ, Bradt, Bailey / 301
3 0 2 / Blair, Russ, Bradt, Bailey
 TABLE 6 Vibrational Transmissivity of Seats

(Cushions + Seat Frames)

Natural Transmissivity @ Frequency @ A/Ao=1

Seat#
Type Condition Frequency A/Ao 6 Hz (A/Ao)
Hz

A- 23°C/50% 1 3.62 4.81 0.93 5.90

Seat RH 2 3.35 4.76 0.64 5.14
Cushions & 3 4.12 4.91 1.58 5.05
Frames 4 5.56 6.41 4.68 7.99
5 4.15 6.46 1.86 7.05
40°C/80% 1 3.26 4.23 0.69 5.14
RH 2 3.14 4.78 0.61 - 5 .0
3 3.76 5.18 1.01 6.01
4 4.96 6.81 2.82 3.41
5 3.66 5.44 1.08 6.08
50°C/50% 1 3.54 5.29 0.88 5 .8
RH 2 3.16 5.65 0.56 4.87
3 3.83 5.39 1.18 6.35
4 5.08 7.64 3.21 4.32
5 3.88 5.70 1.36 6.72
70°C/50% 1 3.29 4.97 0.66 5.28
RH 2 3.04 5.21 0.49 4.72
3 3.60 5.44 0.96 6.00
4 4.84 8.77 3.00 6.32
5 3.65 5.49 1.06 6.16
B- 23°C/50% 1 3.37 4.62 0.74 5.13
Cushions RH 2 3.11 5.52 0.86 5.56
Only 3 4.72 4.15 2.03 7.70
4 5.73 5.85 5.24 5.61
* 5.50 4.49 3.60 8.22
5 3.81 4.81 0.91 5.84
C- Frame + PU 4 5.56 6.41 4.68 7.99
Seat #4 + Fibre
Frame + PU 4 4.35 9.50 2.06 - 7 .0
PU + Fibre 4 5.73 5.85 5.24 5.61
PU 4 5.50 4.49 3.60 8.22

TABLE 7 Ball Rebound Resiliency O f Cushions On Frames*

Seat Number Ball Rebound, %

1 52
2 50
3 42
4A* 24
4B** 54
5 47

• Modified ASTM D 3574 95 (Test H) Ball Rebound on cushion

‘A ’ surface on seat frame
*Seat 4A - Foam + Coconut fibre on seat frame
**Seat 4B Foam only on seat frame

Blair, Russ, Bradt, Bailey / 303

 TABLE 8 Dynamic M echanical Analysis R esults For Foam Blocks

Modulus Change
Foam Type TgSOft, °C TQhard* C Modulus kPa kPa
23-> 23->
at 23°C at 60°C at 90°C 60°C 90°C
TDI, 100 -44.9 128.5 27.5 23.2 19.4 -4.3 -8.1
MDI, 85 -45.3 129.7 15.3 23.2 42.8 7.9 27.5
MDI, 95 -42.2 127.3 103.5 108.9 103.1 5.4 -0.4

3 0 4 / Blair, Russ, Bradt, Bailey

 Since the cushions were manufactured for OEMs headquartered
4.6 D ynam ic M echanical A nalysis in Europe, Japan and North America, we tested each o f them to the
ASTM D3574 -95 standard but included some tests called for in
W e have subjected both block and cushion foams to dynamic the JIS K6400 and German DIN EN ISO 1923:1995 standard.
mechanical analysis (DMA) using a Rheometric Scientific DMTA Once again, the properties tabulated in Table 11 should be
IV analyzer in a Dynamic Temperature Ramp test at a frequency acceptable to m ost OEMs around the world. O f course it must be
o f 1 Hz over the temperature range - 7 5 to - 30 ° C at 2.5 °C/min and borne in m ind that a cushion for a specific OEM has been poured
- 3 0 to 200 ° C at 2.5 °C/min. A typical trace is shown in Figure 22 w ith an optimized formulation to satisfy that customer ’s hardness
for MDI - based foam at 95 index. Table 8 summarizes the results and property requirements.
for the three foam blocks. All foams have about the same soft
segment glass transition ( T g ^ ) temperature (minus 40- 45 ° C) and Hence not all o f the properties listed in the table may be
a hard segment, T g^,* at 127-130°C. This latter Tg is due to the acceptable to another O E M With our optimized formulating
SAN particles present from the copolymer polyol softening. At capabilities, we are able to fine tune formulations to satisfy a wide
23°C, the 85 index MDI based foam has a modulus o f 15.3 kPa. range o f requirements for the OEMs around the w orld
TDI- based foam at 1001 has a higher modulus at 27.5 kPa and the
951 MDI - based foam a m uch higher room temperature modulus at 5.0 Conclusions
103.5 kPa.
By conditioning foams under various controlled temperatures
and relative humidities simulating a very wide range o f climatic
As test temperature is increased over the range 23 to 60°C, TDI
conditions prevalent around the world, we have been able to
foam exhibits a modulus decrease o f 4.3 kPa. The 951 MDI foam
measure foam property responses. Increasing temperature or
increases in modulus over the same range by 5.4 kPa and the low
relative humidity or both simultaneously from the controlled
index MDI - foam shows an increase in modulus o f 7.9 kPa. Over
conditions normally present in a quality testing laboratory (23 ±
the same temperature range, we used in our hardness/resiliency
2 °C, 50 ± 5% RH), results in polymer softening. This results in
study, i.e. 23 to 90°C, the TDI, 851 MDI and 951 MDI foams
foam hardness decreases. For example:
exhibit modulus changes o f - 8 .1 , +27.5 and almost zero,
i) TDI - based foam held at 23°C will exhibit a hardness
respectively. Almost certainly the large modulus increase o f the
decrease o f about lN /% change in RH over the range 23 to
lower index MDI foam is due to progressive polymer cure as
temperature is increased. 70% RH
ii) The same foam held at 50% RH will decrease in hardness
These changes in polymer stiffness with increasing temperature about 0.7N/C degree over the range 23 to 85°C.
will occur in foams conditioned at elevated temperatures prior to iii) MDI -based foam (951) held at 23°C, will decrease in
hardness or resiliency testing and w ill contribute in some degree to hardness about 2N/% RH increase over the range 50 to
the changes we have reported earlier. 95% RH but only 1N/C degree at 50% RH ova: the
temperature range 23 to 70°C.
For the five TDI- based cushions, the DM A results are found in
Table 9. Since these are all SAN - copolymer based foams ware For the five cushions examined in this paper, the cushion
run at about a TDI index o f 100, it would be expected that they supported on sinusoidal springs had the highest hysteresis loss
would have similar soft and hard segment Tgs. This is indeed the presumably due to energy being lost to the cushion and the flexing
case with the Tg** and T g ^ temperature ranges being minus 38.9 springs.
to 41.7 and 139.1 to 143.4°C respectively. At 23°C, there is quite
a wide range in modulus, i.e. 40.8 to 95.4 kPa for seats 4 and 1 The foam/coconut fibre seat had the lowest hysteresis loss but
respectively. Although the formulations used are proprietary, we this may not be a positive attribute since an occupant may
can indicate that the seat with the lowest modulus was made with experience some degree o f seat discomfort as the low hysteresis
the lowest percentage water in the formulation. The higher loss may result in vibrations from the road surface being
modulus foams all have higher percentage water contents in their transmitted to the occupant causing motion sickness.
formulations. Since these cushions were all run at about 100
index, the isocyanate level should not contribute to modulus Cushion seat displacement (at 500N loading) appears to be
differences. somewhat dependent on cushion core density irrespective o f test
temperature or % RH or seat construction.
All foams show a decrease in modulus as temperature is
increased but the range o f decrease is quite wide from almost zero Seat vibration transmissivity measured in the laboratory is an
for seat 4 to - 27.3 for seat 1. W e have not been able to obtain any important guide to in -vehicle seat comfort. The five seats (cushion
correlation between the modulus changes and foam composition or plus frame) tested here all have different transmissivity responses.
density at this time. As mentioned previously, these modulus Thick cushions sitting on metal pans have better vibrational
changes will contribute to foam hardness/resiliency changes after characteristics than a thin cushion with a coconut fibre topper pad.
cushion preparation under various climatic conditions. Increasing conditioning temperature while holding die RH
constant at 50% results in a lowering o f the natural frequency of
4.7 Foam Physical Properties the seat. If the conditions simulating tropical conditions are used,
then even lower resonance frequencies are obtained. It is
We measured a selected number o f foam properties for the test postulated that the lowering o f natural frequency at higher
blocks, see Table 10. These foams have physical properties that temperature and/or RH is due to polymer softening.
would satisfy most OEM specifications.

Blair, Russ, Bradt, Bailey / 305

 The peak transmissivity (A/Ao at resonance) is not significantly
affected by temperature increase for full foam seating but the
foam/fibre seat exhibits higher transmissivity values as the test
temperature is increased. This will result in increased seat
discomfort.

We have also tested cushions removed from their frame

supports. They exhibit similar transmissivities to those found for
cushion plus frame. Therefore testing o f cushions alone should be
sufficient to guide seating engineers/designers as to the vibrational
responses expected for cushions and frames. Thus, testing o f seats
(frame & cushion) may not be necessary and hence testing is less
complicated.

Thicker cushions have lower resonance frequencies but we did

not find any correlation with cushion core density for these five
cushions tested here. However, foams with higher resiliency
(measured as ball rebound from the cushion surface) have lower
natural frequencies.

Composite cushioning such as measured in this study, does not

behave as well as foam - only cushioning. The fibre topper pad
seems to adversely affect the transmissivity responses and this
m ay result in an uncomfortable ride, especially on roads with
uneven surfaces. Thus, we would recommend the use o f all-foam
cushioning to enhance occupant -seating comfort on all types o f
road surfaces from smooth highways to rough unsurfaced roads.

W e believe that the analyses presented in this paper will be a

useful guide for seat designers and test engineers. Knowledge o f
seat hardness and vibrational transmissivity response to changes in
climatic conditions should help seat engineers to design more
comfortable seating.

306 / Blair, Russ, Bradt, Bailey

 TABLE 9 Dynamic M echanical A nalysis R esults For Foam Cushions

Modulus Change
Cushion T gSoft, ° C Tghart, °c Modulus kPa KPa
23 ->90 °C
at 23°C at 60°C at 90°C
1 -41.2 139.1 95.4 77.5 68.1 -27.3
2 -38.9 142.1 50.5 48.3 43.5 -7.0
3 -40.6 141.6 75.4 53.7 50.6 -24.8
4 -41.7 143.4 40.8 40.0 40.2 -0.6
5 -39.1 140.0 65.8 60.9 60.3 -5.5

TABLE 10 Block Foam Physicals

Property Units Test Methods Foam Type

TDI MDI 851 MDI 951
Core Density kg/m3 ASTM D3574 -95 (A) 50 50 50

Tensile Strength kPa ASTM D3574 -95 (E) 200 150 165

Elongation % ASTM D3574-95 (E) 120 140 140

Tear Strength N/m ASTM D3574 -95 (F) 230 205 215

50% Compression set % ASTM D3574 -95 (D) 5 18 17

50% humid aged

Compression Set % ASTM D3574-95 (D, J2) 10 16 16

TABLE 11 Cushion Physical Properties

Property Units Test Method Cushion No.

1 2 3 4 5
Core Density kg/m'5 ASTM D3574 -95 (A) 53 58 40 70 48

Tensile Strength kPa ASTM D3574 -95 (E) 200 140 200 220 210

Elongation % ASTM D3574-95 (E) 125 120 130 150 110

Tear Strength N/m ASTM D3574 -95 (F) 190 165 230 160 200

50% Compression % ASTM D3574-95 (D) 11 10 16 11 11

Set

50% humid aged % ASTM D3574 -95 (D, Ji) 16 15 19 12 14

Compression Set

50% Wet Set % JIS K 6400: 1997 10 7 12 7 13

DIN Tear N/m DIN 53356 210 175 250 345 285

Die aC" Tear kgf/cm JIS K6400: 1997 0.71 0.71 0.82 1.03 0.83

Blair, Russ, Bradt, Bailey / 307

BIOGRAPHIES ACKNOWLEDGEMENTS

G. Ron Blair W e would like to acknowledge the contributions made by

Ron Blair received his degrees in Glasgow Scotland. He Harbans Dhariwal o f the Canadian Automotive Workers who
continued his studies at the University o f British Columbia before tested all the samples for firmness after conditioning; Our
joining Royal Dutch Shell Plastics Laboratory in Holland. During W oodbridge colleagues who supplied us with foam blocks
his six years with Shell, he worked in various functions including (George Ng) or cushions (plants in Brazil) for testing; Carlos
fundamental research, plastics testing and latterly polyurethanes. Rochini and Darrell Robertson who co - ordinated part procurement
In 1976, he joined Monsanto Canada, which became W oodbridge and shipment. W e gratefully acknowledge the DM A work
Foam Corporation in 1978. He has held various processing and supervised by Jing Zhang. W e are especially grateful to Bill gill
chemistry positions and is currently managing the W oodbridge o f Bayer U.S.A. who arranged the vibrational transmissivity
Corporate Quality Laboratory, Woodbridge, Ontario, Canada. testing at his company ’s R&D facility in South Charleston West
Virginia. Special mention is made concerning Kathy Wood who
co - ordinated manuscript preparation and collation.
A. Russ
Alex Russ obtained her BSc degree in Statistics from the We also acknowledge our W oodbridge colleagues ’
University o f Toronto, Canada. She currently holds the position o f encouragement and support and our company ’s permission to
Statistical Methods Specialist, responsible for the application o f pursue this research and allow publication.
statistical methods in research and development
References

D.E. Bradt 1. J.M. Hogan, C.J. Pearson, T.H. Rogers and J.R White,
David Bradt received his Chemical Engineering Diploma from “Humidity Characteristics o f H R . and Hot Urethane
Niagara College o f Applied Arts and Technology. Foams ” , J. Cell Plastics, Sept./Oct. 1973, pp 219 225

During his ten years with Shell Canada, he worked as a process 2. R M . Herrington, and D.L. Klarfeld. 1980. “Humid Aged
technician, and a Coordinator. In 1993, he joined the W oodbridge Compression Set Phenomena in All W ater Blown H .R
Group. He held the position o f Chemical Technician in the Molded Foams ” , Proceedings o f the SPI Annual
Corporate R&D departm ent He currently works in the Corporate Technical/Marketing Conference, p.177
Quality Laboratory as a Laboratory Supervisor, responsible for co ­
ordinating and supervising foam/composite testing from 3. M .R Kinkelaar, K.D. Cavender, J.L. Lambach, R D .
W oodbridge plants, worldwide. Brasington and F.E. Critchfield. U S. Patent 5549841, 1996
August 27. “Process for Manufacturing Foam with
Improved Wet Set Properties ” .
R J. Bailey
Robert J. Bailey is a graduate o f Mohawk College o f Applied Arts 4. J.C. Moreland, G.K. W ilkes and R B . Turner, “Viscoelastic
and Technology ’s three year co operative education Mechanical Behaviour o f Flexible Slabstock Foams as a Function of
engineering Technology program and an Associate member o f the Temperature and Relative Humidity ” , Proceedings o f the
Ontario Association o f Certified Engineering Technicians and SPI PU Conference, Technomic, Lancaster, P.A .1991; pp
Technologists. Robert held 500-508.
the position o f Junior Product Engineer at Morval when it was
acquired by W oodbridge Foam in April 2000, and currently holds 5. K.D. Cavender and M .R Kinkelaar, “Load Bearing
the position o f Senior Quality Specialist at W oodbridge’s Response o f H .R Molded Foam as a Function of
Corporate Quality Laboratory. Processing and Testing Parameters ” , Proceedings o f the
SPI PU Conference, Technomic, Lancaster, P A 1994; pp
248 -251.

6. M .R Kinkelaar, K.D. Cavender and R Brasington,

“Advances in Molded Foam Durability Performance and
Testing:, Proceedings o f the SPI PU Conference,
Technomic, Lancaster, P A 1995; pp 380 -388.

7. R Brasington, M .R Kinkelaar and K.D. Cavender, “New

Polyols for Improved Automotive Seating Durability
Performance ” , Conference Papers, UTECH 96, Crain
Communications: London, 1996, Paper 9.

8. F.M. Casati, R M . Herrington, R Broos, Y. Miyazaki, H.

Phan Thanh and D. Cadolle, “Automotive Seating Foams;
Improving Their Comfort and Durability Performance”,
SPI Conference, Technomic, Lancaster, P A 1998; pp 417 -
431.

308 / Blair, Russ, Bradt, Bailey

9. F.M. Casati, P.R. Berthevas, R.M. Herrington and Y.
Miyazaki, “The Contribution o f Molded Polyurethane 14. G.R. Blair, R. So, A. Milivojevich, J.D. van Heumen, 1998,
Foam Characteristics to Comfort and Durability o f Car “Automotive Seating Comfort; Investigating the
Seats” , SAE International Congress and Exposition, Polyurethane Contribution Phase 1” . SAE Technical P ap a
Detroit, Michigan, 1999 P a p a No. 1999-01-0585; SAE, 980656.
Warrendale, PA.
15. J.D. van Heumen and G.R. Blair, 1998, “Polyurethane
10. G.R. Blair and A.R. Wilson, Polyurethane Automotive Seating Comfort; Analysis o f Foam Formulation
Cushioning; Car Durability vs. Foam Properties, 35th Contribution ” . Polyurethane Expo *98, Dallas Texas, pp.
Annual Polyurethane Technical/Marketing Conference Oct. 507-515.
9- 12 1994, Boston, MA. pp. 478 ^8 8 .
16. G.R. Blair, A. Milivojevich, J.-G. Pageau and J.D. van
11. G.R. Blair and A.R. Wilson, Polyurethane Automotive Heumen, 1999, “Automotive Seating Comfort; Defining
Cushioning; Material Properties After In -Car and Comfort Properties in Polyurethane Foam ” . SAE
Simulated Durability Testing, 36th Annual Polyurethane Technical P ap a, 1999-01-0587.
Technical/Marketing Conference, Sept. 26 - 29, 1995,
Chicago, IL. pp. 413 -417. 17. Milivojevich, R. Stanciu, A. Russ, G.R. Blair, and J.D. van
Heumen, “Investigating Psychometric and Body Pressure
12. G.R. Blair and R.J. Horn, Fleet Durability Testing o f Distribution Responses to Automotive Seating Comfort” ,
Moulded Polyurethane Foam and Competitive Automotive SAE Technical P a p a , 2000-01-0626.
Cushions, Utech 96, March 26 - 28,1996, p a p a 5.

13. A.W. D uff and P.R. Berthevas, Dow Chemical, Meynin,

Switzerland, 1991, Private Communication.

Blair, Russ, Bradt, Bailey / 309

 SESSION H
BLOWING AGENTS

TECHNICAL SESSION ORGANIZERS

Stan Moore, Chair

THE DOW CHEMICAL COMPANY

George Gula
CANNON USA

TECHNICAL SESSION MODERATOR

Craig Sungail
HUNTSMAN POLYURETHANES

TECHNICAL SESSION VICE - MODERATOR

Stan Moore
THE DOW CHEMICAL COMPANY
Aging of Polyurethane Foam Insulation in Simulated
Refreigerator Panels— Three-Year Results with
Third-Generation Blowing Agents

KENNETH E. WILKES, DAVID W. J . RICHARD BOOTH

YARBROUGH, W. A. GABBARD Tennessee Technological University
a n d GARY E. NELSON Department o f Chemical Engineering
Oak Ridge National Laboratory Cookeville, Tennessee 38505
P.O. Box 2008
Oak Ridge, Tennessee 37831-6092

and cyclopentane. Tests with these third-generation blowing

ABSTRACT
agents started in the fall of 1997 and are planned to continue
for the next few years. Results on full-thickness specimens
Laboratory data are presented on the effect of constant-
after aging at controlled temperatures for six months, one
temperature aging on the apparent thermal conductivity of
year, and two years have been reported previously [3-5]. The
polyurethane foam insulation for refrigerators and freezers.
purpose of this paper is to present an update on this study by
The foam specimens were blown with HCFC-141b and with
including results after three years of aging.
three of its potential replacements — HFC-134a, HFC-245fa,
and cyclopentane. Specimens were aged at constant
SPECIMENS
temperatures of 90°F, 40°F, and -10°F. Thermal conductivity
measurements were made on two types of specimens: full­
Two types of specimens are being studied. Specimens of
thickness simulated refrigerator panels containing foam
one type were fabricated as panels that simulate the
enclosed between solid plastic sheets, and thin slices of core
construction of a door or wall of a refrigerator. The panels
foam cut from similar panels. Results are presented for the
are about two inches thick and have lateral dimensions of 24
first three years of a multi-year aging study. Preliminary
inches by 24 inches. The faces of the panels are bounded on
comparisons of measured data with predictions of a
both sides by 0.04 in. thick plastic sheets, with separate sets
mathematical aging model are presented.
of panels being made with acrylonitrile-butadiene-styrene
(ABS) and high-impact-polystyrene (HIPS) plastic. The 0.04
INTRODUCTION
in. thickness was considered to be representative of current
refrigerator production, and plastic was used on both sides to
Polyurethane foam insulation currently used in refrigerators
accelerate the aging experiments by allowing gases to
and freezers in the United States is blown primarily with
permeate through both sides of the panels. It was felt that the
HCFC-141b. Because of its potential for depleting the ozone
effect of an impermeable steel sheet on one side could be
layer, the Montreal Protocol mandates that domestic
simulated using models that are being developed. The edges
production of HCFC-141b cease by the end of 2002,
of the panels were sealed with aluminum foil tape to simulate
requiring that a replacement blowing agent be used after that
the configuration in a refrigerator where there are no
time. Since 1993, the Oak Ridge National Laboratory has
comparable cut exposed foam edges.
been cooperating with the Appliance Research Consortium on
The panels were foamed with HCFC-141b (to provide the
studies of the aging characteristics of polyurethane insulation
base case), HFC-245fa, HFC-134a, and cyclopentane. At the
foamed with various blowing agents. The early studies were
present time, the latter two blowing agents are the only
aimed at the replacement for CFC-11, and results of aging
commercially available non-ozone-depleting replacements for
studies on simulated refrigerator panels blown with CFC-11,
HCFC-141b. HFC-245fa is currently available on a semi­
HCFC-141b, and a blend of HCFC-142b and HCFC-22 have
commercial scale. Commitments have been made to
been reported previously [1,2]. A long-term study is
construct a plant to produce HFC-245fa, and commercial
currently underway on aging of foam blown with HCFC-141b
availability is expected in mid-2002. The panels were
and several potential replacements — HFC-134a, HFC-245fa,

POLYURETHANES EXPO 2001, SEPTEMBER 30 - OCTOBER 3, 2001 313

fabricated by four foam suppliers (denoted here as suppliers core-foam specimens that could be studied, and core-foam
A, B, C, and D), with each supplier providing panels with measurements were concentrated on specimens from Supplier
each of the four blowing agents. A.
To provide a characterization of the foam itself, specimens
were also made that consisted of core foam cut into 12 in. RESULTS AND DISCUSSION
squares and sliced into thicknesses of about 0.4 and 1.5
inches. A stack of four of the 0.4-in.-thick slices made up one Core - Foam Specimens
test specimen.
It should be noted that the foams made in 1997 with the The thermal conductivity of freshly sliced foam at 75 °F is
alternative blowing agents may not have been optimized for shown in Table 1 for the two slice thicknesses. This shows
thermal performance. Visual observations of the sliced that the lowest thermal conductivity was found with HCFC-
specimens showed a very uniform fine cellular structure with 141b, followed by HFC-245fa (4% higher), cyclopentane
HCFC-141b but the presence of many larger bubbles (13% higher), and HFC-134a (21% higher). This relative
interspersed within the fine cellular structure was observed ranking is the same as has been observed by Haworth [7].
for the other blowing agents. Table 1 also shows that the thermal conductivity of thinner
slices was consistently 3 to 4% higher than that of the thicker
EXPERIM ENTAL PROCEDURES slices. This is at least partly due to the larger amount of
damaged surface layers with the thinner slices where air
Thermal resistance measurements were made using heat- immediately displaces the blowing agent in the cut cells.
flow-meter-apparatuses (HFMAs) that conform to ASTM C While the data for the thicker slices are more representative
518 [6]. Intervening layers of foam rubber were placed of that in the full-thickaess panels, the thinner specimens age
between the panel specimens and the hot and cold plates of much more rapidly and are useful for comparing potential
the apparatus to eliminate any undesirable air gaps between differences in long-term performance with the different
the specimens and the plates and also to protect the plates blowing agents.
from the rigid test panels. Thermocouples were taped Curves of thermal conductivity versus time for the thinner
directly to the faces of the panels so that the temperature (0.4 in.-thick) slices have been presented previously [5].
differences across the test panels were measured directly. Figure 1 shows representative curves for aging at 40°F for the
Since the measurements gave the overall thermal resistance four blowing agents. The data were taken over a period of
of the center of the test panel, a small correction (< 1%) was about two years. Since numerous previous studies have
made for the thermal resistance of the plastic sheets to obtain shown that the aging of unfaced foam scales with the square
the apparent thermal conductivity of the foam insulation. of the thickness (see [3] and ASTM C 1303 [8]), the data are
Tests on the core foam specimens gave the apparent thermal plotted as thermal conductivity versus the aging time divided
conductivity directly. The interlaboratory precision of by the square of the slice thickness. With this scaling, data
thermal conductivity data obtained using HFMAs is about 2- obtained over a time period of two years on slices 0.4 in.
4% [6]. The precision is better for relative measurements thick are good predictors of the variation of the thermal
using the same apparatus. conductivity of unfaced 2 in.-thick specimens over a period
The specimens were stored in closed, constant temperature, of 50 years, well beyond the nominal refrigerator lifetime of
atmospheric pressure aging chambers between HFMA tests. 20 years.
Aging temperatures of 90°F, 40°F, and -10°F were used in The curves in Figure 1 show several trends. The curves
order to span most of the range of conditions to which the show an initial rapid increase in thermal conductivity which
foam would be exposed in a refrigerator application. is attributed to diffusion of air components into the cells of
Thermal tests were performed at 45°F and 75°F mean the foam, followed by a more gradual increase which is
temperatures, using a temperature difference of 40°F. attributed to diffusion of the blowing agent out of the cells.
Thermal measurements were made on the full-thickness Except for crossover of the curves for HCFC-141b and HFC-
panels soon after their fabrication and then at six-month 245fa, the relative ranking of blowing agents given in Table
intervals. Since the slices of core foam aged much more 1 is preserved throughout the aging process. Data obtained
rapidly than the full-thickness panels, more frequent thermal at the other aging temperatures showed that the time required
measurements were required. This limited the number of for a given change in thermal conductivity with aging at 40°F

Table 1. Thermal Conductivity of Freshly -Sliced Core -Foam Specimens at 75 ° F

Slice Thickness, in. HCFC-141b HFC-134a HFC-245fa Cyclopentane

0.4 0.132 0.160 0.138 0.150
1.5 0.128 0.155 0.132 0.145
Thermal conductivity units are Btu«in./h»ft2#°F. Each value represents an average over three specimens. Specimens are from
Supplier A.

314 / Wilkes, Yarbrough, Gabbard, Nelson, Booth

Figure 1. Aging of thin (0.4 in.-thick) core-foam specimens blown with third-generation blowing agents by Supplier A. Captions
show aging and testing temperatures.

Figure 2. Average normalized thermal conductivity for full-thickness panels before aging. Data are normalized with respect to
average for HCFC - 141b tests at 75 ° F before aging.

was roughly twice as long as for aging at 90°F, while the time panels that were blown with HCFC-141b, and then dividing
required at -10°F was about 10 times as long as at 90°F.[5] the individual measured thermal conductivity values by this
Data read from curves such as in Figure 1 and companion average. This normalization procedure was considered to be
data for the thicker (1.5 in.-thick) slices are compared with justified since the bias caused by the aluminum foil tape
data obtained on full-thickness panels in the next section. In should be nearly the same for all panels tested.
general, the data obtained on core-foam specimens may be Normalized results before aging and after aging for one and
used to estimate upper bounds for aging of the foam in the three years are given in Table 2. The results in the table give
full-thickness panels. an indication of the variation in thermal conductivity among
specimens from a particular supplier and among the four
Full- Thickness Panels different suppliers. For example, the panels blown with
HCFC-141b have pre-aging normalized conductivities at
Thermal measurements on 96 full-thickness panels have 75°F that range from 0.95 to 1.04, but for a given supplier the
been performed before and after aging under controlled variation is 2% or less. Similar levels of variation are seen
temperatures for three years. Data on the panels are reported for the other blowing agents.
as normalized values because the directly measured values Averaging the pre-aging data on the six panels for each
are biased due to the construction of the panels. From tests supplier and each blowing agent gives the comparisons shown
on standard reference materials, we have found that heat in Figure 2. On average, the 75°F thermal conductivity of
flows due to the aluminum foil tape around the panel edges foam blown with HFC -134a was about 18% higher than for
change the measured center-of-panel thermal conductivities HCFC-141b, while the conductivities for HFC-245fa and
by a few percent. The data were normalized by averaging the cyclopentane were about 7% and 15% higher. For Supplier
pre-aging thermal conductivities measured at 75 °F on the 24 A only, the average conductivities for HFC-134a, HFC-245fa,

Wilkes, Yarbrough, Gabbard, Nelson, Booth / 315

 Table 2. Results of thermal conductivity tests on full thickness panels before and after three years of aging at controlled temperature.
| Values are normalized to the average of the pre -aging results for HCFC - 141b at 75 ° F test temperature . |
□ i_ _ _ i_ _ _ z n
Aging □ z n
Blowing Temp., Tested at 75°F U Tested at 45°F |
Agent Liner °F Supplier A Supplier B Supplier C Supplier D Supplier A i Supplier B Supplier C Supplier D |
0 Yr. 1 Yr. 3 Yr. 0 Yr. 1 Yr. 3 Yr. 0 Yr. 1 Yr. 3 Yr. 0 Yr. 1 Yr. 3 Yr. □ 0 Yr. 1 Yr. 3 Yr. 0 Yr. 1 Yr. 3 Yr. 0 Yr. 1Yr. 3 Yr. 0 Yr. 1Yr. 3 Yr*
lHCFC-141 b ABS 90 1.01 1.06 1.17 0.95 TOO 1.03 1.09 1.21 1.04 1.09 1.22 0.94 0.97 1.07 0.90 0.92 1.05 0.93 0.98 1.11 0.95 0.99 112
|HCFC-141b IABS 40 1.01 1.04 1.09 0.95 1.00 1.09 1.01 1.07 1.17 1.03 1.08 1.15 □ j 0.94 0.93 0.99 , 0.91 0.90 0.99 0.92 0.95 1.05 0.95 0.96 1.04
|HCFC-141b ABS -10 1.01 1.02 1.03 0.95 0.96 0.97 1.02 1.06 ] 1.07 1.04 M.05 1.06 M 0.94 10.93 10.941 0.89 0.881 0.90 I0.93 0.93 0.95 0.95 0.94 0.95 |
IHCFC-141 b HIPS 90 1.01 | 1.13 1.25 0.95 1.06 1.16 1.01 L i Z 1.29 1.02 1.19 1.31 □ 0.94 j 1.04 | 1.15 0.90 0.99 1.09 0.93 1.07 1.19 0.94 1.10 1.21
HCFC-141 b HIPS 40 1.00 1.06 1.14 0.95 1.01 1.11 1.01 1.13 1.24 1.03 1.11 1.24 □ 0.94 0.97 1.06 0.90 0.92 1.03 0.93 1.02 1.13 0.94 1.02 1.14 |
1.01 1.03 1.04 1.07 [ 1.12 0.94 0.95 0.99

316 / Wilkes, Yarbrough, Gabbard, Nelson, Booth

IHCFC-141 b HIPS -10 0 . 9 5 r 0 . 9 5 [098 1 .0 1 1.02 1.05 1.09 M 0.94 0.94 0.96]M O . 90 0.88 0.91 0.93 [0.96 1.01
r I____ i_ _ _ i_ _ _ □ L I
|HFC-134a ABS 90 1.21 1.26 1.37 1.17 1.23 1.33 1.17 1.25 1.36 1.18 1.25 1.35 M 1.09 1.13 1.25 1.07 1.11 1.22 1.06 1.12 1.24 1.08 1.13 1.24
HFC-134a ABS 40 1.20 1.23 1.31 1.14 1.19 1.23 1.17 1.24 1.29 1.19 1.23 1.28! 1.09 1.10 1.18 1.04 1.05 1.10 1.07 1.10 1.16 1.08 1.10 1.15
|HFC-134a ABS -10 1.21 1.22 1.25 1.16 1.18 1.19 1.18 1,22 1.24 1.18 1.22 1.23 1.09 1.09 1.12 1.05 1.06 1.07 1.07 1.09 1.10 1.08 I 1.10 110
HFC-134a HIPS 90 1.20 1.33 1.44 1.15 1.28 1 .3 7 M 1 9 1 ; 1 .3 3 ] J A 4 \ 1.17 1.30 1.40 □ 1.09 [ 1.21 M.32 1.04 1.16 1.26 1.08 1.20 1.32 1.07 1.19 I 1.29
HFC-134a HIPS 40 1.21 1.27 1.35 1.14 1.20 1.27 1.18 1.27 1.35 1.18 1.24 1.31 □ M.09 1.13 1.22 1.04 1.07 1.15 1.08 1.13 1.22 1.08 1.11 118
HFC-134a HIPS -10 1.21 1.23 1.25 1.17 1.21 1.23 1.19 1.24 1.27 1.19 1.22 1.23!□ 1.09 1.10 1.14 1.07 1.08 1.12 1.07 1.10 1.13 1.08 1.09 • Mil
!
HFC-245fa ABS 90 1.06 1.10 1.16 1.04 1.07 1.16 1.07 1.12 1.21 1.09 1.13 1.22 □ 0.97 0.98 1.07 0.95 0.96 1.06 0.97 1.00 1.10 0.99 1.02
|HFC-245fa ABS 40 1.05 1.08 1.12 1.07! 1.11 | 1.24 1.07 1.11 1.16 1.09 113 M 0.96 0.96 1.00 0.96 0.97 1.11 0.97 0.97 1.02 0.99 1.00 1.04
|HFC-245fa ABS -10 1.06 1.06 1.08 1.04 1.06 1.16 1.07 1.10 1.11 1.09 1.12 1.13 M 0.96 0.96 0.97 0.94 0.94 *1.05 0.96 0.97 0.98 0.99 TOO TOO
HFC-245fa HIPS 90 1.05 1.15 1.21 1.05 1.13 1.22 1.08 1.18 1.28 1.08 1.18 1.27!□ 0.96 1.04 1.10 0.94 1.01 1.12 0.97 1.07 1.16 0.98 1.07 116
HFC-245fa HIPS 40 1.06 1 .1 1 ! 1.17 1.05 1.09 1.17 1.08 1.14 1.21 1.08 1.15 1.21 □ 0.97 0.99 1.06 0.95 0.96 1.05 0.97 1.01 1.09 0.99 1.03 1.091
HFC-245fa HIPS -10 1.05 1.08 1.09 1.08 1.31 1.43 1.07 1.11 1.12 1.09 1.11 1.13 □ 0.96 0.97 0.98 0.98 1.20 1.32 0.97 0.99 0.99 0.98 1.00 101
□ I
Cyclopentane ABS 90 1.16 1.22 1.35 1.18 1.22 1.35 1.14 1.22 1.32 1.11 1.18 1.30 M 1.07 1.13 1.27 1.10 1.16 1.27 1.07 1.17 1.27 1.04 1.11 124
Cyclopentane ABS I 40! 1.16 1.19 1.28 1.18 1.21 1.26 1.14 1.21 1.26 1.12 1.16 1.23!□ 1.06 1.07 1.17 1.10 1.12 1.19 1.07 1.11 1.18 1.04 1.05 1.13
Cyclopentane ABS! ___ -1 0 , 1.16 M7, 1.18 M8| 1.20 1.22 1.14 1.20 1.21 1.12 1.14 M5 □ 1.06 1.05 1.08 1.10 1.10 1.12 1.07 MOJ 1.12 1.05 1.04 1.05
ICyclopentane HIPS 90 1.17 I 1.33 1.43 1.18 1.30 1.41 1.161 1.32 11.401 1.11 1.28 1.39 U 1.09 1.25 1.35 11.121 1.25 I 1.35 1.10 1.27 I 1.35 1.02! 1.22 1.32
ICyclopentane HIPS 40 1.15 1.24 1.37 1.18 1.24 1.31 I 1.15 1.26 1.34 1.12 1.23 1.32 L 1.06 1.13 1.27 1.11 1.16 1.25 1.09 1.19 127 1.05 1.14 124
1

0
ID
CO
<1—

Cyclopentane HIPS 1.16 1.19 1.24 1.21 ! 1.24 1.12 1.19 1.07 1.07 1.14 1.09 1.14 1.05 1.07
CO

1.19 I 1’21
T—

T—
—

tT—
—

T—
 and cyclopentane were 20%, 5%, and 15% higher, This is because only one surface in a refrigerator will have a
respectively, than for HCFC-141b. These relative differences plastic sheet, while the other surface will have an
are in good agreement with the values given in Table 1 for the impermeable steel sheet, which would lower the aging rate of
core foam, also from Supplier A. At the 45°F test the panels.
temperature, the average conductivities for HFC-134a, HFC-
245fa, and cyclopentane were 16%, 4%, and 16% higher than COMPARISON OF DATA W ITH M ODEL RESULTS
for HCFC-141b. The ordering of the blowing agents was
nearly the same for all four suppliers. However, there were Since aging data on the full-thickness panels are being
some slight differences. For Supplier B, the conductivity for collected on a real-time basis, there is a need for a model that
cyclopentane was slightly higher than that for HFC-134a at can predict long-term aging behavior. A one-dimensional
both test temperatures. For Supplier C, the conductivity for model for isothermal aging of foam panels with semi-
HFC-134a was higher than that for cyclopentane at the 75 °F permeable surface sheets has been developed, and is being
test temperature, but was slightly lower at the 45 °F test validated against the measured aging data. Preliminary
temperature. results from the model are presented here.
Figure 3 shows the variation of normalized thermal
conductivity with time for panels from Supplier A as they Description of Model
were aged at 90°F, 40°F, and -10°F and tested at 75 °F. This
figure shows that the panels with ABS sheets aged less The model is similar to those described in the literature for
rapidly than did the panels with HIPS sheets, and that aging diffusion of gases through foam [e.g, 9]. However, we have
decreased with decreasing temperature, with only small added the effects of partial condensation of blowing agent
changes at -10°F. In general, the relative ranking of the four within the foam cells. The transient distribution of a gas
blowing agents was maintained during aging over the three during aging is given by Fick’s second law of diffusion:
year period. Exceptions to this occurred for the panels aged
at 90°F, where the curves for HCFC-141b and HFC-245fa
(1)
crossed, and also for the HIPS panels aged at 90°F and 40°F
where the curves for HFC -134a and cyclopentane came
together or crossed. Results for other suppliers were
generally similar; however, certain details differed. For where c is the concentration of gas in moles per unit volume,
example, the pre-aging difference between HCFC-141b and D is the effective diffusion coefficient which is taken to be a
HFC-245fa was larger for Supplier B than for Supplier A (see constant, t is time, and x is distance. The gas concentration,
Figure 2), and the curves for these two blowing agents did not c, and pressure, P, are related through the ideal gas law, c =
cross for Supplier B. P/RT, where R is the gas constant, and T is the absolute
Figure 4 shows percentage increases in thermal temperature. The foam is produced at relatively high
conductivity after three years of aging. These values were temperatures, but the panels are being aged at lower
obtained by averaging the percentage increases for the four temperatures, where a fraction of the blowing agent can be in
suppliers (except for HFC -245fa with HIPS at -10°F, where a condensed state. In this case, it is assumed that the liquid
the anomalously high values for Supplier B were not included is immobile and the only mechanism for transport of the
in the average). Percentage increases for 2 in.-thick core blowing agent is by vapor diffusion. The vapor pressure
foam for Supplier A are also shown for comparison. The curve provides an upper limit on the gas pressure, and a mass
core foam values at three years and 20 years of aging were balance is added to track the distribution of liquid and vapor
obtained by scaling the results for 1.5 in.-thick and 0.4 in.- phases of the blowing agent.
thick core-foam specimens, respectively. The results show The initial conditions for Equation 1 are that the
that the plastic sheets significantly reduce the rate of aging, concentrations of oxygen and nitrogen are zero, while the
with ABS being more effective than HIPS. The difference concentrations of carbon dioxide and blowing agent are
between HIPS and ABS is attributed to the higher gas uniform, with values that are provided by the foam supplier.
permeances for HIPS. Figure 4 also shows that decreasing Boundary conditions for Equation 1 are given by application
temperature results in decreases in the rate of aging, with the of Fick’s first law of diffusion:
changes at -10°F being very small. This behavior is similar
to that observed for the core foam, and is attributed to the ( 2)
effect of temperature on the permeance of the plastic sheets,
the coefficients of diffusion of gases through the foam, and
the vapor pressure of the blowing agent. Both Figure 3 and
Figure 4 show that the percent increases for HFC-245fa tend where K is the permeance of a surface sheet (K = « for bare
to be somewhat less than for the other blowing agents. cut foam, 0 for solid metal surface, and intermediate values
The data presented for the test panels should not be for plastic sheets on the surface), c0 is the concentration of
interpreted directly as quantitative indications of the changes gas in the space surrounding the panel, c and the derivative
that would be expected in the walls and doors of refrigerators. are evaluated at the interface between the foam and the

Wilkes, Yarbrough, GabbardNelson, Booth / 317

Figure 3. Aging of foam in full-thickness panels blown with third-generation blowing agents by Supplier A. Tested at 75°F.

318 / Wilkes, Yarbrough, Gabbard, Nelson, Booth

Figure 4. Percentage increase in thermal conductivity of foam in full-thickness panels after three years of aging. The legend
shows the surface material: ABS or HIPS plastic sheets. Values for panels are averages over four foam suppliers. Values for
core foam are scaled from aging curves for slices of foam from Supplier A.

Wilkes, Yarbrough, Gabbard, Nelson, Booth / 319

surface sheet, and the plus and minus signs on the left hand
side of Equation 2 apply to the right and left hand sides of the
panel, respectively. The model has been formulated to allow (6)
for different permeances on each side of the panel. The
surroundings were taken to consist of nitrogen and oxygen at
concentrations that correspond to 0.21 atm for oxygen and
0.79 atm for nitrogen. M aterial Properties for Model
Equations 1 and 2 are used to calculate concentrations
through the thickness of the specimen during aging. Gas In addition to properties of the diffusing gases, the model
pressure profiles under the thermal test conditions are then requires values of the effective diffusion coefficients for the
calculated from the concentration profiles using the ideal gas foam and gas permeances for the surface sheets. Permeances
law and the vapor pressure curve for the blowing agent. of C 0 2, 0 2, and N2through specimens of the 0.040 in.-thick
Next, the Lindsay-Bromley formulation of the Wassiljewa ABS and HIPS plastic sheets were measured by a commercial
equation is used to calculate the thermal conductivity of the laboratory [17], and the results are given in Table 4. These
gas mixture as a function of position through the specimen results show that the permeances for HIPS are four to six
[10]. The Wassiljewa equation is times larger than for ABS, and that the permeances at 40°F
were 0.4 to 0.7 of the 90°F values. Diffusion coefficients of
gases through specimens of the foam insulation are in the
(3) process of being measured using a gravimetric method that
was developed by Booth [18]. Data that have been obtained
to date are given in Table 5. The large differences among the
gases should be noted with both the permeances and diffusion
where kgas is the thermal conductivity of the gas mixture, kgi coefficients being ordered as C 0 2 > 0 2 > N 2. Table 5 also
and are the thermal conductivity and mole fraction of the ith shows that the diffusion coefficient of C 0 2 through foam
gas, and Ay is given by the Lindsay-Bromley equation, blown with cyclopentane is about three times larger than
through foam blown with HCFC-141b. This is presumably
due to differences in foam structure with the different
blowing agents.
Values of the diffusion coefficients and permeances at
temperatures other than those at which values were measured
were estimated from a fit of the data to the Arrhenius
equation, as, for example,

where and Sj are the dynamic viscosity, molecular (7)

weight, and Sutherland constant for the ith gas, and T is the
absolute temperature. Also, S{= 1.5 T bi, where Tbi is the
normal boiling point of the ith gas, and Sy = ( S ^ f 2 for non­
polar gases. Gas properties needed for specimens blown with Comparison of Mode! with D ata
HCFC-141b are given in Table 3.
The total apparent thermal conductivity, k, at a plane Figure 5 compares model predictions with measured
within the specimen is then calculated from the equation thermal data for the thinner slices of core foam blown with
given by Glicksman [16] as HCFC-141b. Here, the thermal conductivity is plotted versus
the square root of the aging time in order to expand the early
k = K ol + k rad + 6 k gas = k0 + 8*ga, (5 ) parts of the plots where changes are large and to compress the
later parts where changes are smaller.
The comparisons in Figure 5a are based on the minimum
where kSOI represents conduction through the solid polymer number of adjustable parameters. In addition to the
cell walls and struts, k rad represents radiation through the parameters given above, the model used only two other
foam, 6 is the void fraction (taken to be about 0.97), and kgas adjustable parameters. These are the coefficient of diffusion
is the thermal conductivity of the gas mixture. Glicksman for HCFC-141b, and the sum of the radiative and solid
gives formulae for estimating the values of k ^ and k,ad -; contributions to the thermal conductivity (Iq, in Eq. 5). For
however, for the purposes of this study, their sum is estimated aging of the thinner slices at 90°F, the model predicted that
from the measured thermal data. The average thermal diffusion of all gases except HCFC-141b had been completed
conductivity of the specimen is then calculated from an by about 300 days. After this, the slight increase in thermal
integral over the thickness (L) of the specimen, conductivity is due to loss of the blowing agent. By a trial
and error process, a diffusion coefficient of about 2.5 x 10'10
cm2/s was found to provide a good match for the data.

320 / Wilkes, Yarbrough, Gabbard, Nelson, Booth

 Table 3. Gas Properties for Calculation of Thermal Conductivity of Gas Mixture at 75°F

Thermal
Conductivity, Viscosity, Molecular Weight Normal Boiling
Gas Btu»in./h*ft2»°F centipoise* Point, K

Nitrogen 0.178 [11] 0.0174 28.01 [14] 77.4 [15]

Oxygen 0.179 [11] 0.0200 32.00 [14] 90.2 [15]

Carbon dioxide 0.115 [12] 0.0148 44.01 [14] 194.7** [15]

HCFC -141b 0.067 [13] 0.0102 116.95 [14] 305.2 [15]

* Estimated by Corresponding-States method [14].
** Normal sublimation point.

Table 4. Permeances of Gases through 0.04-in.-thick ABS and HIPS Plastic Sheets

HIPS ABS
Gas
40° F 90°F 40°F 90°F

C02 55.6 82.9 8.78 13.6

o2 8.47 16.2 1.92 4.32

n2 1.64 3.58 <0.32 0.78

Units are cc(STP)/100 in.^atm-day

Table 5. Coefficients of Diffusion of Gases through Foam Produced by Supplier A

Foam Blowing Agent Diffusing Gas 77°F (25°C) 104°F (40°C)

O

Cyclopentane 2.69 x 10^ 4.5 2 x 1 O8

'
o
ro

HCFC-141b C02 8.15 x 10'7 1.40 x 10-*

HCFC-141b 02 3.30 x 10 7 5.03 x 10'7

HCFC-141b n2 1.77 x 10'8 3 .14 x 10-8

Units are cm2/s

Figure 5. Comparison of model predictions with measured data for thin (0.4 in. thick) specimens of foam blown with HCFC- 141b
from Supplier A. Symbols are measured data; lines are model predictions, (a) Model with minimum of adjustable parameters;
(b) with initial C 02 concentration reduced by factor of three and D for 0 2 reduced by a factor of four.

Wilkes, Yarbrough, Gabbard, Nelson, Booth / 321

(Gravimetric measurements are underway to provide an time period.
independent estimate of this parameter.) For other Figure 6b shows a similar comparison for the panels with
temperatures, the ratio of diffusion coefficients for the HIPS surfaces. The agreement between the model and data
blowing agent and nitrogen was assumed to be the same as at is good at 40°F and -10°F. At 90°F, there is good qualitative
90°F. This assumption is not critical, since the importance of agreement, but the model underpredicts the aging by several
diffusion of the blowing agent decreases markedly with percent. The agreement could be improved by allowing the
decreasing aging temperature, due partly to a decrease in the gas permeances to be adjustable parameters. Comparisons on
diffusion coefficient, but also due to a large decrease in more panels, and over a longer time period are needed for a
blowing agent gas pressure because of condensation. more thorough assessment of the adequacy of the model for
Variation of ko shifts the curves up and down, but does not predicting long-term aging of panels.
change their shape. Values of 0.065, 0.066, and 0.061 Physical arguments suggest that the time to produce a given
Btu*in./h*ft2*°F were used for the comparisons shown in change in thermal conductivity would be two to four times
Figure 5 for aging at 90°F, 40°F, and -10°F, respectively. longer when one plastic sheet is replaced with a steel sheet.
This comparison shows that with a minimum of adjustable The factor of two would apply if the permeance of the plastic
parameters, the model predicts the long-term aging of core sheets were the dominating resistance to gas transport, while
foam very well, but further refinements are needed at short the factor of four would apply if diffusion through the foam
times. were the dominating resistance. The model was used to
Two other parameters were varied to improve the estimate the effect of a steel sheet on one side. For HIPS on
comparison of the model with the data as shown in Figure 5b. one side and steel on the other, the model estimated that a
The initial C 0 2 content was arbitrarily reduced by a factor of time period about 3.0 to 3.3 times longer would be needed to
three in order to match the model with the data at zero time. produce the same change as with HIPS on both sides. Similar
Also, the diffusion coefficient for 0 2was reduced arbitrarily simulations for ABS gave a factors of 2.2 to 2.7. These
by a factor of four in order to provide a better match at short results show that for HIPS, the gas transport is controlled
times. Neither of these changes influenced the long-term approximately equally by permeation through the plastic
aging predictions, and both changes may be thought of as sheets and by diffusion through the foam. For ABS, gas
ways to calibrate the model to the data on the core foam. transport is controlled more by permeation through the plastic
The model as calibrated in Figure 5b was used to predict sheets.
the increases in thermal conductivity for the full-thickness
panels blown with HCFC-141b. The model simulations SUMMARY AND CONCLUSIONS
included a period of time between the production of the foam
and the start of controlled-temperature aging, during which Thermal conductivity measurements have been performed
the panels were stored near room temperature. Figure 6a over a three year period on a set of full-thickness panels
shows that the model predicts the aging of the panels with containing polyurethane foam confined between solid sheets
ABS surface sheets very well. Aging of the panels is of ABS and HIPS plastics. The panels simulated the walls
controlled mostly by diffusion of C 0 2, 0 2, and N2, with very and doors of a refrigerator or freezer except that the steel
little change in the concentration of blowing agent over this sheet normally on one side was replaced with a plastic sheet.

Figure 6. Comparison of model predictions with measured data for full-thickness panels containing foam blown with HCFC-141b
from Supplier A. Symbols are measured data; lines are model predictions.

322 / Wilkes, Yarbrough, Gabbard, Nelson, Booth

Measurements at 75°F before the start of controlled REFERENCES
temperature aging showed that the conductivity of foam
blown with HFC-134a, HFC-245fa, and cyclopentane 1. Wilkes, K.E., D. W. Yarbrough, and F. J. Weaver. 1997.
averaged 18%, 7%, and 15%, respectively, higher than that “Aging of Polyurethane Foam Insulation in Simulated
with HCFC-141b. Refrigerator Walls,” International Conference on Ozone
Tests on the panels after three years of aging at controlled Protection Technologies, pp. 253-262.
temperatures have been completed. For both ABS and HIPS
plastics, the conductivity increases were less than those 2. Wilkes, K. E., W. A. Gabbard, and F. J. Weaver. 1998.
predicted for unenclosed full-thickness core-foam, showing “Aging of Polyurethane Insulation Foamed with Second- and
that the plastic liners reduce the rate of aging. The panels Third-Generation Blowing Agents,” Proceedings o f the 49th
with HIPS sheets showed average increases of 17 to 26% Annual International Appliance Technical Conference, pp.
with aging at 90°F, 10 to 19% at 40°F, and 2 to 6% at -10°F. 387-398.
The panels with ABS sheets showed smaller increases of 12
to 18% at 90°F, 7 to 12% at 40°F, and 1 to 5% at -10°F. 3. Wilkes, K. E., W. A. Gabbard, and F. J. Weaver. 1998.
These differences in aging rates correlate with measurements “Aging of Polyurethane Foam Insulation in Simulated
of gas permeances of the plastic sheets. The percentage Refrigerator Panels - Initial Results with Third-generation
increases for HFC-245fa tended to be somewhat less than for Blowing Agents,” The Earth Technologies Forum, pp. 280-
the other blowing agents. 289.
Thermal conductivity measurements have been made over
a two year period on cut slices of polyurethane foam 4. Wilkes, K. E., W. A. Gabbard, and F. J. Weaver. 1999.
insulation blown with HCFC-141b, HFC-134a, HFC-245fa, “Aging of Polyurethane Foam Insulation in Simulated
and cyclopentane. Initial results at 75°F mean temperature Refrigerator Panels - One-Year Results with Third-generation
before aging showed that the thermal conductivities of foam Blowing Agents,” The Earth Technologies Forum, pp. 303-
blown with HFC-134a, HFC-245fa, and cyclopentane were 313.
21%, 4%, and 13%, respectively, higher than with HCFC-
141b. The aging rate was sensitive to aging temperature, with 5. Wilkes, K. E., W. A. Gabbard, F. J. Weaver, and J.
aging at 40°F being about one-half as fast as at 90°F, and Richard Booth. 2000. “Aging of Polyurethane Foam
with aging at -10°F being about one-tenth as fast. The Insulation in Simulated Refrigerator Panels - Two-Year
relative ranking of the blowing agents was generally Results with Third-generation Blowing Agents,”
preserved through the aging process. Polyurethanes 2000, p. 417-428.
Preliminary comparisons of a mathematical model with
aging data show that the model is capable of predicting the 6. C 518, “Standard Test Method for Steady-State Heat Flux
aging behavior of thin slices of core foam very well. The Measurements and Thermal Transmission Properties by
model is also able to give good qualitative predictions of the Means of the Heat Flow Meter Apparatus,” in 2000 Annual
aging of full-thickness panels with solid plastic surfaces. In Book o f ASTM Standards. West Conshohocken, PA:
most cases, very good quantitative agreement was obtained American Society for Testing and Materials, Vol. 04.06, pp.
between the model and the full-thickness data. Comparisons 173-184.
on more panels, and over a longer time period are needed for
a more thorough assessment of the adequacy of the model for 7. Haworth, G. J. 1996. “Next Generation Insulation Foam
predicting long-term aging of panels. Blowing Agents for Reffigerators/Freezers,” Proceedings,
47th International Appliance Technical Conference, pp. 271-
ACKNOWLEDGMENTS 282.

The authors would like to acknowledge the contributions of 8. C 1303, “Standard Test Method for Estimating the Long-
a number of organizations. Funding for this project was Term Change in the Thermal Resistance of Unfaced Rigid
provided by the Appliance Research Consortium, the U. S. Closed Cell Plastic Foam by Slicing and Scaling Under
Environmental Protection Agency, and the U.S. Department Controlled Laboratory Conditions,” in 2000 Annual Book o f
of Energy, Office of Building Technology, State, and ASTM Standards. West Conshohocken, PA: American
Community Programs under contract number DE-AC05- Society for Testing and Materials, Vol. 04.06, pp. 726-734.
OOOR22725 with the Oak Ridge National Laboratory,
managed by UT-Battelle, LLC. Test panels were fabricated 9. Hoogendoorn, C. J. 1994. “Thermal Ageing,” in Low
by BASF Corp., Bayer Corp., Dow Chemical Corp., and Density Cellular Plastics: Physical Basis o f Behaviour, N. C.
Huntsman Polyurethanes, Inc. Aging chambers were Hilyard and A. Cunningham, eds. London: Chapman & Hall,
contributed by Sub-Zero Corp. pp. 153-186.

Wilkes, Yarbrough, Gabbard, Nelson, Booth / 323

10. Reid, R. C. and T. K. Sherwood. 1966. The Properties David W. Y arbrough
o f Gases and Liquids, 2nd Ed. New York: McGraw-Hill,
pp. 483-484. David Yarbrough is an adjunct
research participant at ORNL
11. ASHRAE Handbook-Fundamentals. 1997. Atlanta: and also Professor and Chairman
American Society of Heating, Refrigerating, and Air- of the Chemical Engineering
Conditioning Engineers, pp. 19.75-19.79. Department at Tennessee
Technological University (TTU).
12. Eckert, E. R. G. and R. M. Drake, Jr. 1959. Heat and He earned B.Ch.E., M.S., and
Mass Transfer. New York: McGraw-Hill, p. 506. Ph.D. degrees in chemical
engineering from the Georgia
13. Shankland, I. R., D. J. Williams, M. C. Bogdan, and P. Institute of Technology. He has
B. Logsdon. 1998. “Next Generation Blowing Agents for over 20 years of experience in thermal insulation research, is
Thermal Insulating Foams,” in 60 Years o f Polyurethanes, J. active in ASTM Committee C 16, and is a registered
E. Kresta, E. W. Eldred, eds. Lancaster, PA: Technomic professional engineer.
Publishing Co., Inc., p.86.
W. Alex G abbard
14. Reid, R. C. and T. K. Sherwood. 1966. The Properties
o f Gases and Liquids, 2nd Ed. New York: McGraw-Hill, p. Alex Gabbard is a research staff
405. member at ORNL. He received
his B.S. and M.S. in physics at
15. ASHRAE Handbook-Fundamentals, 1997. Atlanta: North Carolina State University.
American Society of Heating, Refrigerating, and Air- He has been at ORNL since
Conditioning Engineers, p. 18.3. 1980 and has worked on
programs dealing with
16. Glicksman, L. R. 1994. “Heat Transfer in Foams,” in thermonuclear fusion, nuclear
Low Density Cellular Plastics: Physical Basis o f Behaviour, fuels, carbon, and thermal
N. C. Hilyard and A. Cunningham, eds. London: Chapman insulation.
& Hall, pp. 105-152.
Gary E. Nelson
17. Permeance measurements were made by Modern
Controls, Inc., Minneapolis, Minnesota. Gary Nelson is a research
technician at ORNL. He has
18. Booth, J. R., R. S. Graves, and D. W. Yarbrough. 1996. been at ORNL and the Oak
“Effective Diffusion Coefficients for CFC-11 by Gravimetric Ridge Y-12 Plant for 25 years,
Depletion from Thin Slices of PIR Foams,” in Thermal where he has worked in
Conductivity 23, K. E. Wilkes, R. B. Dinwiddie, and R. S. analytical chemistry and physical
Graves, eds. Lancaster, PA: Technomic Publishing Co., Inc., properties measurements.
pp. 352-337.

BIOGRAPHIES

Kenneth E. Wilkes J. Richard Booth

Ken Wilkes is a senior research Richard Booth received his B.S.

staff member at the Oak Ridge from Virginia Polytechnic
National Laboratory (ORNL). Institute and State University and
He holds a B.S. in Mechanical his M.S. and Ph.D. from
Engineering from Duke Clemson University, all in
University, an M.S. in chemical engineering. He
Mechanical Engineering from worked for Dow-Badische and
Purdue University, and a Ph.D. in the Dow Chemical Company
Condensed Matter Physics from until 1994. He is currently an
The Ohio State University. He Adjunct Associate Professor of
has performed research on thermal insulation for over 20 Chemical Engineering at TTU. He has performed research
years. He is a registered professional engineer and is active on the properties of foam insulation for over 20 years.
in ASTM Committee C 16 on Thermal Insulation.

324 / Wilkes, Yarbrough, Gabbard, Nelson, Booth

Analysis of the Evolution of PIR Foams in the Context of the
Phaseout of HCFCs

PIERRE DOURNEL LOTHAR ZIPFEL

SOLVAY Research and Technology SOLVAY FLUOR und DERIVATE
310, rue de Ransbeek Hans-Bockler Allee, 20
B-1120 Brussels D-30173 Hannover
Belgium Germany

ABSTRACT

PIR foams are known for their outstanding fire properties

and for their good insulation performances. The fire INTRODUCTION
behavior can be attributed to the aromatic structure of the
isocyanurate ring and to the non-flammability o f the In the process of substitution of HCFC-141b, two
blowing agent that is used nowadays. alternatives have been identified: hydrocarbons as a
For environmental reasons, HCFC-141b will have to be flammable option and hydrofluorocarbons as a non ­
phased out as o f 01.01.2003 for the production o f PIR flammable option. Whilst Hydrocarbons are already
laminates boards, in Europe and in the United States. applied, liquid HFCs, e.g. HFC-365mfc, HFC-245fa and
This evolution represents a challenge for the producers as blends are under development. In this paper we concentrate
they need to find a high performance alternative in a on skills how to apply best HFCs, notably the non
relatively short period of time. As major requirement, this flammable blend 365/227 and the flammable azeotropes of
alternative will have to maintain the properties o f the PIR 365mfc and hydrocarbons.
foams, notably the fire behavior and the k -value at a high In our previous publications, we have reported on the use
quality level. o f HFC -365mfc and blends thereof as physical blowing
Several blowing agent alternatives have been presented, agent for polyurethane foams. It is now well established
and some of them have already been implemented in that HFC -365mfc is an important component o f blends
production plants. covering a wide range of physical properties such as
Among them, pentane is seen as a candidate for economical thermal conductivity and boiling point. These blends have
reasons, but because of its high flammability, its use shows become necessary for the PU industry, because none of the
some limitations. The non-flammable alternative HFC- existing individual products can match the long list of
365mfc/HFC -227ea has proven to be an interesting optimum properties that characterized the previous
alternative to obtain a good insulation value and good fire generations of products (HCFCs and CFCs).
properties. More and more the concept o f compromise between the
Besides, the existence of azeotropic blends between HFC- various properties and the production costs has become a
365mfc and the pentane isomers increases the degree of general trend in the polyurethane industry.
freedom o f the system and makes the optimization process In this paper, we will look especially at the substitution of
easier. When using the right combination of blowing agent, HCFC -141b for the production of polyisocyanurate foams.
it is possible to reach an interesting compromise between During this change, it is important that the main properties
the insulation characteristics, the fire performance and the o f this type o f foam are maintained at the optimum level.
cost o f the system. This concerns both the insulation properties and the fire
An analysis o f properties o f PIR foams produced with the properties.
several alternatives will be presented in this paper. Therefore, we propose here below to give a survey on the
several available alternatives based on HFC -365mfc and to
establish a comparison with alternative products such as
isopentane.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 325

 STATUS ON HFCs NON-FLAMMABLE BLEND: HFC-365/227:

HFC-365mfc is produced since January 2000 in our pilot Characteristics of HFC-365/227

plant facility located in Tavaux (France). The average
yearly capacity is several hundred of tons. During the year Table 2. Main characteristics of HF0365/227 blend
2000, the availability o f the product allowed us to perform HFC-365/227
industrial trials at customer facilities and to evaluate the Structure CF3-CH 2-CF 2-CH 3
performance of the product in real conditions. In 2001, CF3-CHF-CF 3
sales have started especially for producers that are not Amount of HFC-227ea (%) 6.0
anymore allowed to use HCFC-141b in specific region of Molecular weight 149
Europe. This step can be considered as a smooth transition Boiling point, [°C] 30.0
towards foil commercialization of the product that is Flash point [°C] None
Thermal conductivity of gas 10.9
planned for the second half o f2002 in order to be in line
phase @ 25 °C (mWAn.K)
with the phase-out schedule planned by the European Flame Limits [% by volume] 3.8 - 1 3 . 3
legislation.
This blend has a similar behavior as HCFC-141b. It has no
flash point but it still has flammability limits in air. This
OVERVIEW
means that in practice, the energy required to ignite the
vapor of the blend at the equilibrium is relatively high.
SOLVAY is committed to supplying a liquid zero ODP
Being a blend of products with different boiling points,
blowing agents for plastic foams as substitutes for the
HFC -365/227 will have however a different behavior than
currently used HCFC-141b. Since 1995 we have published
HCFC-141b. Indeed, HFC-227ea being the most volatile
a series of papers on the development of HFC-365mfc as a
compound will have a higher concentration in the vapor
substitute for HCFC-141b (1) (2) (3) (4) (5) (6). HFC-
phase. The following diagram gives an example of
365mfc has a boiling point of 40°C and a vapor thermal
concentration of HFC-365mfc and HFC-227ea in the liquid
conductivity of 10.6mW/m.K, as shown in table 1.
and in the vapor phase at the equilibrium.
Table 1. Main characteristics of HFC-365mfc
HFC-365mfc
Structure CF3-CH 2-CF 2-CH 3
Molecular weight 148
Boiling point, [°C] 40.2
Melting point [°CJ 3 2 .0
Critical Temperature [°C] 187.7
Critical pressure kPa 2754
Thermal conductivity of gas 1 0 .6
phase @ 25 °C (mW/m.K)
Heat of vaporization at boiling 26.2
point, [kJ/mol]
Minimum Ignition Energy 10.4 mJ
Flame Limits [% by volume] 3.8 - 1 3 . 3
Figure 1: Vapor pressure curve of the blend 365/227 in
The developments of blends can be justified by two main comparison to HCFC-141b
reasons.
First of all, HFC -365mfc is a flammable product and very
often, a non-flammable alternative is preferred by the end
user. In that context, the blend HFC -365mfc/227ea has
been introduced as ideal substitute for HCFC-141b.
In addition, for such plants being already converted into
explosion proof facilities that can allow handling of
flammable compounds, low boiling azeotropic blends
based on HFC-365mfc and pentane isomers (n-pentane and
isopentane) have been developed. These blends are of
special interest for the polyisocyanurate foams because
they can help to improve the physical properties of the
foam (k-factor and fire behavior), compared to foams
blown with straight hydrocarbons only.
Figure 2 : Composition of vapor phase in equilibrium with
liquid at several temperatures

326 / Dournel, Zipfel

For that reason, we recommend to use the blend in a closed In this paragraph, we will examine the main characteristics
loop in order to avoid losses o f HFC-227ea via the vapor o f the three types of blowing agent HFC-365mfc/n-
phase. Evaporation of HFC-227ea might lead to a pentane, HFC-365mfc / isopentane and HFC-365/227 in
flammable composition. Indeed, it has been established that order to check their suitability for the production of
if the blend contains less than 5 % wt o f HFC -227ea then, polyisocyanurate foams.
it has a flash point. We will especially have a look at the vapor pressure of the
compounds, at their blowing efficiency, at their thermal
conductivity and finally at their compatibility with standard
FLAMMABLE BLENDS: HFC-365mfc/pentane polyols used in PIR foams.

Several azeotropic compositions containing HFC -365mfc Vapor pressure and condensation
and hydrocarbons have bean identified. Among them, two
are o f great interest for the production of polyisocyanurate For higher boiling blowing agents such as HFC -365mfc or
foams. These two blends are HFC-365mfc / n-pentane and hydrocarbons and for a lambda-test at low temperatures,
HFC -365mfc / isopentane. e.g. at 10°C (European condition) a partial condensation of
H ie properties o f these two blends are listed in table 3. the blowing agent can effect the lambda value. However,
this is not true any more at 23°C (US-conditions).
Table 3 : main properties of365mfc/pentane blends HFC-365mfc has a rather high boiling point. Because of
Properties 3 65m fc/ 365mfc/ this property, the partial pressure of HFC-365mfc at 10 °C
n-pentane isopentane is quite low (0,026 MPa). This means that there is a risk of
Composition (wt%) 58/42 45/55 partial condensation of HFC-365mfc inside die cells at low
Molecular weight 102.7 93.8 temperature whenever the partial pressure in the cells
Boiling point °C 27.5 22.5 exceeds the vapor pressure o f HFC -365mfc at a given
Vapor pressure @ 10 °C 0.050 0.061
temperature. The use of the azeotropic blends based on
(MPa)
Thermal conductivity @ 13.0 13.0
HFC-365mfc and pentane isomers can minimize the effect
24 °C (mW/m.K) of this condensation.
As the vapor pressure of these blends is higher than the
The vapor pressure curve of these two blends is given in vapor pressure of HFC-365mfc, it will be possible to use a
figure 3 with HFC-365mfc as reference. higher loading of physical blowing agent without saturation
o f the vapor phase that can lead to condensation.
Figure 4 gives examples of cell gas composition that can be
used for the two azeotropic blends HFC -365mfc / n-
pentane and HFC-365mfc / isopentane and for the non ­
flammable blend HFC-365/227.
These optimized cell gas compositions for an average
temperature of 10 °C have been calculated according to the
following form ula:

The partial pressure of the blowing agent is a physical

characteristic that depends on its boiling point. For this
calculation, we assumed that the total pressure in the cell is
equal to 0.07 MPa. This figure shows clearly that in the
case of HFC-365/227, the condensation at 10 °C can not be
Figure 3 : vapor pressure curves of azeotropic blends in avoided. Indeed, such a high amount of carbon dioxide in
comparison to HFC~365mfc
the cells would be difficult to reach in the case of a high
index isocyanate foam (PIR).
USE OF HFC-365mfc / PENTANE AZEOTROPIC
On the contrary, the azeotropic blends give a more
BLENDS IN PIR FOAMS
reasonable value and especially the blend HFC-
365mfc/isopentane that combines a low boiling point and a
In general, it is admitted that PIR foams can not be
low content of HFC-365mfc. In that case, it is possible to
produced with a high amount of water. For this reason, it is
use more than 87 mole % of physical blowing agent in
reasonable to think that the physical blowing agent (HFC
comparison of 75 % for straight isopentane. This increase
or hydrocarbon) will be the main component in the foam
cells. o f physical blowing agent combined with the good k-value

Dournel, Zipfel / 327

 If we compare HFC-365/227 to HCFC-141b and pentane, 1
pbw o f this non-flammable blend is equivalent to 0.79 pbw
of HCFC-141b and 0.48 pbw o f pentane. This means feat
this blend will be less efficient as blowing agent
On fee contrary, fee blend HFC -365mfc/isopentane has
efficiency comprised between fee one of HCFC-141b and
fee one o f straight pentane. Indeed, 1 pbw o f this blend is
equivalent to 1.25 pbw o f HCFC-141b and to 0.77 pbw of
isopentane.
In practice, this means that fee amount of blend HFC-
365mfc/isopentane feat needs to be used will not differ a
lot from what is being used nowadays in production.

Therm al Conductivity of Blowing Agents

Figure 4 : Optimum ceil gas composition to be used for
several physical btowing agents in order to avoid The thermal conductivity in fee gas phase is fee key
condensation at 10 °C. property for a blowing agent. Indeed this property will have
a big impact on fee final thermal conductivity o f fee foam.
o f HFC-365mfc should lead to a significant improvement The k-factor of the several alternatives is shown in figure 6.
o f the thermal conductivity of the PIR foam. In fee case of the azeotropic blends, this value has been
This optimum cell gas composition depends of course of calculated.
the average temperature at which the foam has to perform
or o f the average temperature that is used to measure the k-
factor according to a specified standard
If the average temperature is higher then, the condensation
is not present anymore. As example if the measuring
temperature is 24 °C, then it is possible to use up to 81
mole % o f HFC -365mfc in die cells instead of 40 % for an
average temperature of 10 °C.

Blowing efficiency

The blowing efficiency is related to the molecular weight

o f the product and it will determine the amount of blowing
agent that will be needed in order to reach a given density.
In the following figure, this blowing efficiency is
calculated according to two scales. In one case HCFC-141b
is taken as the reference whilst in the second case, pentane
(normal or iso) is the reference. Figure 6 : thermal conductivity of btowing agents in the gas
phase.

The gas phase thermal conductivity of the azeotropic

blends is quite high compared to the blend HFC-365/227,
which has a typical value of a fluorinated gas (close to 10
mW/m.K). This increase of the thermal conductivity
obtained in fee case fee azeotropic blends is due to fee
relatively high amount of pentane in fee blend, which can
increase heat transfer because o f its low molecular weight.

Solubility of blowing agents in polyester polyols.

The solubility o f HFC -365mfc blends in standard polyester

polyols has been checked. Figure 7 shows the behavior of
these polyols
Figure 5 : Btowing efficiency of several alternatives

328 / Dournel, Zipfel

 Examples of formulations used for this study are listed in
table 4.

Table 4 : typical formulations

Formulation A B
Polyol blend 100 100
TCPP 17 17
L6900 2 2
DMCHA 0.3 0.3
DABCO TMR 3 3
H20 1.25 1.25
HFC-365mfc 17 18
Azeotropic blend 365/isopentane 13
Isopentane 10
ISO index 250 250

Figure 7 : solubility of HFC-365mfc blends in polyester The cell gas composition for those two formulations is
polyols.
given in figure 8.
Generally speaking, the non-flammable blend HFC-
365/227 is much more soluble in aromatic polyester
polyols than the isopentane.
The azeotropic blend HFC -365mfc / isopentane has an
intermediate behavior although in some cases, the
solubility in lower than expected.

PROPERTIES OF PIR FOAMS BLOWN WITH HFC-

365mfc / PENTANE AZEOTROPIC BLENDS

Description of the Formulations

Polyisocyanurate foams have been prepared in the

laboratory at an isocyanate index o f 250 and using several
Figure 8 : Initial cell gas composition for different blowing
combinations o f blowing agents. agent blends
Isopentane has been selected because the azeotropic blend
HFC -365mfc / isopentane has the lowest boiling point.
In the first case (Table 4, formulation A), we made the Foams Thermal Conductivity
hypothesis that HFC-365mfc and isopentane behave
independently in the cells. The quantity o f HFC-365mfc The foams thermal conductivity has been measured at 24
has been adjusted according to the vapor pressure of HFC- °C at the initial time. These results have been correlated to
365mfc at 10 °C (0.026 MPa). the initial cell gas content and more especially to the
In the second case (Table 4, formulation B), we assumed relative amount o f HFC -365mfc and isopentane.
that the azeotropic blend HFC -365mfc / isopentane is not A summary o f the results is listed in table 5.
affected by additional quantities of HFC-365mfc. The
quantity o f HFC -365mfc has then been calculated using the Table 5 : cell gas composition and k-tactor
vapor pressure curve of the azeotropic blend and the vapor co 2 21 % 21 % 22% 21 % 21 %
pressure curve o f HFC -365mfc. HFC-365mfc 35% 48% 7 5%
Isopentane 44% 31 % 7 8%
HFC-227ea 4%
HCFC -141b 7 9%
Initial k-factor @ 25.9 24.7 27.1 23.7 22.8
24 °C (mWAn.K)
Values are in mole %

This table shows that the initial lambda value can be

correlated to the amount of isopentane or HFC -365mfc
present inside the cells.
Indeed, when the cells contain only isopentane as physical
blowing agent, the initial lambda value at 24 °C can be as

Doumel, Zipfel / 329

high as 27 mW/mK. On the opposite, if the cells contain condensation in this range of temperature. This situation is
HFC-365mfc and HFC-227ea, the initial lambda value is identical to HCFC-141b. This fact can validate the
close to 24 mW/m.K at 24 °C. hypothesis that we mentioned in the beginning of this
This evolution can also be observed in figure 9. paragraph that says that when additional HFC~365mfc is
added to the azeotropic blend, the azeotropic behavior is
maintained.
On the contrary, when the non-flammable blend is used,
some condensation can be observed at 10 °C because the
difference with the reference is higher at 10 °C than at 24
°C (9 % at 10°C instead of 4 % for 24 °C). However, this
increase stays within reasonable limits and it is anyhow
lower than the increase observed when using straight
isopentane.

Designed of Optimized PIR Formulations

Optimization of PIR foams should be done by calculating

accurately the partial pressure of each component in the
cells. This calculation of optimum cell gas content has to
Figure 9 : Evolution of the initial thermal conductivity with be done by taking into consideration the partial pressure of
the amount of isopentane in the cells (the cells contain carbon dioxide, of HFC-365mfc and of the azeotropic
initially 21 % of CO2). blend HFC-365mfc / isopentane. Indeed, it has been
verified that HFC-365mfc and the azeotropic blend HFC-
In order to better understand how these blends could 365mfc / isopentane have an independent behavior as cell
modify the insulation properties of PIR foams, it is also gas.
interesting to compare the k-factor of these foams to the k- Basically, the water content used in PIR foams is low
factor o f a standard HCFC-141b blown PIR foam. This has (between 0.5 pbw and 1.0 pbw) and, for optimum
been done in figure 10. insulation performance, the amount of blowing agent has to
be calculated according to the vapor pressure curve of this
compound and to the temperature at which the k-factor will
be measured. As example, if the k-factor is measured at 10
°C, then the partial pressure of HFC-365mfc in the cells
should not exceed 0.026 MPa but if the k-factor is
measured at 24 °C, then this same partial pressure can be as
high as 0.050 MPa.
Once this calculation has been done, then the total amount
o f blowing agent is adjusted by adding the azeotropic blend
HFC-365mfc / isopentane in order to get the required foam
density. There is no limitation on the amount of azeotropic
blend that can be used. Indeed, this blend has a low boiling
point and no condensation can occur in the temperature
range that we are considering in this paper (between 10 and
24 C°).
Typical optimized cell gas composition for PIR foams at
index 250 are given in table 6.
Figure 1 0 : Thermal conductivity of foams prepared using
new blowing agents blends in comparison to a HCFC-141b
Table 6 : Optimum cel! gas for best k-factor @ various T
blown foam (the cells contain initially 21 % of CO2).
@ 1 0 °C @ 24 °C
co2 9% 9%
The thermal conductivity of the isopentane blown foam is
HFC-365mfc 35% 71 %
about 20 % higher than the thermal conductivity of the HFC-365mfc/isopentane 56% 2 0%
reference foam based on HCFC-141b. If we increase
Compositbns are given in mole %.
gradually the ratio of HFC-365mfc in the cells, this
difference decreases leading to only 4 % at 24 °C when the Figure 11 shows the corresponding composition of the
non flammable blend HFC-365/227 is used without blowing agents blend in the polyols blend.
isopentane.
We can also notice that die difference in k-factor is
identical at 10 °C and at 24 °C for die two blends HFC-
365mfc / isopentane. This suggests that there is no

330 / Dournel, Zipfel

 When using isopentane as only physical blowing agent, the
flame height is between 10 and 11 cm depending in the
ISO index. If we add 35 mole % o f HFC -365mfc then, the
flame height is decreased down to 8.0 cm. Values close to
7.5 cm can even be obtained with an amount o f HFC-
365mfc of 45 mole %.
If the non-flammable blend HFC-365/227 is used, then the
flame height becomes very close to the flame height o f a
PIR foam blown with HCFC-141b. The difference is
indeed 1.2 cm at an ISO index o f 250 and 0.5 cm for an
isocyanate index of 320.
This study shows clearly that even if HFC -365mfc is
classified as a flammable product, the use o f this blowing
Figure 11: Optimum blowing agent composition (in wt %) in agent does not affect that much the flammability o f the
the polyol blend for different evaluation temperatures. foam itself. In general, similar fire properties as with
HCFC-141b can be easily obtained.
As conclusion, if we want to get the best R -value for the
foam at 10 °C, the amounts of HFC-365mfc and azeotropic General Properties
blend HFC-365mfc / isopentane to be used in the polyols
blend are very similar. In order to obtain the best k-factor The evolution of the PIR foam after the phase out of
o f the foam at 24 °C, then it is better to use a higher HCFC-141b will depend on the selected alternatives.
amount o f HFC -365mfc (up to 83 %) and a limited amount Several blowing agent blends give promising results as
o f azeotropic blend HFC-365mfc / isopentane (15 %). they can maintain the properties of the PIR foam at a
HFC-365mfc is needed in order to reduce the k-factor of comparable level as today’s product.
the foam but the quantity is limited by the partial pressure Figure 13 shows the several alternatives studied in this
o f this product at a selected temperature. The azeotropic document and establishes a classification based on the two
blend HFC -365mf / isopentane is used in order to adjust the key properties of the PIR foams: the thermal conductivity
density to the desired value. and the fire behavior.

Fire Properties

PIR foams are known for their outstanding fire properties.

The use o f highly flammable blowing agents such as
hydrocarbons can have a negative impact on this key
property.
The impact o f the substitution of HCFC-141b on the fire
properties of PIR foams has been checked according to the
DIN 4102 standard (B2 test) and using several
combinations o f blowing agents. The results are
summarized in figure 12.

Figure 1 3 : classification of the several blowing agent

alternatives based on the final properties of the PIR foams.
(365/iC5_1 blend contains 31 mole % of isopentane and
365/iC5_2 blend contains 44 mole % of isopentane).

Logically, the classification follows the amount of

isopentane present inside the cells. The non-flammable
blend HFC-365/227 gives similar properties to HCFC-141b
blown foams. The addition of isopentane leads to an
increase of the thermal conductivity and of the flame
height.
Obviously, a classification of these blowing agents based
on economical aspects would show the opposite trend, as
isopentane is the less expensive component in the
Figure 12: Evolution of the fire behavior of PIR foams with formulation. For this reason, optimization of the PIR foams
the amount of HFC-365mfc. (formulations contain 1.25 pbw will require some compromise depending on the desired
of water). end use properties and on the average expected costs.

Doumel, Zipfel / 331

CONCLUSIONS REFERENCES
1. P. Doumel et al., HFC-365mfc b le n d s .Status and
HFC-365mfc has been developed as third generation Development as Blowing Agent for High Performance
blowing agent with no ozone depletion potential. Several Insulation Foams, Polyurethanes Conference 2000,
blends based on this component have been designed in Proceedings p. 171.
order to substitute HCFC-141b in various applications 2. P. Doumel et al, HFC-365mfc: Optimization of
where the process and the required end use properties are Formulations Based on HFC -365mfc and Blends Thereof
not always identical. Polyurethanes EXPO’99, Proceedings p. 227
The non-flammable blend HFC-365/227 has been designed 3. L. Zipfel et al, HFC-365mfc: A Versatile Blowing Agent
in order to give the possibility to our customers to work for Rigid Polyurethane Foams,
safely with a non-flammable liquid blowing agent. The Polyurethanes EXPO’98, Proceedings p. 51
evaluation o f this product tells us that the foam properties 4. L. Zipfel et al, The Next Generation Blowing Agent,
that can be achieved are very similar to the properties of a from a Single Product to a Product Range,
HCFC-141b blown foam. Polyurethanes EXPO’98, Proceedings p. 265
In parallel, flammable azeotropic compositions with 5. L. Zipfel et al, HFC-365mfc and HFC-245fa - Progress
hydrocarbons can be o f great interest especially for the in Application of New HFC Blowing Agents, in
production of PIR foams. Among these blends, the Polyurethanes World Congress497, Proceedings p .176
azeotropic composition HFC -365mfc / isopentane (45 / 55) 6. P. Barth&emy et al, Progress in the Development o f a
can be used as a low boiling compound having a high New HFC Blowing agent - A Closer Analysis,
blowing efficiency. The foam properties that can be Polyurethanes EXPO’96, Proceedings p. 404
obtained depend on the relative amount of isopentane and 7. P. Barthelemy et al, Latest Results in the Development
HFC-365mfc that is present inside the cells. Optimum o f Next Generation HFC-Blowing Agents, Polyurethanes
foam properties can be obtained using straight HFC- Expo 1995, Proceedings p. 26
365mfc (or eventually the non-flammable blend HFC-
365/227) as insulation gas with as limitation the partial
pressure of this compound at the temperature where the k- BIOGRAPHIES
factor will be measured. If the amount of HFC -365mfc that
can be introduced without reaching the condensation area is Pierre Doiirnel
too low in order to get the right density, we can recommend Pierre Doumel received a
to adjust this density using the required amount of PhD in Chemistry from the
azeotropic blend HFC -365mfc / isopentane. University o f Bordeaux,
By doing this, it is possible to obtain optimum insulation France. He joined SOLVAY
and fire properties. in 1991 and is now
As for the low isocyanate index PUR foams, HFC-365mfc responsible for technical
represents a wide range of opportunities for the production customer service in the area
o f optimum PIR foams having properties which are very o f foam blowing agents, as
similar to the ones that are produced nowadays with well as for the brominated
HCFC-141b. polyol IXOL®.

DISCLAIMER

All statements, information, and data given herein are Lothar Zipfel
believed to be accurate and reliable but are presented Lothar Zipfel received a PhD
without guarantee, warranty or responsibility o f any kind, in Chemistry from the
express or implied. University of Hanover,
During the handling of the blends described in this paper, Germany. He joined the
the user should not assume that all safety measures are Fluorinated Products Group
indicated, or that other measures may not be required. o f SOLVAY in 1982. He is
Manager of the technical
service for plastic foams of
ACKNOWLEDGMENTS SOLVAY FLUOR und
DERIVATE GmbH. His
The authors wish to thank Mr A.Schrouben (SOLVAY responsibilities include
Research and Technology in Brussels) for the realization of technical support for SOLKANE® blowing agents and for
the several analyses and for the preparation of the several the brominated polyol IXOL®.
foam samples.

332 / Doumel, Zipfel

 HFC 365mfc-Blown Spray Foams: Market Positioning and
Environmental Benefits

HERMANN KRAHLING LOTHAR ZIPFEL

SOLVAY Management Support GmbH SOLVAY FLUOR und DERIVATE
Hans-Bockler Allee, 20 Hans-Bockler Allee, 20
D-30173 Hannover D-30173 Hannover
Germany Germany

• Products must be safe during production and use.

ABSTRACT • The environmental performance of the product should be
acceptable, in that natural resources are used in an effi­
The zero-ODP product SOLKANE 365mfc is currently cient manner, and impacts or hazards are minimized.
being developed to replace HCFC -141b as a blowing and Consequently, product development is an optimization
insulation agent in high-performance polyurethane (PU) task with respect to all relevant aspects throughout the
foams for the thermal insulation of buildings. In a life cy ­
whole life cycle. Unacceptable burdens to the environment
cle assessment (LCA) study, HFC-365mfc-blown PU must be avoided. For instance, the Kyoto protocol (1997)
sprays were investigated in comparison with water/C02-
monitors a »basket« of gases with respect to their potential
blown PU sprays. contribution to global warming. Additionally, the reference
The climatic conditions were recognized as being deci­ high-performance blowing agent HCFC-141b will be with­
sive for the absolute environmental benefits achievable drawn from the market because of its contribution to the
with high-performance insulation products. In order to
depletion of the stratospheric ozone layer.
support the positioning of these new products in distinct
The newly developed blowing agent HFC-365mfc pos­
markets, the environmental profiles were calculated for dif ­
sesses no ozone depletion potential (ODP). It does, how ­
ferent climatic zones in Spain and Portugal.
ever, contribute to global warming if released into the at­
The SOLKANE 365mfc-blown rigid PU sprays were
mosphere. Political acceptance for HFC use requires com ­
found to be environmentally advantageous or at least com ­
prehensive environmental data. Emphasis was given to
petitive as compared to the water/C02-blown PU sprays
those environmental criteria which are currently considered
with respect to all environmental criteria and in the inves­
to demand particular attention, e. g. total energy demand,
tigated climatic zones. They show the potential to set the
global warming, acidification, nutrification, and summer
future standard of environmentally sound high-
smog.
performance insulation products with SOLKANE 365mfc
In a previous life cycle assessment study, HFC-365mfc-
as blowing and insulation agent. blown PU products had been found to be environmentally
competitive in a wide range o f settings [1], However, the
INTRODUCTION question remained as to which environmental profile the
HFC-365mfc-blown insulation products for buildings show
Background and Motivation in different climatic zones. Regarding the forthcoming
market entrance of these products, this aspect was consid­
Insulation systems help us to provide adequate tempera­ ered to be of significant interest.
tures in our buildings in order to ensure health and com ­
fort. High-performance systems offer optimum efficiency Project Partners
in this respect. When aiming at the development of sus­
tainable products, the benefits and burdens of the entire
Recognizing their joint responsibility, the following
product system need to be taken into consideration. Several
partners commissioned the life cycle assessment (LCA)
important requirements must be met throughout the entire
study reported herein:
life cycle o f the products:
• the Spanish association for PU foams, Asociacidn
• An advantageous performance and service should be
T£cnica del Poliuretano Aplicado (ATEPA);
feasible at competitive costs.
• Elastogran SA (BASF Group), Spain;

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 333

• Solvay Fluor und Derivate GmbH, Germany; the other hand as a not-in-kind alternative the water/C02-
• Group Synthesia Espaflola SA, Spain; blown PU sprays.

• and Advanced Chemicals SL (Spain) as a cooperating Relevance of Climatic Conditions

company.
Because of the considerable market importance of PU
AIMS, SCOPE AND KEY SETTINGS sprays for thermal insulation in southern Europe, the envi­
ronmental profiles were calculated for Spain and Portugal.
The LCA study presented herein addresses the environ­ It should be noted that there is a comparable relevance of
mental performance of PU spray thermal insulation sys­ these applications for the U.S. market. The investigated
tems for roof and wall applications in different climatic climatic zones range from those with high demands on
zones of Spain and Portugal [2]. The study aimed to guide heating and/or cooling to those with limited seasonal
the product development and market positioning of the changes and moderate temperatures.
new blowing agent SOLKANE 365mfc and to provide According to the Spanish regulation NBE-CT-79 for the
sound information for the dialogue with customers, politi­ thermal insulation of buildings, the country is divided into
cians, authorities, and the general public. It should be five different climatic zones based on the heating degree
noted, that in this early phase of application development it days. Similarly, the corresponding Portuguese ordinances
is still open whether at the end HFC-365mfc alone or define equivalent climatic zones. These climatic zones
blends with e. g. HFC-227ea or HFC-245fa will be used. were represented by the selected cities Almeria, Barcelona,
At present, however, there is not sufficient information on Toledo, Madrid, Avila for the Spanish peninsula, and Lis­
these development options to reliably calculate such sce­ bon exemplifying Portugal. Based upon the respective
narios. heating and cooling degree days for each of these regions,
The life cycle assessment was performed in accordance the energy demand during the service life of the thermal
with ISO 14040, and the study was subject to a critical re­ insulation was calculated.
view by an independent accredited verifier.
Scenario and Sensitivity Analyses
Life Cycle and Application of Insulation Foams
Table 1 summarizes the key settings for the calculation
The scope of the study encompassed the entire life cycle of the life cycle inventory.
of the thermal insulation systems, from cradle to grave.
The production of the raw materials was represented by the In the framework of the study, scenario and sensitivity
respective eco-profiles of each constituent. The production analyses were conducted for relevant parameters in order
of the foam as such and its application for the thermal in­ to verify the reliability and transferability of the results:
sulation of roofs and walls were considered in conjunction • the variation of the thermal conductivity of the
with the service life. HCFC/HFC blowing agents;
After a service life of 50 years, the building is assumed
• the service life of the dwellings was varied between 10
to be renovated or demolished. Regarding the waste man ­
and 80 years (mean: 50 years);
agement options for the thermal insulation, recovery and
disposal were only briefly discussed because of the respec­ • the diffusion rate of the blowing agents was varied up to
tive uncertainties and negligible relevance according to the an assumed complete loss during the service life (see
investigation. below);
The investigated scenarios reflect typical construction • the thickness of the insulation foam was varied to ac­
habits involving rigid polyurethane sprays, and the char­ count for spatial restrictions (see below);
acteristic climatic conditions of the Iberian peninsula. The
• the different climatic conditions (heating and cooling
following two applications for rigid polyurethane sprays degree days) were represented by the various regions as
were considered; explained above;
• A pitched warm roof ­ plasterboard on rafters, thermal
insulation, ventilated cavity, membrane, battens and • different heating systems (given heating mix as standard,
slates. gas, oil, or electricity heating as scenarios) were consid­
ered.
• A cavity wall ­ plastered supportive brick wall, thermal
insulation, ventilated cavity, and outer ventilated brick. • Because of uncertainties regarding the future regulation
of waste management in Spain and Portugal and the re­
The reference products included on the one hand PU
spective practices, recovery and municipal waste incin­
sprays blown with HCFC-141b as today’s standard, and on
eration were considered as the most realistic scenarios.

334 / Krdhling, Zipfel

 Table 1. Settings and parameter ranges for the calculation of the life cycle inventory.
Parameter Unit Mean Range considered

Thermal conductivity W/(m.K) HCFC-141b: 0.026 0.025 - 0.027

HFC-365mfc: 0.028 0.028 - 0.029
Service life a 50 10 - 80
Blowing agent loss during % 5 —
production
Diffusion rate %/a 1.2 0.0 - 2.0
Thickness m 0.03 0.01 - 0.05
HDD CDD
Heating and cooling degree K.d/a Almeria: 673 873
days HDD/CDD (base 18 °C)
Barcelona: 1403 448
Toledo: 1660 778
Madrid: 1979 541
Avila: 2868 99
Lisbon: 870 527

supply, a mix of fuel sources typical of the Iberian penin­

RESULTS sula was used. In each case, rigid PU foam sprays of equal
thickness are employed for thermal insulation, but the
Total Energy Demand blowing agents differ: the first column represents insula­
tion products using the reference blowing agent HCFC-
Figure 1 shows the total energy demand of the pitched 141b, the second column the not-in-kind alternative wa-
warm roof of a residential building calculated for die five ter/C 02, and the third column represents PU sprays using
cities representing the climatic zones in Spain, and for Lis­ the new blowing agent HFC-365mfc.
bon exemplifying the situation in Portugal. For energy

Figure 1. Cumulative energy demand of the pitched warn roof application. Columns represent, from left to right, the variants
with HCFC-141b-, H2 O/CO2-, and HFC-365mfc-blown PU foam, respectively.

Krdhling, Zipfel / 335

 The energy demand for the production of the foam is The calculations confirmed that the diffusion of the
negligible in comparison to the energy demand of the life blowing agent HFC -365mfc affects the total global warm ­
cycle. The greatest part of the energy demand is caused by ing upon its release by approximately 10 %. The predomi­
the space heating, with quantities growing from the warm ­ nant source of greenhouse gas emissions, however, are the
est regions (Almeria, Lisboa) to the coolest regions (Av­ energy supply processes. Again, the total quantities of
ila). Conversely, the cooling energy demand diminishes, greenhouse gas emissions grow from warmer to cooler re ­
but is comparatively of less importance in these balmy gions as expected. Similarly, the differences between the
climates. In all regions considered, PU sprays blown with HFC- and water/C02-blown products become more dis­
HFC-365mfc show advantages compared to the water/C02 tinct. On the whole, the PU sprays blown with HFC-
system and attain almost the beneficial profile of HCFC- 365mfc were found to be superior compared to the wa-
14lb -blown sprays that are today’s standard. The absolute ter/C02 systems or - in the warmest regions with low de­
differences become more pronounced in the cooler regions. mand for thermal insulation - at least competitive. As
Results for the insulation of a cavity wall are similar to the above, the results for the insulation of a cavity wall are
findings for the pitched warm roof. similar to these findings.

Global W arm ing Potential O ther Im pact Categories

Similar to the findings for energy, the production of the The acidification potential (AP), nutrification potential
foam contributes negligibly to the total global warming (NP), and photochemical ozone creation potential (POCP)
potential (GWP). Beside the C 0 2 equivalents from energy were found to be determined almost exclusively by the
supply for space heating and cooling, however, the diffu­ emissions caused by the supply of thermal energy during
sion of blowing agents during foam processing and service the use of the PU sprays. Therefore the qualitative results
life of the products was expected to play a certain role. resemble the findings for the cumulative energy demand.
Figure 2 shows the GWP (calculation base of 100 years)
of the same pitched warm roof in the six different regions.

Figure 2. Global warming potential of the pitched warm roof application. Columns represent, from left to right, the variants with
HCFC-141b-, H2 O/CO2-, and HFC-365mfc-blown PU foam, respectively.

336 / Krdhling, Zipfel

 GWP, the respective sensitivity analysis is also worth dis­
Sensitivity Analyses cussing. Again, the scenario of the pitched warm roof lo­
cated in Madrid was considered. Figure 4 shows the global
As examples of the sensitivity analyses, the variation of warming potential of the three systems. The diffusion rate
the thickness of the insulation foam and the variation of the of the blowing agent within the foam was varied from zero
diffusion rate of the blowing agents are presented. Figure to 2% annually, corresponding to a complete loss within
3 shows the savings of greenhouse gas emissions incurred the service life of 50 years.
by HCFC-141b and HFC -365mfc in relation to the alterna­ While the water/C02 system is hardly affected by the
tive water/C02 system considering a variation of the foam diffusion of the blowing agent, both HCFC-141b- and
thickness. The scenario is the pitched warm roof, located in HFC-365mfc- show a linear increase of the total GWP, as
Madrid as an appropriate average climatic zone of the Ibe­ expected. However, both blowing agents still perform bet­
rian peninsula and as an important market. ter than the water/C02 system across the whole range of
It is obvious that products containing the blowing agents diffusion rates, i. e. also in the case of a complete loss of
HCFC-141b and HFC -365mfc perform better than the the blowing agent during service life. Again, it is note­
water/C02 system throughout the whole range of foam worthy that the other parameters of the environmental pro­
thickness from 1 up to 5 cm. The greatest advantages of file of HFC-365mfc-blown sprays remain advantageous for
HFC-365mfc around 10 % savings are achieved in a range the entire range of diffusion rates under consideration.
of 1-3 cm thickness. With increasing thickness of the As was shown for the variation of the above parameters,
thermal insulation spray, this advantage diminishes, until all other scenario and sensitivity analyses confirmed the
the performance is about equal to the water/C02 system at general advantage (or, at least, competitiveness) of the
5 cm thickness. It should be noted that the advantages of HFC-365mfc-blown PU sprays as compared to the wa-
HFC-365mfc-blown sprays regarding the other environ­ ter/C 02 system in a wide range of settings.
mental impact categories remain valid for the entire range With respect to probable waste management scenarios,
of thickness considered. the respective impacts arising from a treatment in munici­
Since the diffusion of blowing agents from the foam was pal waste incinerators were found to affect the results only
found to be of a certain importance for the parameter by less than 2 %.

Figure 3. Global warming potential for the pitched warm roof application, shown vs. insulation thickness, using the variants
with HCFC-141b- and HFC-365mfc-blown PU foam, respectively. Savings in comparison with the H20/C02-blown PU foam.

Krdhling, Zipfel / 337

Figure 4. Global warming potential for the pitched warm roof application, shown vs. diffusion rate of the blowing agent during
service life, using the variants with HCFC-141b-, H2O/CO7- and HFC-365mfc-blown PU foam, respectively.

REFERENCES
CONCLUSIONS 1. Solvay Management Support GmbH. March 2000.
“HFC-365mfc as blowing- and insulation agent in
The following conclusions can be drawn from the LCA polyurethane rigid foams for thermal insulation,” LCA
study: report. Hanover, Germany.
• In all five climatic zones of the Spanish peninsula as 2. Solvay Management Support GmbH. March 2001.
well as in the Lisbon area exemplifying Portugal, the “High-performance thermal insulation for buildings in
SOLKANE 365mfc-blown rigid PU sprays are environ­ Spain and Portugal,” LCA report. Hanover, Germany.
mentally advantageous or at least competitive regarding
all criteria considered as compared to the water/C02- BIOGRAPHIES
blown PU sprays.
Hermann KrShling
• In particular, assuming Madrid as an appropriately rep­
resentative region, savings of energy amount to 15 %. Hermann Krahling received a PhD in
This corresponds to 1 year of free heating every 7th year. Biophysics from the University of Stutt­
The associated contribution to global warming is re ­ gart, Germany. He joined the Solvay
duced by 10 % by SOLKANE 365mfc as compared to Group in 1986. He is Department Head of
the water/C02 system. This means that we save one Environmental Management—Products
year‘s worth of greenhouse gas emissions every decade. and Markets at Solvay Managment Sup­
port GmbH. His responsibilities include
• Other emissions causing environmental impacts such as support for Solvay business units regarding LCA, product
acid rain, nutrification, and summer smog, as well as stewardship and sustainable development.
amounts of waste are reduced in a manner similar to the
energy savings mentioned above. Lothar Zipfel
With respect to the particular aims of this study, it was Lothar Zipfel received a PhD in Chemistry from the
shown that in geographic zones with demanding climatic University of Hanover, Germany. He joined the Fluori-
conditions the high-performance HFC-365mfc-blown rigid nated Products Group of Solvay in 1982. He is Manager of
PU insulation sprays offer the most pronounced environ­ the technical service for plastic foams of
mental advantages. The investigated HFC-365mfc-blown Solvay Fluor und Derivate GmbH. His re ­
foams were found to have the potential to set the future sponsibilities include technical support for
standard of environmentally sound high-performance in­ SOLKANE® blowing agents and for the
sulation products in a wide range of settings. The climatic brominated polyol IXOL®.
conditions associated with the intended market region con­
stitute a decisive factor for the achievable environmental
benefits.

338 / Krahling, Zipfel

HFC-134a as a Coblowing Agent with Pentanes

JINHUANG WU a n d DOUG DILLON

ATOFINA Chemicals, Inc.
900 First Avenue
King o f Prussia, Pennsylvania 19406

ABSTRACT reasonable boiling point, no flash point, low

flammability, low thermal conductivity and low
The Montreal Protocol regulates the uses of ozone permeability out of foam. However, HCFC-141b will
depletion substances (ODSs). In the developed eventually be phased out due to its residual ozone
countries, the phaseout of high ozone depletion potential depletion potential (ODP) of 0.11. Its phaseout date is
(ODP) CFCs in rigid foam applications was made governed by the Montreal Protocol and local regulations.
possible with the introduction of lower ODP HCFCs. There are two approaches in phasing out HCFC-141b.
One is to go from HCFC -141b directly to zero ODP
HCFC-141b was chosen as an excellent replacement for
alternatives. The other approach is to use lower ODP
CFC-11 in many rigid foam applications. HCFC-141b
HCFCs as a transition. For some regions, it may be
blown rigid foams offer excellent thermal insulation, fire
feasible to reduce the total consumption of ozone
resistance and dimensional stability.
depletion substances by switching to blowing agents
According to the Montreal Protocol, HCFC-141b will
having lower ODP. These compounds include HCFC-
eventually be phased out. Lower ODP HCFCs ( i.e.,
142b (ODP = 0.065), HCFC-22 (ODP = 0.055) and their
HCFC-142b and HCFC-22), zero ODP HFCs and
blends. These lower ODP HCFCs are particularly useful
hydrocarbons are considered to be alternatives.
for appliances and some system applications because
Unfortunately, neither HFCs nor pentanes can match
they can handle gaseous blowing agents (GBAs)
performance and cost of HCFCs.
relatively easily. Furthermore, these lower ODP HCFCs
An HFC and pentane(s) coblowing system can
offer similar low thermal conductivity as HCFC-141b,
potentially offer some advantages over purely pentane
especially for low temperature applications.
blown foams including lower foam density (therefore
In the transition from HCFC -141b to zero ODP
lower cost), better thermal insulation and better fire
blowing agents, the alternatives seem to be more
performance than purely pentane(s) blown foams. The
diversified. Currently, two categories of compounds are
coblown foam has lower cost than HFC blown foams.
being pursued. One category is HFCs, ranging from
In this study, we choose HFC -134a, a commercially
already commercialized HFC -134a to products such as
available blowing agent, as a coblowing agent with
HFC-365mfc and HFC-245fa, which are not presently
pentane(s). We make polyisocyanurate (PIR) foams with
commercially available. The physical properties of
HFC-134a as a coblowing agent with n-C5, i-C5, i-C5/c-
HFCs range from gas (HFC -134a, b.p. = -26.4 °C) to
C5, and the levels of HFC -134a are 10 and 20 mole%.
near liquid (HFC -245fa, b.p. = 15.3 °C) to true liquid
We determine thermal conductivity (or k-factor) of
(HFC-365mfc, b.p. = 40.2 °C). The advantages of HFCs
foams from - 8 to 40 °C. The results show that the
are that they are non -flammable or they have low
effects of HFC-134a on the low temperature k-factors are
flammability, and they provide low thermal conductivity
different, depending on the pentane isomers.
which is critical to rigid polyurethane insulating foams.
Furthermore, we screen a series of surfactants to
The other category of zero ODP alternatives is
determine their effects on the k-factors.
hydrocarbons. They include cyclopentane, isopentane,
normal-pentane as well as butane isomers, which are
INTRODUCTION
mostly used as coblowing agents. Hydrocarbons have
advantages of being low cost, low GWP and are
Since the phaseout of CFC-11 in rigid polyurethane commercially available. The disadvantages of
foam applications, HCFC -141b has become the hydrocarbons are that they are highly flammable and
predominant alternative due to its desirable properties: they are volatile organic compounds (VOCs), and may

POLYURETHANES EXPO 2001 , SEPTEMBER 30 -OCTOBER 3 , 2001 339

be regulated. Safety and environmental measures needed RESULTS AND DISCUSSION
to handle these issues result in additional costs to the
final products or may preclude their use. Table 1 lists physical properties of several current and
Cyclopentane, isopentane and normal-pentane are next generation rigid foam blowing agents. Generally
liquid at ambient temperature. While the liquid behavior speaking, alternative blowing agents do not have as low
is desirable for handling, it is not ideal for certain foam gas thermal conductivity as HCFCs. The alternative
properties. For instance, one may have to increase foam blowing agents also face issues such as flammability,
density because of the condensation of blowing agent(s) volatile organic compound (VOC), GWP and increased
in the cells, which may result in shrinkage of foams. cost.
Furthermore, the condensation of blowing agents in the Table 2 lists the formulations of pentane foams used in
cells will increase thermal conductivity because a liquid this study. When HFC -134a is added, the same molar
has much higher thermal conductivity than a gas. The amount of pentane or mixture was removed from the
condensation of blowing agents is especially undesirable formulation.
for low temperature applications such as cold storage, Pentanes, in particular c-C5, have relatively high
freezers and refrigerators. One of the solutions to this boiling points. While there is an advantage to be a liquid
issue is to use a gaseous blowing agent as a coblowing for handling, blowing agents with high boiling points
agent with liquid pentane(s). For example, normal- cause condensation of blowing agent vapor in the foam
butane and iso-butane have been used together with cells. This results in increases in thermal conductivity at
cyclopentane to reduce foam density [1, \ Unlike low temperatures because the thermal conductivity of a
pentanes, butanes are gas and are difficult to handle. It is liquefied blowing agent is much higher than a vapor [3].
therefore useful to employ a nonflammable gas such as The onset of vapor condensation of a blowing agent
HFC -134a as a coblowing agent with pentanes. depends on, among other things, its boiling point and its
Incorporation of a nonflammable gas such as HFC-134a solubility in polymer matrix.
into pentane is also expected to improve fire Figures 1-4 show initial k-factors of c-C5, i-C5, n-C5
performance of purely pentane(s) blown foams because and c-C5/i-C5 (50/50) at -8 , 0, 10, 24 and 40 °C. For n-
of reduction of fiiel contribution from highly flammable C5, i-C5 and c-C5/ i-C5blown foams, the inversion of k-
pentanes. factor (k -factor increases as temperature decreases)
begins around 24 °C. For C-C5 blown foam, an inversion
EXPERIMENTAL of k-factor starts at 40 °C, and the degree of the inversion
is not as significant as of n-C5 and i-Cs blown foams.
All foams were produced with a high pressure Interestingly, k-factor of c-C5 blown foam begins to
impingement foam machine by Edge-Sweets. The “A” decline again as the temperature decreases below 10 °C,
side contains polymeric methane diphenyl diisocyanate which is dramatically different from i-C5, n-C5 and c-
(Mondur 489) only. The “B” side consists of a polyester C5/i-C5 (50/50). It is unclear why HFC-134a/c-C5
polyol, a surfactant, a fire retardant, catalysts and system has different behavior from others. We speculate
blowing agent(s). The temperature of the “A” and “B” that high solubility of c-C5 in polymer matrix may
sides were kept at 75-80 °F. contribute to this unusual behavior.
The ratio or iso index of the “A ” and “B” sides was Figures 1-4 also show initial k-factors of C-C5, i-Cs, n-
adjusted by varying the speed of “A” and “B" side C5 and c-Cs/i-C5 (50/50) with 10 and 20 mole% of HFC-
pumps. The “A” and “B” streams were mixed together at 134a as a coblowing agent. The addition of HFC-134a
a pressure above 2000 psi. The predetermined amount of has mixed effect on the initial k-factors of coblown
mixed “A” and “B” sides, controlled by the shot time, foams measured at 40 °C and 24 °C, depending on the
was dispensed into a pre -heated mold (150 °F). The type of pentanes. For example, the addition of 10 and 20
mold was kept closed for few minutes before releasing mole% of HFC -134a has no significant effect on the
the foam. initial k-factors of the coblown foams measured at 40 °C.
Thermal conductivity, or k-factor, measurements As to k -factors measured at 24 °C, the incorporation of
(ASTM C518) were conducted by LaserComp Heat Flow HFC-134a, in most cases, improves the initial k-factors.
Meter Instrument (Model Fox 314a) at several mean As shown in Figures 1-4, the improvement for k-factors
temperatures. They are - 8, 0, 10, 24 and 40 °C. For at lower temperatures (< 10 °C) is consistent. In
each temperature, the temperature difference between hot addition, the improvement in k-factors are more
and cold plates is kept constant at 24 °C. All samples are significant with higher level of HFC -134a (20 mole% vs.
core foams with thickness of 1.0 inch. Initial k-factors 10 mole%), regardless of the type of pentanes and
are taken within 24 hours after removing the foam skin mixtures. Aging does not seem to alter the effect of
with band saw. Aging is conducted with the core foam HFC-134a on the k-factors of pentane blown foams, as
at ambient condition. shown in Figure 5.

340 / Wu, Dillon

 Table 1. Physical Properties of Alternative Blowing Agents
Lower ODP Zero ODP
141b 142b 22 c-Cs n-Cs i-Cs 134a 245fa 365mfc
B .P (°C) 32 -9.2 -41 49.0 36.0 28.0 26.5 15.3 40.2
ODP 0.11 0.065 0.055 0 0 0 0 0 0
G W P l0 0 v r 630 2000 1700 11 11 11 1300 820 840
voc No No No Yes Yes Yes No No No
kaa«@ 25 (°C) m W /m .K 9.7 11.5 11.0 12.0 15.0 15.0 13.6 12.2 10.6
Flash Point (°C) None N/A N/A -37 <-40 <-40 N/A None 25
Flam. Limit (% Vol) 7.4-15.5 6.9-15.5 None 1.5-8.7 1.4-7.8 1.3-7.6 None None 1.3-7.6

Table 2. Formulation of pentane(s) blown foams

C om ponent (part) i-C5 c-C5/i-C5* c-C5 n-C5
STEPANPOL® PS 23521 100 100 100 100
P O L Y C A T -5 * 0.19 0.19 0.19 0.19
PO LYCAT* - 46* 0.33 0.33 0.33 0.33
D A B C O * K -15 “ 2.98 2.98 2.98 2.98
TEGOSTAB® B - 8462 ’ 2.0 2.0 2.0 2.0
A n tib la z e *8 0 s 10.0 10.0 10.0 10.0
i- C5 24.91 12.46
c - C5 12.11 24.22
n - C5 24.91
M on d u r* 4 89 ' 170.51 170.51 170.51 170.51
Iso index 300 300 300 300
*: a mixture of 50 mole/50 mole
1Stepan Company; 2,3,4Air Products; 5Goldschmidt Chemical Corporation; 6Rhodia, Inc.;
7Bayer Corporation

Figure 1: Initial k-factor of c-C5 and HFC-134a/c-C5 coblown foams

Wu, Dillon / 341

 Figure 2: Initial k-factor of c-C5/i-C5 and HFC-134a/c-C5/i-C5 coblown foams

Figure 3: Initial k-factor of i-C5 and HFC-134a/i-C5 coblown foams

342 / Wu, Dillon

 Figure 4: Initial k-factor of n-C5 and HFC-134a/n-C5 coblown foams

Figure 5; One month aged k-factor of c-C5fi-C5 and HFC-134a/c-C5fi-C5 coblown foams

Wu, Dillon / 343

 Traditionally, k-factors have been measured at a mean by 4%, c-C5 foam by 5% and i-C5/c-C5 (50/50) foam by
temperature of 24 °C, and most comparisons are made at 12% .
this temperature. In reality, this approach may not be
valid as foams are used in applications under wide REFERENCES
temperature ranges, especially for low temperature
applications such as cold storage facilities, freezers and 1. P. Altoe, G.H. Dean, V. Parenti and P. Clavel.
refrigerators. Our results demonstrate that the addition 1998. Proceedings of Polyurethanes Expo ’98, p
of HFC-134a may not improve k-factors of pentanes 343
blown foams measured at ambient temperature, however 2. V. Parenti, G. H. Dean, P. Clavel, A. Selin, L.
it dramatically improves the k-factors at sub-ambient Zancai, R. Del Perugia. 1997. Proceedings of
temperatures. the Polyurethanes World Congress 1997, p 308
We evaluated effects of several silicone based 3. A. Birch, V. Parenti, K. Duin, G. Smits and P.
surfactants (TEGOSTAB® B -8462 and B-8465 from Clavel. 1995. Proceedings of the Polyurethanes
Goldschmidt Chemical Corporation; Niax® L-5135, L- 1995 Conference, p 448
6912 and L-6900 from CK Witco Corporation; Dabco® 4. J. Wu, D. Dillon, I. Wheeler and A. Albouy.
DC-5598 from Air Products) on the k-factors of 1999. Proceedings of Polyurethanes Expo ’99, p
pentane(s)/HFC -134a coblown foams. We found no 617
significant difference among surfactants.
In this study, we employed up to 20 mole% (about 26 BIOGRAPHIES
wt%) of HFC -134a as the coblowing agent. The amount
of HFC -134a did not cause violent dispensing with the Jinhuang W u is a Research Scientist at Atofina
high pressure impingement mixing head. With 10 Chemicals Technological Center in King of Pmssia. He
mole% of HFC -134a, the dispensing is close to liquid or received his PhD in Macromolecular Science and
creamy like. We expect to see improvement (less froth) Engineering from the University of Michigan, and joined
when one uses polyether polyols to replace some Atofina in March 1997.
portions of polyester polyols because HFC -134a has Doug Dillon is a Research Chemist at Atofina
better solubility (characterized by lower vapor pressure) Chemicals Technological Center in King of Prussia. He
in polyether polyols than polyester polyols [4] received his BS in Chemistry from Juniata College, and
joined Atofina Chemicals in May 1997.
CONCLUSION
APPENDIX - M OLECULAR FORMULA OF
The use of HFC -134a as a coblowing agent with FLUOROCARBONS
pentanes significantly improves the thermal conductivity
of pure pentane(s) blown foams measured at sub-ambient CFC-11, CC13F; HCFC-22, CHC1F2;
temperatures. The degree of improvement depends upon HCFC-141b, CH3CCI2F; HCFC-142b, CH3CC1F2;
the type of pentanes and mixtures. For k-factors HFC-134a, CF3CH2F; HFC-245fa, CF3CH2CHF2;
measured at 0 °C, the addition of 10 moie% of HFC- HFC -365mfc, CF3CH2CF2CH3
134a improves k-factor of n-C5 foam by 6%, i-C5 foam

344 / Wu, Dillon

 SESSION J
INNOVATION AND NEW APPLICATION IN CASE

TECHNICAL SESSION ORGANIZERS

MikePfister, Chair
ARCH CHEMICALS

Kurt Frisch, Jr.

BAYER CORPORATION

Chris G odinich
HUNTSMAN CORPORATION

TECHNICAL SESSION MODERATOR

Steve Seneker
ANDERSON DEVELOPMENT CO.

TECHNICAL SESSION VICE - MODERATOR

Jayne Mallwitz
ARCH CHEMICALS
New UV-Curable Coatings for Caul Paper Manufacturing

DIRK BONTINCK a n d MIKE IDACAVAGE

GUIDO VANMEULDER UCB Chemicals, Inc.
UCB Chemicals 2000 Lake Park Drive
33 Anderlechtstreet Smyrna, Georgia 30080
B-1620 Drogenbos
Belgium

ABSTRACT oligomers must be embossed and cured at the same time,

which is only possible when curing through the back of the
The use of waterborne lacquers to produce high-performing paper to avoid sticking to the embossing roll. Therefore, only
coatings is gradually increasing in importance. Many high-energy radiation curing (electron beam) can be used.
applications for wood and plastic coatings have been Embossing and curing of the coating must be carried out at the
described and realized. In many industries, the development of same time because of its liquid nature before and its non­
technologies having a lower chemical impact on occupational thermoplastic behaviour after curing. The result is a difficult
health and environment is welcomed. In spite of serious and expensive manufacturing process.
efforts of the chemical industry to meet these challenges, the
switch to these new technologies is difficult and slow. In many UV-CURABLE POLYURETHANE DISPERSIONS
cases a higher raw material cost and more critical application
conditions are cited as the most important drawbacks. The combination of polyurethane dispersions and Radiation
This paper will focus on UV -curable polyurethane Curable technology has led to the development of waterborne
dispersions for the manufacturing of caul paper industry. It UV-curable polyurethane dispersions (UV-PUD). The
will be shown that in this particular case, the characteristic chemistry is outlined in Figure 1. An anionic, isocyanate-
behaviour of UV -PUD can offer distinct application terminated polyurethane prepolymer is reacted with an
advantages, resulting in a less expensive, more versatile hydroxylated acrylate, and is then dispersed into water using a
manufacturing process, in addition to ecological and tertiary amine as a neutralizing agent for the carboxylate
occupational health issues. functions. Prepolymer chain length and functionality of the
hydroxy acrylate can be varied in order to come up with the
CAUL PAPER desired crosslink density. The use of non-irritating acrylating
oligomers enables the manufacturing of low irritation coating
Caul paper is a technical term that encompasses several type systems.
of embossing papers and casting foils, either paper or plastic A comparison of UV -PUD with PUD and 100 % UV
based. These ‘papers’ are used to transfer a functional surface systems is shown in Table 1. It is clear that waterborne UV
texture or grain onto a coating. Typical examples include benefits from the advantages of both 100 % UV and PUD
artificial leather (PU or PVC based), furniture foils, shoes, systems, without suffering from any of their obvious
upholstery, handbags, attache cases and suitcases. disadvantages. Waterborne UV-PUD exhibit a superior
There are three categories of embossing ‘papers’ based on balance in chemical resistance and cold flexibility. The
the level of performance: conventional coatings where release improved flexibility can be explained by the fact that their
properties are obtained by silicone or fatty acid chromium UV-crosslink density is lower than for 100 % UV systems,
complexes, extruded film/paper laminates, and Ultracast® meaning that the average molecular weight between crosslinks
based on 100% acrylate resins cured by electron beam. The is higher (Figure 2). This enables the design of UV-PUD that
latter is designed for high perfromance, and is the most are tack-free before cure, meaning that they are non-blocking
widespread in the embossing industry [1] [2]. after water flash-off and before UV-curing. This feature is
It can be argued however that the Ultracast® technology has very interesting for the manufacturing of caul paper: the
a high operational cost due to the electron beam curing and a surface can be embossed thermally before curing, without
narrow application window. This can be derived from the fact sticking to the embossing roll. If necessary, the uncured paper
that a liquid layer of a mix of acrylated monomers and can even be wound up and cured with conventional UV-lamps

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 347

 Figure 1. Chemistry of UV-PUD

100 % UV UV-PUD PUD

Organic cosolvent-free OK OK 3-14 % NM P
Monomer/Oligomers Reactive diluents Water Water
96% EtOH resistance OK OK Limited
Stain resistance OK OK OK
Hard yet flexible Limited OK OK
Sprayable Limited OK OK
Thickness control Limited OK OK
Tack free before cure Limited OK OK
60° gloss range 30-90% 5 -80% 5 70%
Monomer deep -cure Problematic OK NA
Adhesion to plastics Problematic OK OK

Table 1. Comparison of technologies

UV-PUD #1 UV-PUD #2 UV-PUD #3 UV-PUD #4

meq. Acrylic/g of solids 1.91 3.29 3.93 3.89
Approx. Mn 15,000 12,000 8,000 8,000
% solids 36.0 36.3 40.0 42.5
Mean Av. part, size (nm) 56 51 85 91
pH 7.1 7.1 7.2 7.1
Brookfield Viscosity 34 29 45 85
25°C at 50 RPM (mPa.s)
Isocyanate Aliphatic Aliphatic Aliphatic Aliphatic
Polyol Polyester Polyester Polyester Polyester

Table 2. Characterization of UV-PUD used in this study

348 / Bontinck, Vanmeulder, Idacavage

 Figure 2. Relative crosslinking density representation of a 100 % UV system (left) and a UV-PUD (right)

Supplier Parts Application on paper is preferably done by roller coating in

UV-PUD UCB 100.0 reverse mode. Flash -off for a 90 g/m2 wet coating is achieved
TS100 Degussa 1.0
in 1 minute at 80°C. Curing is carried out with two Hg-lamps
Byk 346 Byk 1.0
Byk 028 Byk 0.5 at 80 W/cm, at a typical line speed of 10 m per minute.
Irgacure 500 Ciba 1.5
Ucecoat XE430 UCB 0.7 INFLUENCE OF CROSSLINK DENSITY AND
MOLECULAR W EIG H T BETW EEN CROSSLINKS
Table 3. Formulation for waterborne UV-PUD
It is obvious that a UV -cured coating becomes harder and
more rigid as the number of acrylic double bonds increases.
at another time or at another place since the coating on the The chemical and solvent resistance should be expected to
casting foil does not have the tendency to stick to the backside increase accordingly. The influence of crosslink density and
of the foil. Furthermore, since these waterborne UV- molecular weight between crosslinks on the performance of
crosslinkable polyurethane dispersions can be flattened to a UV-PUD was reported during Polyurethanes 2000 in Boston
broad range of gloss values, the final gloss of the paper is not [3]. Although this study related to the use of UV-PUD
linked to the embossing roll. In practice, this means that with coatings for resilient flooring, many of the characteristics are
one and the same textured roll, different gloss levels can be also valid for caul paper: an optimum level of solvent
obtained. The use of UV -PUD results in a higher degree of resistance and flexibility remains of paramount importance.
freedom in both process and design, as gloss, thickness and The UV-PUD products used in this study are characterized
surface embossing can be arranged for on the same coating in Table 2 and were formulated according to Table 3. All
equipment. grades are aliphatic and polyester based polyurethanes.
The lack of crosslink density, necessary to obtain good solvent
and chemical resistance, is balanced by the presence of a EMBOSSING
polyurethane. The typical urethane hard domains provide
excellent chemical resistance, even at low UV-crosslink Several UV-PUD, differing in UV-crosslinking density and
density. hence coating hardness, were coated on a 150 g/m2 type of
A typical formulation for waterborne UV-PUD is shown in paper. Typically, 90 g/m2 of wet UV-PUD was applied,
Table 3. Next to the binder itself, a matting agent, a wetting yielding 30 g/m2 of dry coating after drying at 80°C for 1
agent, a wax, a photo -initiator and a rheological modifier are minute. All coatings represented in Table 4 showed to be tack-
being used. This formulation results in a Brookfield viscosity of free after water flash-off: wiping the surface with talc and
1000 mPa.s and a 60° gloss of 25%. cotton-wool leaves no trace of talc particles.

Bontinck, Vanmeulder, Idacavage / 349

 The embossing is carried out at 150°C for 4 seconds with a
pressure of 14 kg/cm2. According to the degree of tackiness at Supplier parts Release Migration
150°C, related to the molecular weight before UV-curing, (150°C)
embossing is poor (lower molecular weight) to OK for the UV-PUD #3 UCB 100 Difficult -

+ Byk 333 Byk 0.2 Difficult No

highest molecular weight UV -PUD #1. After embossing, all
+ Byk 306 Byk 0.2 Difficult No
caul paper samples could be wound or stacked without any + Byk 307 Byk 0.2 Difficult No
sticking or loss of definition at room temperature. UV-curing + Add 51 Dow Corning 2.0 Difficult Yes
can be carried out at any given time, either on-line or after + Add 51 Dow Corning 5.0 OK Yes
transportation to a UV-line. + Add 51 Dow Corning 10.0 OK Yes
UV-PUD #4 UCB 100 Excellent No

Approx. Mn Embossing Table 6. Silicone modifications of UV-PUD #3

UV - PUD #1 15,000 OK
UV - PUD #2 12,000 Acceptable
UV - PUD #3 8,000 Poor SOLVENT RESISTANCE
Table 4. Embossing performance of UV-PUD In order to avoid the attack of the caul paper by solvents
during the application of the top coat in a transfer coating
process, solvent resistance is of paramount importance. The
RELEASE lifecycle time of the paper is greatly influenced by its
resistance to solvents.
This caul paper property is perhaps the most important one; All UV-PUD described in this paper were found to be
the affinity and adhesion towards the embossed coating should resistant to toluene, isopropanol and dimethylformamide.
be as low as possible. The caul paper needs to be ‘hard’ These as the most aggressive solvents used in transfer coating.
enough to ensure fast and correct release of the applied Table 7 shows the solvent resistance results after 5 cycles of
coating during the image transfer process. transfer coating using typical solvent based polyurethanes
Since in practice, the transfer is carried out immediately (Ucecoat FN 301 and Ucecoat TCM from UCB Chemicals) on
after drying the applied coating, temperatures can be as high the same sample of caul paper. None of the UV -PUD showed
as 150°C. It was found that only the hard, densely UV- any damage to the embossed texture of the paper; gloss and
crosslinked PUD (#3) is temperature resistant enough (no detail were not affected. The caul papers with the highest UV-
softening) to ensure a correct release at 150°C. The relative crosslinking density however showed less softening by the
release performance of the UV -crosslinked PUD is solvents.
represented in Table 5, and reflects the easiness with which
the dried coating can be torn from the paper.
5 Transfer cycles Ucecoat TCM Ucecoat FN301
(Solvent) (DMF) (Toluene/IPA)
Release Release UV - PUD #1 OK OK
(100 ° C) (150 ° C) UV - PUD #2 OK OK
UV - PUD #1 Borderline Impossible UV - PUD #3 Excellent Excellent
UV - PUD #2 OK Very difficult UV - PUD #4 Excellent Excellent
UV - PUD #3 OK Difficult
Table 7. Solvent resistance of UV-PUD caul paper
Table 5. Release performance of UV-PUD

MULTILAYER UV -PUD CAUL PAPER

The surface tack of UV -PUD #3 can be decreased using
silicones for easy release. Table 6 reveals that silicone gum The results reported so far seem to indicate that none of the
Add51 (Dow Corning) is particularly suited to improve the UV -PUD can fully meet the requirements regarding
release properties of UV -PUD #3 coated caul paper. However, embossing and release. However, using a softer, less UV-
after 3 days, all silicone additives had migrated to the surface croslinked PUD main coat (UV -PUD #1) and a thin finish of a
and were eventually removed during the coating process, harder, more densely UV -crosslinked PUD, a caul paper
again resulting in difficult release behaviour at higher exhibiting both properties (thermoplasticity for the embossing
temperatures. It was found that the incorporation of 2.5 % of a step and hardness for the correct release after transfer coating)
silicone modified urethane acrylate (IRR 154 from UCB can be made. Table 8 describes the composition of the
Chemicals) into UV-PUD #3 (combination shown as UV-PUD multilayer UV-PUD caul paper.
#4), during UV -irradiation, provides the necessary reduction
of surface tack without risk of migration.

350 / Bontinck, Vanmeulder, Idacavage

 1SI Layer 2na Layer Embossing Release BIOGRAPHIES
(150°C)
3 0 g /m 2 o f #1 OK Poor
D irk Bontinck
3 0 g /m 2 o f # 4 Poor OK
2 0 q /m 2 o f #1 1 0 g /m 2 o f # 4 OK OK
Dirk Bontinck is currently
Table 8. Multilayer UV-PUD caul paper Research Manager
Polyurethanes at UCB
Chemicals in Drogenbos,
GLOSS Belgium. He joined UCB in
1987, when he was appointed
As the gloss level of Ultracast® paper is determined by the Lab Manager Thermoplastic
gloss level of the embossing roll, the manufacturing of caul Polyurethanes. He received a
paper with different gloss levels needs to be carried out with PhD in Chemistry from the
different embossing rolls. Since UV -PUD can be formulated Ghent State University,
into a wide gloss range, the resulting paper gloss level is Belgium, in 1987. Since then,
independent from the embossing roll. In Table 9 , it is shown he has been responsible for the
that a 60° gloss range between 10 and 80 % is accessible. development of thin TPU film
for lamination, waterborne
PUD and UV-PUD for clear lacquer applications.
1st Layer 2™ Layer OK 412 (%) 60° Gloss
(dry matter) (dry matter) (Degussa) (% ) Mike Idacavage
2 0 g /m 2 o f #1 1 0 g /m 2 o f # 4 0 80
2 0 q /m 2 o f #1 1 0 q /m 2 o f # 4 2 10 Mike Idacavage is currently
Vice President Research and
Table 9. Gloss as a function of flattening agent (2nd layer) Technology with UCB
Chemicals Corp. in Smyrna,
Georgia. He joined UCB in
CONCLUSION 1996, when he was appointed
Technical Manager Radcure. He
In addition to the technical advantages cited above, received his PhD in Chemistry
waterborne UV -PUD systems have a clear advantage over from Syracuse University, NY
many competing technologies in relation to occupational in 1979. He has been in the
health and environmental issues. Radiation Cure industry since
It has been shown that UV -PUD bring process benefits 1984. His current
without compromising the performance of caul paper: responsibilities cover the areas
embossing and release can be obtained with a two-layer UV- of Radiation Curing and Powders Research, Analytical and
PUD system. Both UV -PUD layers can be cured in one single Engineering for UCB Chemicals in the Americas.
pass under a UV-lamp. Water based UV may never fully
replace 100 % UV systems for caul paper manufacturing, but Guido Vanmeulder
they could well be an interesting alternative in terms of
improved gloss control and process freedom, requiring only 1
Guido Vanmeulder is currently
embossing roll per type of texture. The tack-ffee before cure Fabric Coating Application
properties of UV -PUD allow for the separation of coating and
Engineer at UCB Chemicals in
curing. Since the bottleneck obviously is the embossing Drogenbos, Belgium. He joined
process of the paper, UV -curing could be reduced to one UCB in 1984 as a Fabric
dedicated line at higher speeds, avoiding the installation of Coating Application Technician.
expensive EB tunnels on all production lines.
He graduated form the
Technische Hogeschool Ter
REFERENCES Biest in Sint-Truiden, Belgium
in 1982. His present
1. US patents 4,289,821 - 4,322,450 - 4,327,121 - 4,840,575
responsibilities include the
2. Gray, L. 1988. “Chemical and Cure Parameters That
introduction and Technical
Influence Functional Coatings,” presented at the RadTech
Service of polyurethane fabric
Conference, 1988.
coating systems in Asia.
3. Bontinck, D., M. Idacavage. 2000. “New UV-Curable
Coatings for the Resilient Flooring Industry,” presented at
Polyurethanes Conference 2000, October 8-11, 2000.

Bontinck, Vanmeulder, Idacavage / 351

A Novel Plasticizer-Free PU Gel Technology for
Improved Comfort

LUIGI PELLACANI a n d VERENA M. T. THIEDE

Dow Deutschland GmbH & Co. OHG
Theodor-Schwarte Strasse 2
59227 Ahlen
Germany
effort is needed in producing elastomeric components. PU
ABSTRACT elastomers are generally defined as segmented copolymers
consisting of soft segments (typically featuring polyether or
The versatility of polyurethanes, both in finished product polyester backbone), and hard segments. The urethane or
properties as well as in ease of production and application, urea linkages deriving from the isocyanate component are
is unmatched among the various families of polymers. The the main constituents of the hard segments. Phase separation
polyurethane chemistry allows the manufacture of a wide occurs in these polymers due to the thermodynamic
range of products from soft foams to compact elastomers, incompatibility between the hard and the soft segments. The
from rigid foams to integral skin items. This paper describes extent of the phase separation and the ratio of hard to soft
a novel formulation technology that is suitable for the segments have a direct impact on the physical mechanical
manufacture of a special type of super-soft polyurethane properties of the PU elastomers. Polyurethane gels may be
elastomer, also called polyurethane gel. Dow successfully considered as a special class of polyurethane elastomers,
commercializes these polyurethane gels under the trademark characterized by a very low hardness, namely by a very low
VORASTAR*.
ratio of hard to soft segments.
VORASTAR* polyurethane gels are designed for use in a PU gels represent a class of polymers that today is
large variety of consumer and industrial applications. They attracting increasing attention from a widening audience,
are typically applied as non-foamed derivatives. Thanks to due to the potential of these products from both a scientific
their molecular architecture, VORASTAR* polyurethane and commercial standpoint. PU gels have the valuable
gels feature improved properties as well as improved ability of combining softness, comfort, appearance,
performance in the final application. vibration-dampening properties, and versatile processing, all
This paper describes some of the recent developmental
at a very effective cost/performance balance. These
work that took place within Dow Polyurethane Systems
properties are highly valued in many different applications,
business, and that resulted in a novel polyurethane gel
such as bicycle saddles, mouse wrist pads, furniture seating,
technology.
vibration dampening inserts, and footwear in-soles.
Dow’s objective was to develop a range of PU gel
When referring to the novel PU technology described in
products capable of meeting current application comfort and
this paper, the term “gel” may be misleading, since it
safety standards, along with improved dampening properties
generically refers to a physical state of matter whose
and good dimensional stability at high temperatures.
properties lie between those of a liquid and those o f a solid.
VORASTAR* polyurethane gels are designed to be used
More accurately, a gel may be defined as a coherent mass
without plasticizers and are available in wide range of
consisting of a liquid in which particles too small to be seen
hardness. The applications described in this paper range
in an ordinary optical microscope are either dispersed or
from shoe insoles to anti-pressure medical cushions, from
arranged in a three-dimensional network structure
mouse wrist pads to bicycle saddles.
throughout the mass. According to this definition, a gel may
be very elastic and soft (such as gelatin), or it may be quite
solid and rigid (such as silica gel, a material that looks like
INTRODUCTION
white sand and that is used as a dehumidifier). With
reference to the above description, the term “polyurethane
Thanks to its castability from a viscous liquid to a solid
gel” would indicate a material that includes a dispersed
via a crosslinking chemical reaction, polyurethane has
urethane polymeric network, formed by the reaction of a
gained high acceptance for use where a minimal fabrication

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 353

poly-functional polyol with a poly-functional isocyanate, and it represents a very valuable feature in the final
plus a liquid component that acts as the dispersant for the application.
three-dimensional network structure. This definition is in A further advantage of the novel PU gel technology
line with the traditional PU gel technology, which is described in this paper is that the versatility of the gel
typically characterized by the presence of a solid technology lends itself to extensive formulation tailoring.
elastomeric matrix, combined with a liquid non-reactivt Accordingly, the novel PU gel technology is available based
component. The non-reactive component acts as plasticizer, on TDI, or on MDI, or, for the manufacture of UV stable
and is used to soften the PU elastomeric matrix. The grades, also on aliphatic isocyanates such as IPDI. In the
plasticizer may be for instance a non-reactive diluent, such present paper, the focus is on the MDI formulation
as a dialkyl-phthalate, or it may be unreacted polyol or a approach. As part of the formulation versatility, it is possible
low MW urethane-polyol derivative if the PU gel has been to change the reactivity, as well as the physical-mechanical
manufactured through reaction of the isocyanate component properties such as softness, elasticity, tear and tensile
with a stoichiometric excess of polyol component (low- strength, of the novel PU gel two-component system over a
index reaction). Since the plasticizer is not bonded to the wide range. Furthermore, the novel PU gel described in this
polymeric matrix, it may migrate. The migration of the paper is also suitable to be loaded with fillers or pigments or
plasticizer may cause various negative effects, such as other additives, in order to provide the gel with specific
leakage of plasticizer, as well as a progressive loss of characteristics, such as further improved dampening
elasticity and increase in brittleness of the PU gel itself. properties. An additional feature of this novel PU gel is that
This paper outlines the development work that took place it may also be formulated to be completely transparent: in
within Dow Polyurethane Systems R&D laboratories, and this case, customized colors may be used to increase its
that resulted in a novel PU gel formulation technology. The jewel-like appeal. The transparency of the PU gel can be
most important difference between the traditional PU gel used to enhance or customize the appearance of the final
technology and the novel PU gel technology developed by item. In order to manufacture bubble-free material for
Dow, is that the latter does not require the presence of aesthetic reasons, the "see through" transparent PU gel may
plasticizers or of any other non-reactive component in the require a special manufacturing procedure, such as the use
final product. Given the absence of the liquid dispersant of vacuum on the machine tanks and/or on the molds.
component, the novel PU gel technology developed by Dow Table 1 lists the benefits to be considered when
is related more to the manufacturing of super-soft identifying possible applications of the novel PU gel
elastomers, rather than to the manufacturing of a traditional technology.
“gel”. The absence of plasticizers is an important advantage,

Table 1. PU gel: traditional technology ve. novel technology

Feature Traditional PU gel technology New PU gel technology
C an b e m olded in p resen c e of inserts Yes Yes
Property ch an g e upon cooling Small Small
Dimensional stability Typically, medium to poor Typically, medium to good, even at high
tem perature
Optical a p p earan ce Transparent: may be opague T ransparent
Type of chem istry Typically TDI - polyether polyol Available in various versions: TDI or
MDI or IPDI based - polyether polyol
Shock absorption capabilities Yes Yes
Ability to reproduce mold surfaces Good Good
P re sen ce of plasticizer or other non - Typically p resen t Typically ab sen t
reacted ingredient
Machine eguipm ent 2K low pressure casting m achine 2K low pressure casting m achine
Curing conditions Medium to room tem perature Medium to room tem perature
Mold tem perature and demoiding time Medium to high tem perature may be Medium to room tem perature - few
needed - few m inutes demolding minutes demolding
Demolding Relatively ea sy Easier, thanks to the reduced stickiness
Possibility to adjust the reactivity with Yes Yes
catalyst
Possibility to tailor the physical mechanical Yes Yes
properties

KEY PROPERTIES AND APPLICATIONS

Typically, the hardness of PU gel materials goes from 0
The super-soft PU elastomers described in this paper are ShA (very soft gels) up to about 20 ShA (relatively hard
polymeric materials that are defined as PU gels, featuring a gels). When the hardness rises up to and above 20 ShA, then
combination of two key physical-mechanical properties, rather than referring to PU gels, we speak of soft elastomers
namely the softness and the visco-elasticity. These two (see Figure 1). As the hardness rises even more, then we
properties are at the basis of the specific applications where speak of elastomers and of very hard materials such as
PU gels are being used today. casting resins. The formulation technologies as well as the

354 / Pellacani, Thiede

 applications for elastomers and for casting resin materials • Office supplies, such as the filling in mouse wrist pads.
are different from those typical for PU gels. With the increasing amount of time spent using the
Hardness is measured as indentation hardness using the computer both in the office and at home, wrist injuries
appropriate hardness-measuring device (durometer). Each are becoming more and more prominent among the
type of durometer has a different combination of indenter Repetitive Strain Illnesses (RSI). Wrist injuries are due
shape and spring strength, and the ShOO durometer is to, among other things, improper positioning of the
suitable for measuring the hardness of soft materials, such as forearms, wrists, and hands while operating the computer
PU gels (see Figure 1). In traditional PU gels, the low keyboard. The use of a quality support, in the form of
hardness is achieved thanks to the use of plasticizers or of PU gel padding for the wrist/palm, helps the operator use
other non-reactive components. In the novel PU gel the keyboard more comfortably. As in the previous
technology developed by Dow, the softness of the polymer applications, an additional important advantage of PU
is simply the result of an innovative formulation approach: gel is the possibility of formulating transparent material,
no plasticizers nor other non-reactive components are thus increasing designers creativity and appeal for
needed to achieve low hardness properties. customers.
On top of featuring very low hardness, PU gels are also • In­soles for footwear application. The layer of PU gel
characterized by being non-perfect elastic elastomers. They visco-elastic material helps absorbing shock and
behave as visco-elastic materials, and as such they are for distributing pressure away from problem areas. The PU
instance capable of converting a repeated cyclic gel in-sole also provides an increased degree of comfort
deformation, such as a vibration, into heat. The phenomenon in occupations that require long standing, walking or
according to which the un-recovered energy is converted climbing. For instance, during normal walking, it has
into heat is called hysteresis. PU gels are characterized by been recorded that at each heel strike, the peak forces
relatively high hysteresis, and have therefore a high generated at the tibia (about 5 gravities), are still capable
capability of absorbing and dampening vibrational energy. of reaching the skull (about 0.5 gravities). The
These two key properties, namely the softness and the dampening suggests that there is substantial shock
capability to absorb vibrations, explain why PU gels have attenuation along the muscle-skeletal system between the
found use in the following applications: tibia and the skull, however this is not enough to
• Padding in saddles for bicycles and motorcycles. This completely eliminate the shock transmitted to the skull.
is one of the first applications that gained broad By using visco-elastic shoe inserts made of PU gel, it is
commercial importance. Soft PU gels have high level of possible to achieve a further dampening of the
pressure distribution and shock and vibration absorbing transmitted shock right at its source. Due to the high load
properties, which make them ideal for this application. applied to the gel in-sole, the PU gel material for this
Additional key attributes of some PU gel materials application is typically relatively hard, around 1 0 - 2 0
include softness, shape retention, and transparency. The ShA.
latter is very important in allowing for customized see- • Potting compounds. This is a special application for PU
through inserts. gels, leveraging on the electrical insulating properties of
• Padding in various seating and furniture polyurethanes. Accordingly, in a properly formulated
applications. Thanks to their combination of softness variant, PU gels are also suitable as two-part soft
and elasticity, PU gels can be advantageously used to encapsulating compounds. An additional advantage of
provide extra seating comfort. For instance, layers of PU PU gels in this application is that they may be
gel may be used in combination with layers of PU foam, formulated to be transparent, which makes possible to
with the former “distributing” the weight pressure and perform non-invasive visual external examination of the
preventing it from being concentrated in few points. The status of the encapsulated electrical part.
PU gel is effective in easing any localized pressure on • General noise and vibration dampening. By proper
the back and legs of the person sitting. This is valuable usage of visco-elastic materials such as PU gel, it is
in a broad range of applications, which include furniture possible to reduce the intensity of mechanical vibration
and automotive seating, as well as medical mattresses for and noise, thus increasing the comfort of people exposed
use in operating theatres and in reanimation rooms. In to it. Thanks to their ability of absorbing and converting
the latter case, the PU gel helps prevent the formation of vibrational energy into heat, PU gel materials may find
decubitus sores, thus improving the condition and the use in applications such as anti-vibration gloves, as well
comfort for the patient. The bright and jewel like as in many different types of noise related problems.
transparency represents an additional characteristic of In most of these applications, due to the “jelly -like”
some PU gel materials, as it allows for customized see- characteristics and to the stickiness of the PU gel, a covering
though applications (for instance, transparent PU gel material must be used. Typically, the PU gel is used in
inserts in the arm rests of “premium” chairs). The use of “encapsulated” form, protected by some sort of flexible
see-through inserts and parts will also enhance the covering material such as PVC foil, TPU foil, synthetic
aesthetics of the final product, increasing its leather, and similar. In order to minimize direct handling of
attractiveness for designers and customers alike. the PU gel, the preferred way of manufacturing it is by

Pellacani, Thiede / 355

directly casting the PU gel on the flexible covering material. worsening in the shape retention properties of the PU gel,
Instead, when manufacturing the PU gel separately, to later which may favor the occurrence of permanent deformation,
position it in the final item, the use of some precautions will particularly after prolonged stress or after exposure to high
be necessary in order to handle the PU gel material. The temperatures (70 - 80°C). As a matter of fact, the
latter include, applying some release agent to the mold plasticized PU gel material is held together by bonds that are
before casting the PU gel, and treating the PU gel material weakened by the non-reactive component. At temperatures
with mineral powder to reduce its stickiness. as high as 70 - 80°C, the bonds may weaken to a point that
When using the novel PU gel technology developed by the gel becomes a viscous fluid rather than a soft solid. As a
Dow described in this paper, thanks to the absence of consequence, the PU gel may start losing its dimensional
plasticizers, if the covering material is damaged or stability, showing unrecoverable shape change under the
punctured, there will be no leaking of plasticizer. On the action of stress or even simply of gravity. On the other hand,
other hand, in traditional PU gel materials the plasticizing or the novel PU gel technology described in this paper shows
non-reactive components typically used represent an good shape memory retention properties, and minimal shape
important drawback. Migration of the plasticizer may cause change, even when subjected to prolonged stress or to
leaking as well as a progressive shift in the physical- elevated temperatures. This is a consequence o f the novel
mechanical properties of the PU gel itself (“drying”). The PU gel being a polymer featuring a network structure held
plasticizer may also interact with the covering material, together mainly by strong covalent bonds, and not weakened
causing for instance deterioration in its mechanical by plasticizers or other non-reactive components.
properties. Furthermore the plasticizer may cause a

Figure 1. Range of hardness- applications

water level. Analytical techniques used for the purposes of

MATERIALS AND METHODS this work include:
♦ isocyanate content measurement according to the
Chemicals dibutylamine titration method (ASTM D 5155);
♦ determination of polyol hydroxyl number via titration
All the PU gel systems used for this work are described in (ASTM D 4274);
Table 2. All systems are 100 % solids, formulated two ♦ measurement of water content in polyols according to
component systems, based on MDI and featuring polyether the Karl Fischer titration method (ASTM E 203);
backbones. System A and B are MDI-based systems for the ♦ basicity and acidity determination in polyols and
manufacturing respectively of very soft and of medium soft isocyanates via titration (modified ASTM D 4667);
PU gels. System C is another MDI-based system for the ♦ determination of polyol and isocyanate viscosity as
manufacturing of PU gel with medium softness, and it is per DIN 53019; and
characterized by very low viscosity components. System D ♦ determination of polyol and isocyanate density as per
is a filled system, for use in specialized applications. Before DIN 53217.
use, all reactants were degassed: when necessary, polyols
were subjected to drying treatment in order to control the

356 / Pellacani, Thiede

 Table 2. System s evaluated fo r the purposes o f this w ork
S ystem A S ystem B S y ste m C S y ste m D
D e s c rip tio n MDI-polyether MDI- polyether MDI- polyether MDI- polyether
polyol based polyol based polyol based polyol based
system system system system
K ey F e a tu re High softness Medium softness Medium softness Medium softness
PU gel PU gel PU gel - low PU g el-filled
system viscosity polyol component
Polyol viscosity 23 °C [mPas] 1000 950 600 2800
Isocyanate viscosity 23°C [mPas] 5000 5000 900 5000
Polyol density [g/cm3] 1 1 1 1.6
Isocyanate density [g/cm3] 1.1 1.1 1.1 1.1

samples were manufactured by casting on non-adhesive

Preparation of the isocyanate terminated prepolymers substrates at room temperature.
Both DIN and ASTM test methods were used in the
All the PU gels evaluated for this work were characterization of the physical-mechanical properties of the
manufactured according to the prepolymer method. The PU gels studied.
isocyanate-terminated prepolymers were prepared by ♦ Density was measured according to DIN 53420
controlled reaction of an excess of isocyanate with the (ASTM D 792).
isocyanate reactive compound. A 6-liter reaction vessel ♦ Shore Hardness was measured as ShA according to
equipped with chemicals addition inlet, heating mantle, DIN 53505, as well as ShOO according to ASTM D
electrical stirrer, thermometer, and gas inlet and outlet for 2240. Measurements are performed only on
continuous flow of nitrogen, was charged with the specimens of sufficient thickness to ensure a proper,
isocyanate, and heated to approximately 70°C. sensitive test: generally, for softer material, a thicker
The reaction was performed under stirring, feeding the sample is preferable. For the purposes o f this work,
isocyanate reactive compound into the reaction vessel at the samples to be tested had a minimum thickness of
controlled rate, while maintaining the temperature in the at least 0.4 inch (10 mm). The hardness o f the PU gel
vessel at about 70 - 75°C. Higher temperatures can cause is measured on the “naked” gel, after having removed
harmful side reactions that may affect the final performance any covering material, if present.
of the prepolymer. Samples were periodically withdrawn to ♦ Tear strength was measured according to DIN 53515
determine the isocyanate content. After a total reaction time at 500 mm/min. on a suitable tester device. Each data
of about 4 hours, the isocyanate content was at the is the average of 3 measurements. Measurements
theoretical value. The prepolymer was unloaded after were also performed according to ASTM D 624 - die
stopping the reaction by cooling. C.
The commercial manufacturing scaling up of the ♦ Tensile strength was measured according to DIN
isocyanate-terminated prepolymers for use in the novel PU 53504 on a suitable tester device. Each data is the
gel technology was successfully performed according to a average of 5 measurements. Measurements were also
similar procedure. performed according to ASTM D 412.
♦ Elongation at break was measured according to DIN
Preparation and testing of the PU gels 53504 on a suitable tester device. Each data is the
average of 5 measurements. Measurements were also
Room temperature processing conditions were used for performed according to the ASTM D 412.
the PU gel samples prepared for this work. Samples were ♦ Shape retention properties was measured according to
manufactured according to hand mixing technique, as well the following procedure: a test specimen having a
as using a low-pressure casting machine equipped with thickness of at least 10 mm (0.4 inch), was placed on
vacuum and static mixer. Especially in the case of the hand a metal plate, and then it was bordered with a line.
mixed samples, vacuum treatment before mixing and after The plate with the specimen was placed in an oven at
mixing-casting was used to remove the bubbles caused by a fixed temperature, checking at regular intervals
stirring as well as the dissolved gases. In order to assess the whether the specimen had changed its shape and
reactivity profile o f the PU gels studied for this work, moved over the borderline.
reaction parameters such as gel time were measured. (The ♦ Static compression set test under fixed loading
gel time is the time when a string of the curing polymer can conditions. This test determines the compression set
be pulled up from the reactive mixture using a glass rod, and remaining after deformation of a PU gel cushion by
the string does not break for at least 10 seconds.) The loading it for a period of time with a certain load. In

Pellacani, Thiede / 357

 this case, the applied load was around 2 psi (0.014 The novel PU gel formulation technology approach
N/mm2), and the period of time was 4 days. The test described in this paper allows to successfully combine an
was performed inside a climatic room (22°C ± 1°C, at elevated softness with a good level of physical-mechanical
a humidity level of 50 % ± 2 %). properties, while maintaining a plasticizer-free environment.
♦ Dynamic mechanical properties: the spectra were Table 3 reports selected physical mechanical properties
recorded on a Rheometric Spectroscopy (DMS) RSA measured for the four PU gel systems considered in this
II spectrometer. Analysis were performed in study. Data were measured on PU gel specimens prepared at
compression-mode with a command strain of 0.2 % index 100. The reactivity at which the samples were
and a sinusoidal signal having frequency of 6.28 rad/s manufactured is not specified, since the reactivity is easily
(1Hz). Testing was performed on a cylindrical adjusted by changing the catalyst level in the polyol. The
sample, placed between the two compression clamps reactivity also depends on the temperature of the
of the DMS device, covering the temperature range components. The suggested temperature range is 25 - 45°C.
from -120 to 130°C, with a ramp rate of 5°C/min. The samples manufactured for this study had a gel time of
All the physical mechanical properties were measured on about 8 minutes and a demolding time of about 20 minutes.
PU gel specimens prepared at index 100. The versatility of The four PU gel systems described in Table 3 have
the novel PU gel formulation approach allows to operate medium to low hardness. They are rather weak polymers, as
under-index or over-index, corresponding respectively to indicated by the rather low numbers on the strength
reacting the isocyanate component with a stoichiometric properties. At the same time, some of the systems in Table 3
defect or with a stoichiometric excess of formulated polyol show very high elongation properties (systems C and D).
component. However, in line with the desire to minimize the System C is a recently developed grade based on the
presence of non-reacted molecules in the final product, and innovative formulation concept described in this paper.
to maximize the physical-mechanical properties, the System C further expands on this innovative formulation
preferred ratio at which the isocyanate is reacted with the concept by lowering the viscosity of the components as well
formulated polyol component is the stoichiometric one, as achieving an improvement in overall physical-mechanical
corresponding to an isocyanate index of 100. properties.
It should be pointed out that, due to very high softness of Any PU elastomer system tends to show some sensitivity
the material being tested, the ASTM or DIN test methods to stoichiometry, meaning that the physical mechanical
for tear, tensile and elongation need to be applied with properties are influenced by the ratio of reactive groups in
extreme care in order to maximize the accuracy of the the reactive mixture. This is also true in the case of PU gels.
measured properties. More specifically, in order to achieve the required softness
and visco-elasticity properties, some of the traditional PU
gel materials need to be manufactured around 40 - 70 index.
RESULTS AND DISCUSSION This means that the working ratio is such that the isocyanate
component contains only 40 - 70 % of the NCO groups that
Physical mechanical properties - hardness, tear, tensile are present in the isocyanate reactive component. On the
and elongation other hand, the novel PU gel technology described in this
paper allows to manufacture good performing PU gel
PU elastomers combine properties such as high material even when operating at perfect stoichiometry (i.e.
elongation, high tensile strength and high modulus. This isocyanate index 100). Therefore, the novel PU gel
combination determines their high toughness and durability. technology is preferably to be applied at isocyanate index
The physical-mechanical properties of PU elastomers are 100, since this is expected to maximize the final physical-
largely dependent on the extent of the phase separation mechanical properties. Furthermore, when operating a low-
between hard and soft segments, as well as on the ratio of index reaction (isocyanate component + stoichiometric
hard to soft segments. In the case of the PU gels, the need to excess of polyol component), the PU gel may contain some
achieve very high softness means that the amount of hard non-reacted polyol or some low MW urethane-polyol
segments must be minimized. Phase separation loses its derivatives. The presence of all these derivatives is instead
importance, which is expected to have a negative impact on minimized when manufacturing the PU gel at isocyanate
most physical-mechanical properties, in particular on the index 100.
polymer strength. Indeed, PU gel are typically rather weak,
and in most applications they need to be protected by some
flexible layer, such as for instance a skin of transparent
TPU.

358 / Pellacani, Thiede

 Table 3. Properties of PU gel systems
S ystem A S ystem B S y stem C S y ste m D
Application Technique Hand mix Hand mix Hand mix Handmix
P ro c e s s in g p a ra m e te rs
Iso index 1 1 1 1
Iso.Pol
’ ratio by weight 23/100 23/100 40/100 11/100
Mixing temperature (2C - 2F) 2 5 -7 7 2 5 -7 7 2 5 -7 7 2 5 -7 7
Gel time, minutes (*) n n n
P h y s ic a l p ro p e rtie s
Density (g/cm3) 1.07 1.07 1.07 1.60
Hardness (ShA) 0 1 1 2
Hardness (ShOO) 0 30 30 40
Tensile (MPa - psi) - DIN 53504 0 .1 - 1 5 0.2 - 30 0.2 - 30 0.3 - 45
Elongation (% ) - DIN 53504 270 250 540 410
Tear (kN/m - pli) - DIN 53515 0.6 - 3.4 0.8 - 4.6 1 .3 - 7 .4 2 .0 - 1 1 .4
Tensile (MPa - psi) - ASTM D 412 0 .1 - 1 5 0.2 - 30 NM NM
Elongation (% ) - ASTM D 412 600 450 NM NM
Tear (kN/m - pli) - ASTM D 624 die C 0.6 - 3.4 1 .2 - 6 .8 NM NM

(*) The reactivity may be adjusted by adjusting the catalyst level in the formulated polyol. The gel time may be as fast as 1 minute,
with about 2 - 4 minutes demolding time, or it may be as slow as 1 0 - 3 0 minutes, with proportionally longer gel time.
NM = not measured

Shape retention and compression set properties At the same time, in spite of the absence of plasticizers or of
other non-reacted components, the novel PU gel still
PU gels are non-perfect elastomers, which may be maintains an elevated visco-elastic nature. This makes it
described as visco-elastic materials. The elevated softness very effective for applications requiring not only high
corresponds to an elevated visco component within the dimensional stability, but also a high level o f pressure
visco-elasticity balance. Particularly super soft PU gels distribution or high vibration and shock absorption
(ShOO = 0) and “traditional” PU gels containing plasticizers capabilities.
or other non-reactive components, may behave more as a Similar test performed on PU gel samples manufactured at
liquid than as a solid material. In practical terms, this means decreasing isocyanate index (increasing excess of
that the PU gel may be unable to maintain its dimensional formulated polyol component), did show a progressively
stability, and its shape will change with the passing of time. worse dimensional stability performance.
Applying a load and/or subjecting the PU gel to treatment at System A was also tested according to the compression
elevated temperature may accelerate this shape change set test described previously. The applied load was around 2
phenomenon. psi (0.014 N/mm2), corresponding to around 25 %
Systems A and B were tested according to the shape compression. After being subjected to this load for 4 days,
retention - dimensional stability test described previously. when removing the load, the novel PU gel described in this
Table 4 reports the corresponding results, which show that paper returns very quickly to its original shape, showing
the novel PU gel described in the paper features an negligible set. The extremely low creep during the 4 days
outstanding degree of shape retention properties, even when testing period further supports the excellent compression set
manufactured at very low hardness. The material, as detailed performance.
in the previous paragraph, also has a highly elastic character.

Table 4. Dimensional stability o f PU gel systems

S y stem A S y s te m B
T e m p e ra tu re o f d im e n s io n a l s ta b ility te s t
702C for 3 weeks stable stable
An additional week at 802C stable stable
Two additional weeks at 902C NM stable
Two additional weeks at 1002C NM stable

NM = not measured

Pellacani, Thiede / 359

Dynamic mechanical properties applied strain. DMS testing may be performed in various
ways, for instance it may be performed under tension-mode,
As visco-elastic materials, PU gels at rest have no or under compression mode, or under shear mode. The test
preferential microstructural orientation. The shear forces output will be somewhat different in each case. For the
due to flow cause an overall anisotropic reorganization of purposes of this work, the DMS study on the PU gel System
the microstructure of the visco-elastic material, which B, was performed by holding the frequency constant at 1 Hz
develops a preferential directional orientation. There are two and varying the temperature from - 120°C up to 150°C.
types of energy being spent in producing the anisotropic Higher temperatures were not considered as PU gels are
structural change: a recoverable energy associated with normally used at temperatures lower than 150°C. The DMS
structure formation which is identified with the elasticity, graph was generated under compression mode, measuring
and a lost energy dissipated in structural formation and the storage modulus E ’ and the tan -delta = E”/E’, with E”
sliding which is associated with the viscosity. The energy being the loss modulus. Tan -delta is a measure of the ratio
lost due to the viscous component is transformed into heat, of the energy lost to the energy stored in a strain cycle.
and goes to increase the temperature of the visco-elastic The DMS graph for System B is reported in Figure 2, and
material. The phenomenon according to which the un­ it shows three regions only:
recovered energy is converted into heat is called hysteresis. - The glassy region at low temperature, where the
Rheology is the study of the deformation and flow of molecular motion of the chain segments is essentially frozen
matter. Studying the response of a material to a forced in. Modulus is very high, and tan -delta is very low,
shearing flow is the basis for determining its specific indicating a high elasticity character.
rheological properties. Rheological studies on visco-elastic - At about - 50°C, a transition region corresponding to the
materials are preferably performed using oscillatory flow at glass transition of the soft phase, consisting of the polyether
a selected frequency, because this will reveal both the polyols chains. It is a rather sharp transition, also due to the
viscous and the elastic properties. On the other hand, steady absence of phase mixing (absence of hard segments).
flow would reveal only the viscous properties. - A rubbery plateau: the low tangent delta indicates again
For the purposes of this work, rheological study on the PU very good elasticity. The material shows a low modulus,
gel - System B, was performed using Dynamic Mechanical with a slight decrease in hardness as the temperature
Spectroscopy (DMS). The latter measures the ability of a increases. There is no indication of the presence of a hard
material to store and dissipate mechanical energy. DMS segment phase.
measurements are based on the differences in response of
the viscous and elastic components to a small sinusoidally

Figure 2 DMS graph for PU g e l system B

360 / Pellacani, Thiede

 BIOGRAPHIES
As reported in Table 3, system B has a relatively high
hardness for a PU gel. This is reflected in the relatively high Luigi Pellacani
elasticity, namely in the relatively low value of tan-delta in
the temperature range around room temperature, as reported After completing his
in Fig 2. When considering PU gel material with reduced University education in
hardness, the visco component will increase in importance, chemistry, Luigi Pellacani joined
resulting in increasing values for tan-delta. Dow in 1987 in Italy, at the Dow
Italia site located in Correggio,
Italy. Here, he was involved in
CONCLUSIONS various technical service &
development activities in
Dow is now introducing into the market a novel flexible as well as in SRIS PU
generation of PU gel systems with improved performance. foam applications. In 1992 he
Dow has developed a polymer technology approach that transferred to the Dow's PU
allows manufacturers to prepare PU gels combining high R&D laboratories in Temeuzen,
elasticity with very high softness, as well as excellent shape The Netherlands, where he has
retention and compression set properties. These PU gel been involved in research activities including rigid polyols
systems can be processed according to casting technology, development and products for CASE applications. He is
using suitable low -pressure machines. The most important presently located in Ahlen, Germany, with Dow
novelty aspect o f the new PU gel technology is the Deutschland GmbH Co. OHG., with global development
possibility of manufacturing items that are completely free responsibility for CASE products.
from plasticizers or other non-reacted components.

Verena M. T. Thiede
ACKNOWLEDGMENTS
Verena Thiede joined Dow
We would like to thank for their contributions to this Deutschland Inc. in 1998 at
work: Geert Bultink - Dow Temeuzen Material Science the site located in Ahlen,
group; Torsten Schmidt, Ingolf Pastemok, Mark Lord, Germany after having
Michael Janssen and Holger Paschke - Dow Deutschland received on the same year her
GmbH Co. OHG; Carlo Cocconi and Valerio Sodini - Dow Ph.D. in chemistry from the
Italia S.p.A.; and Netza Lopez and Stacey Blake - The Dow University of Muenster,
Chemical Company, U.S.A.* Germany. After some
research activities in the
areas of flexible PU foam
applications, she moved to
the CASE Research &
Development division, where
she is now responsible for
product development.

* Trademark o f The Dow Chemical Company.

Pellacani, Thiede / 361

 New Non-Isocyanate Curatives for 2K Solvent-Borne
Urethane Coatings

K. B. CHANDALIA, S. L. GOLDSTEIN, M. J . MORGAN a n d R. T. W OJCIK

Lyondell Chemical Company
3801 West Chester Pike
Newtown Square, Pennsylvania 19073

ABSTRACT composition is then sprayed onto a substrate and cured

with heat. The most commonly used catalyst for this cure
LYONDELL has developed a family of novel non ­ is an organotin compound such as dibutyltin dilaurate. The
isocyanate resins. These resins are polyuretidione mechanism involved is thought to be via the cleavage of
polyurethanes that are designed to have low solution the uretidione ring to release the masked isocyanate groups
viscosities for meeting the demands o f high solids coatings which then react with the polyol hydroxyl groups to give
technologies. These resins have high functionality —in the the coating polymer (see Figure 2).
range of 3 to 4 —yielding high performance polyurethane In the mid ­1990’s, we started the development of
coatings when formulated with typical polyol resins in high polyuretidione polyurethane based curatives that were
solids, solvent-borne 2K and pseudo-IK formulations. The suitable for use in solvent and waterborne coatings
resulting coatings have the excellent appearance, systems. The following is an interim report on this work in
mechanical performance and chemical resistance of progress.
conventional isocyanate based 2K urethane systems. This
paper will describe the processing conditions, including
catalysis and bake temperature, used to achieve the
ultimate performance of these curatives in formulations as
well as their application to a number of plastic substrates.

INTRODUCTION

The use of polyuretidione polyurethanes as curatives for

polyols, in the preparation of polyurethane coatings, is well
known. They provide the advantages of having no free
isocyanate groups, thus reducing the potential of worker
exposure, and avoid the problems associated with
conventional blocking agent emissions, while giving
coatings with the excellent properties typical of
polyurethanes. Their major disadvantage is that relatively
high temperatures - 130 to 150°C - are required to achieve
a cure. These curatives are usually prepared by the reaction
o f IPDI uretidione (dimer) with low molecular weight diols
such as 1,4-butanediol (see Figure l ) 1.
As such, these products are solids and find use in
solvent-free powder coatings formulations. The
polyuretidione polyurethanes are blended with solid acrylic
or polyester polyol resins, catalysts and other additives and Figure 1. IPDI/ 1 ,4-Butanediol Polyuretidione Polyurethane
then ground to a fine, intimately mixed powder. This

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 363

 method for preparing polyuretidione polyurethanes based
on HDI dimer that have low solution viscosities2. We
wanted to use HDI dimer (Luxate® HD - 100) because it is
readily available, low viscosity and easy to manipulate. In
addition, we excepted that the reduced steric hindrance of
the hexyl group (from the HDI) flanking the uretidione
ring, as compared to the isophoronyl ring (from IPDI),
offered the potential of achieving lower cure temperatures
and a higher effective functionality during the film forming
reaction.
Typical linear diols, such as 1,4-butanediol and 1,6-
hexanediol, failed to give stable solutions when reacted
with HDI dimer. However, branched diols, such as 1,3-
butanediol and 2-methyl-1,3-propanediol (MPDiol®
glycol) did yield stable, low viscosity 50% solids solutions
in common coatings solvents. Thus, we were able to
prepare two classes of these curatives. In the first, n
Figure 2. Polyol / Polyuietidione Polyurethane Cure by equivalents of HDI dimer was reacted with n+ 1
Thermal Cleavage equivalents of a branched diol to give a mixture of
hydroxyl terminated polyuretidione polyurethane
POLYURETIDIONE POLYURETHANE SYNTHESIS
oligomers. In the second class, n equivalents o f HDI dimer
was reacted with n-\ equivalents of a branched diol and 1
Key to the development of polyuretidione polyurethane
equivalent of a branched monol to give a mixture of alkyl
based curatives, that are suitable for use in solvent and
terminated polyuretidione polyurethane oligomers. (See
waterborne coatings systems, was the discovery of a
Figure 3.)

Figure 3. Preparation HDI Polyuretidione Polyurethanes

364 / Chandalia, Goldstein, Morgan, Wojcik

 If some o f the monol used contains hydrophilic groups, DIAMINE CURED COATINGS
such as a monoalkyl ether o f poly(oxyethylene)gylcol, the
resulting product can be dispersed in water. The ratio of When cross-linked with low molecular weight diamines,
alcohols to dimer was chosen so as to yield an average of these curatives produced coatings with excellent
at least three uretidione groups per oligomer chain. This properties. Example formulations with three polyuretidione
avoids the coating property reduction problems associated polyurethanes cured with isophorone diamine (IPDA), and
with mono- and non-functional oligomers. (Commercial the properties of the resulting films, are shown in Table 2.
HDI dimer, such as HD - 100, contains some isocyanurate Similar properties were obtained with methylene
trimer, so the average chain functionality is somewhat di(cyclohexylamine) (H12MDA). The cross-linking
higher than the idealized three or four.) Typical properties reaction involved is the opening of uretidione ring with the
o f some of these curative solutions are shown in Tables la formation o f a biuret linkage. (See Figure 4.)
and lb.

Table 1a. Alkyl Residue Terminated Polyuretidione Polyurethanes

Composition and Properties - Solutions: 50% in n-Butyl Acetate
Relative Solution
Equivalents On Solids Viscosity,
Solution # Diol Monol Diol Monol Iso %Uretidione %Urethane cP @ 25*C
44-1 1,3-Butane Diol 2-Butanol 4 1 5 14.5 25.5 511
44-2 1,3-Butane Diol 2-Butanol 6 1 7 14.7 25.7 1150
48-1 1,3-Butane Diol 2-Ethyl-1- 4 1 5 13.9 24.3 218
hexanol
84-1 2-Ethyl-1,3- 2-Butanol 4 1 5 13.3 23.3 165
hexane Diol
88-1 2-Ethyl-1,3- 2-Ethyl-1- 4 1 5 12.7 22.3 135
hexane Diol hexanol
88-2 2-Ethyl-1,3- 2-Ethyl-1- 6 1 7 12.9 22.6 348
hexane Diol hexanol
N8-1 2,2-Dimethyl-1,3- 2-Ethyl-1- 4 1 5 13.6 23.8 235
propane Diol hexanol
M8-1 2-Methyl-1,3- 2-Ethyl-1- 4 1 5 13.9 24.3 275
propane Diol hexanol

Table 1b. Hydroxy Alkyl Residue Terminated Polyuretidione Polyurethanes

Composition and Properties - Solutions: 50% in n-Butyl Acetate
Relative Solution
Equivalents On Solids Viscosity,
Solution # Diol Diol Iso %Uretidione %Urethane cP @ 25°C
40-1 1,3-Butane Diol 5 4 14.2 24.9 126
40-2 1,3-Butane Diol 6 5 14.3 25.2 251
M0-1 2-Methyl-1,3-propane Diol 5 4 14.2 24.9 155
MO-2 2-Methyl-1,3-propane Diol 6 5 14.3 25.2 284
N0-1 2,2-Dimethyl-1,3-propane Diol 5 4 13.7 24.0 194
NO-2 2,2-Dimethyl-1,3-propane Diol 6 5 13.9 24.3 385
80-1 2-Ethyl-1,3-hexane Diol 5 4 12.4 21.7 195
80-2 2-Ethyl-1,3-hexane Diol 6 5 12.5 22.0 312

Chandalia, Goldstein, Morgan, Wojcik / 365

 Table 2. Solvent Borne Formulations and Physical Properties: IPDA Cure
P o ly u r e tid io n e P o ly u r e th a n e
S o lu tio n # 44-1 44-1 44-1 88-1 88-1 88-1 8 8 -2 8 8 -2 8 8 -2
E q u ivalen ts R atio,
U retid io n e:A m in e 1:1 1.5:1 2:1 1:1 1.5:1 2:1 1:1 1.5:1 2:1
A p plicatio n M eth od D rw D w n D rw D w n D rw D w n D rw D w n D rw D w n D rw D w n D rw D w n D rw D w n D rw D w n
T a c k F re e T im e , min 10 10 20 10 20 10 20 10 10

F IL M P R O P E R T IE S
U n po lish ed b o n d erite p an els, T w o w e e k s 2 5 ° C /5 0 % R H
G loss
CM
O
o

83 87 75 86 87 82 87 89 85
o
CD
O

104 104 101 10 4 104 101 103 101 104

P encil H a rd n e s s B F HB B HB HB HB HB B
M an d re l Bend p ass p ass pass p ass pass p ass pass p ass pass
C ross H atch A d h esio n
% Rem oved 0 0 0 10 0 0 0 0 10
Im p act, in/lb
D irect >160 >160 >160 >160 140 >160 >160 >160 >160
R e v e rs e >160 >160 >160 >160 12 0 >160 >160 >160 >160
C h em ic a l & S o lv e n t
R e s is ta n ce , 3 0 Min
M EK m ark slm rk slm rk m ark slm rk slm rk m ark slm rk m ark
X y le n e slm rk p ass slm rk slm rk slm rk slm rk slm rk slm rk slm rk
1 0 % A c e tic A cid pass p ass pass p ass pass pass p ass p ass pass
10% HCI pass p ass pass p ass pass p ass p ass p ass pass
10% NaO H p ass p ass pass pass pass p ass p ass p ass p ass

Figure 4. Diamine / Polyuretidione Polyurethane Cure by Biuret Fonvation

366 / Chandalia, Goldstein, Morgan, Wojcik

 With primary amines, this reaction occurs very rapidly POLYOL CURED COATINGS
at room temperature. In fact, the formulation pot life is so
short that these systems are best applied using dual When formulated with polyester or acrylic polyols and
component spray systems. Note that greater than common organometallic catalysts, such as dibutyltin
stoichiometric amounts of the diamines were required to dilaurate and dibutyltin mercaptide, or amine catalysts,
give the coatings with the best properties. This suggests such as l,4-diazabicyclo[2.2.2]octane (DABCO) or 1,5-
that the reduced mobility to the growing polymer chains diazabicyclo[4.3.0]non-5-ene (DBN), these HDI
inhibited the achievement o f optimum crosslink density in polyuretidione polyurethanes failed to yield acceptable
the film. coatings even when baked at high temperatures for
While these coatings performed well in the cross-hatch prolonged periods. We evaluated more exotic catalysts and
adhesion test on unprimed metal, we found that they had found a class of compounds that will catalyze the
poor inter-coat adhesion on primed and base-coated formation of coatings using standard polyester and acrylic
substrates if the undercoat was allowed to cure for more polyols with these HDI polyuretidione polyurethanes.
than one day. This may be explained by the fact that no These catalysts allowed us to prepare coatings with
free isocyanate groups are formed during the cure process excellent properties under the relatively mild cure
and are, therefore, not available to promote good adhesion. conditions of 80°C for 30 minutes. Example formulations
Further, the coatings are expensive because of their high using two polyuretidione polyurethanes, at several
isocyanate content compared to a conventional polyol uretidione to hydroxyl group ratios, and the properties of
based coating. the resulting films are shown in Table 3. In these systems,
the cross-linking reaction involved is the opening of
uretidione ring with the formation of an allophanate
linkage. (See Figure 5.)

Table 3. Polyuretidione Polyurethane Structure and Index

P o ly u re tid io n e P o ly u re th a n e
S o lu tio n # 88-1 88-1 88-1 88-1 44-1 44-1 44-1 44-1
Gram s 10.2 12.8 15.4 17.9 8.9 11.2 13.5 15.7
Uretidione:OH 0.8 1.0 1.2 1.4 0.8 1.0 1.2 1.4
G - C ure 109H 75, gm 10 10 10 10 10 10 10 10
Catalyst I, 50% in acetone, gm 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9
Application Method DrwDwn DrwDwn DrwDwn DrwDwn DrwDwn DrwDwn DrwDwn DrwDwn
Cure Tem perature, ° C 80 80 80 80 80 80 80 80
Cure Tim e, minutes 30 30 30 30 30 30 30 30

F IL M P R O P E R T IE S
Unpolished bonderite panels, 2 4 hours 2 5 ° C /5 0% RH
Thickness, mils 2.3 2.4 2.5 2.4 2.6 2.4 2.4 2.5
Gloss
20 ° 94 97 89 90 97 96 97 96
60 ° 106 109 102 106 110 109 110 109
P E N D U L U M H A R D N E S S , Swings 87 75 68 42 15 36 46 47
1/8 M andrel Bend pass pass pass pass pass pass pass pass
Cross Hatch Adhesion
% Rem oved 0 0 0 10 0 0 0 10
Impact, in/lb
Direct >160 > 160 > 160 >160 > 160 > 160 > 160 > 160
Reverse 60 > 160 > 160 >160 >160 > 160 > 1 60 > 160
Chem ical & Solvent R esistance
Spot Test, 30 Minutes
M EK slmrk slmrk slmrk slmrk slmrk slmrk slmrk slmrk
Xylene slmrk pass slmrk slmrk slmrk slmrk slmrk slmrk
10% A cetic Acid pass pass pass pass pass pass pass pass
10% HCI pass pass pass pass pass pass pass pass
10% N aO H pass pass pass pass pass pass pass pass
M EK , 100 Double Rubs slmar slmar slmar slmar slmar vslm ar vslm ar slmar

Chandalia, Goldstein, Morgan, Wojcik / 367

 Figure 5. Polyol / Polyuretidione Polyurethane Cure by Allophanate Formation

In this case, closer to stoichiometric amounts of polyol required is the addition of the catalyst to initiate the cure
gave the coatings with the best properties. This suggests process.
that the elevated temperature cure increased the mobility For the present, we have focused on the somewhat less
of the growing polymer allowing a more optimum active catalysts because they provide a very long pot life
crosslink density in the film at this lower ratio. (more than two days) and because we found that coatings
While these systems overcame the inter-coat adhesion prepared at lower temperature (60°C, even for longer
and cost issues o f the diamine-cured formulations, this times) did not perform as well. They were somewhat
class of catalysts was judged to be unsuitable for softer and did not have the same degree of chemical
commercial use. They are expensive and hard to handle. resistance compared to those cured at the higher
They are air sensitive and/or prone to cause severe temperatures. (See Table 4 for a comparison o f the effect
yellowing o f the coating. of varying catalyst level and cure conditions.) This again
However, a second class of catalysts has been found suggests that a certain amount of heat is required to keep
that avoids these problems. They have high activity and the growing polymer chains flexible enough to form a
lower potential for color formation, are low cost and very fully cross-linked matrix. As before, the cross-linking
compatible with formulation components. Again, we were reaction involved is the opening of uretidione ring with
able to prepare excellent coatings with an 80°C / 30 the formation of an allophanate linkage.
minute cure. O f course, depending on the catalyst loading To demonstrate that polyester polyols would also give
and the exact structure of the catalyst, even milder cure acceptable films when cured with these polyuretidione
conditions are possible. Since the isocyanate groups are polyurethanes, Table 5 shows the results from a
not available for reaction, polyol / polyuretidione comparison of a polyester system and an acrylic system at
polyurethane mixtures are indefinitely stable. Thus, they several uretidione to hydroxyl group ratios.
can be formulated as pseudo - IK systems where all that is

368 / Chandalia, Goldstein, Morgan, Wojcik

 Table 4. Catalyst Level and Cure Conditions
P o ly u re tid io n e P o ly u re th a n e
S o lu tio n # 88-1 88-1 88-1 88-1 88-1 88-1 88-1 88-1
G ram s 3 8 .4 3 8 .4 3 8 .4 3 8 .4 3 8 .4 3 8 .4 3 8 .4 3 8 .4
UretldionerO H 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0
G - C u re 1 0 9 H 7 5 , gm 30 30 30 30 30 30 30 30
E E P , gm 24 24 24 24 24 24 24 24
C atalyst II, 5 0 % in t - B utanol, gm 0 .6 0 .9 1.2 1.2 0.9 0 .9 1.2 1.2
Application M ethod Spray Spray Spray Spray Spray Spray S p ray S pray
C ure T em p eratu re, ° C 80 80 80 80 60 60 60 60
C ure T im e, m inutes 30 30 30 60 30 60 60 90

F IL M P R O P E R T IE S
Unpolished bonderite panels, 2 4 hours 2 5 ° C /5 0 % RH
Thickness, mils 2.1 2.6 2 .5 2.3 2.6 2 .3 2 .2 2.5
Gloss
20 ° 92 94 89 90
60° 105 110 101 109
P E N D U L U M H A R D N E S S , Swings 35 45 64 62 32 38 47 52
1/8 M andrel Bend pass pass pass pass
Cross Hatch A dhesion
% R em oved 0 0 0 0
Im pact, in/lb
Direct >160 >160 >160 >160
R everse >160 >160 >160 >160
C hem ical & Solvent R esistance
Spot T est, 3 0 M inutes
MEK slmrk slmrk slmrk slmrk
X ylen e pass pass slmrk slmrk
10% A cetic Acid pass pass pass pass
10% H C I pass pass pass pass
10% N a O H pass pass pass pass
M E K , 100 D ouble Rubs fail fail vslm ar vslm ar fail fail m ar slm ar

Because o f the low cure temperatures we were able to spray coated to a dry film thickness of 2.8 to 3.0 mils of
use, we decided to test the performance of these formu­ clearcoat. All panels were cured in an oven at 80°C for 30
lations on plastic substrates. Example formulations and the minutes. After curing, all panels were then conditioned at
adhesion performance of the resulting films are shown in 25°C and 50% relative humidity for at least 24 hours
Table 6. With the exception of PVC, the formulations before testing.
cured with ‘Catalyst II’ showed very good adhesion to all
of the plastics tested. The acrylic film on SMC appeared to Coatings on Plastic Substrates
be defect free, suggesting that the coating completely Film crosshatch adhesion was tested on sample panels of
sealed the surface. the following plastics:
PVC (Flexible PVC Vista 8001);
EXPERIMENTAL TPO (Himont Republic Inc, HMN ETA 3183 Rim
Flexible Exterior);
Coatings on Metal Substrates Urethane (Bayer, BayFlex 110-25 Glass filled 15%);
The formulations were applied to test panels by ABS (Dow, DowPulse 830 ABS Rigid Exterior);
drawdown or with a Binks Model 7 spray gun. For SMC (GNC 7113 IMC:EC 600 SCM Rigid Exterior);
appearance and mechanical property measurements and and, Polycarbonate (GE, Lexan EM 1210-754).
chemical resistance tests, 4”x l2 ” polished cold rolled steel All of the samples were pretreated by wiping with
panels were coated to a diy film thickness of about 2 to 3 isopropanol and then allowed to air dry. The SMC samples
mils. For QUV exposure tests, 4”x l2 ” U2SRW035K were then sanded with 3M Imperial Grade Microfine 2000
primed panels with an E86WE466L white basecoat were paper followed by wiping with isopropanol and then

Chandalia, Goldstein, Morgan, Wojcik / 369

 Table 5. Polyol Structure and Index
P o ly u r e tid io n e P o ly u r e th a n e
S o lu tio n # 88-1 88-1 88-1 88-1 88 -1 88 -1

G ra m s 3 0 .6 3 8 .4 4 6 .2 3 2 .8 4 1 .0 4 9 .2
U re tid io n e :O H 0 .8 1.0 1.2 0 .8 1 .0 1.2
R u co flex X R - 3 5 1 A , gm 15 15 15
R u co flex X R - 3 5 0 , g m 4.1 4.1 4.1

G - C u re 1 0 9 H 7 5 , g m 30 30 30

E E P , gm 24 24 24 22 22 22

C a ta ly s t II, 5 0 % in t - B u tan o l, gm 1.2 1.2 1.2 1.2 1.2 1.2

A p plicatio n M eth o d S p ra y S p ra y S p ra y S p ra y S p ra y S p ra y
C u re T e m p e ra tu re , ° C 80 80 80 80 80 80
C u re T im e , m in utes 30 30 30 30 30 30

F IL M P R O P E R T IE S
U n p o lish ed b on derite p an els, 2 4 hours 2 5 ° C /5 0 % R H
T h ic k n e s s , m ils 2.1 2 .2 1.9 2 .4 2 .6 2 .5

G loss
CM
o
o

94 97 89 90 97 96

60° 1 06 10 9 10 2 10 6 110 109

P E N D U L U M H A R D N E S S , S w in gs 78 69 65 52 51 40

1 /8 M a n d re l B end pass pass p as s p as s p ass p as s

C ro ss H a tch A d h e s io n
% Rem oved 0 0 0 10 0 0

Im p act, in/lb
D irect >160 >160 >160 >160 >160 >160
R e v e rs e 120 >160 >160 >160 >160 >160
C h e m ic a l & S o lv e n t R e s is ta n c e
S p o t T e s t, 3 0 M in u tes
MEK slm rk slm rk slm rk slm rk slm rk slm rk
X y le n e slm rk pass slm rk slm rk slm rk slm rk

1 0 % A c e tic A cid pass pass p as s p ass p as s p as s

10% HCI pass pass p as s p as s p as s p as s
10% N aO H p ass p ass p as s p as s p as s p as s
M E K , 1 0 0 D o u b le R u b s slm a r slm a r s lm a r s lm a r v s lm a r v s lm a r
Q U V - B , D ry, 1 5 0 0 H ours
2 0 ° G lo s s R eten tio n , % 95 94
6 0 ° G lo s s R eten tio n , % 98 97

In c re a s e in Y l 1.2 1.3

allowed to air dry. After wiping with isopropanol, the TPO CONCLUSIONS
samples were further treated by washing with liquid Tide,
dried for 30 minutes at 70°C, then spray coated with a We have developed polyuretidione polyurethane
50/50 blend of resin and CP-343 Adhesion promoter curatives that are suitable for use in solvent and waterborne
diluted to 5% solids in toluene and allowed to air dry. The coatings systems. They provide the advantage of having no
test formulations were applied to the panels with a Binks free isocyanate groups, thus reducing the potential of
Model 7 spray gun to a dry film thickness of 2 to 3 mils of worker exposure. Because they are ‘internally5blocked,
clearcoat and cured in an oven at 80°C for 30 minutes. All they avoid the problems associated with conventional
panels were then conditioned at 25°C and 50% relative blocking agent emissions, while giving coatings with the
humidity for at least 24 hours before testing. excellent properties typical of polyurethanes. We have also
identified a class of catalysts that can be used to prepare
coatings under the relatively mild cure conditions of 80°C

370 / Chandalia, Goldstein, Morgan, Wojcik

 Table 6. Adhesion to Plastic BIOGRAPHIES
Polyuretidione Polyurethane
Solution # 88-1 88-1 88-1 K iran B. Chandalia
dram s 353 38.4 TO5
UretidioneiOH 1.0 1.0 1.0
Kiran received his M.S. in
Rucoflex XR - 351A, gm 15
Rucoflex XR - 350, gm 4.1
Chemical Engineering from LSU,
G -Cure 109H75, gm 30 30 Baton Rouge and B.S. in Chemical
EEP, gm 24 24 22 Engineering from Bombay
Catalyst I, 50% in acetone, gm 2.7 University. He has held leadership
Catalyst II, 50% in t-Butanol, gm 1.2 1.2 positions in Research and
Application Method Spray Spray Spray Development, Technical Service,
Cure Temperature, ° C 80 80 80
Business Development, Global
Cure Time, minutes 30 30 30
Marketing and Business
Thickness, mils 2.1 2.1 2.6
Management for Polyurethanes,
Cross Hatch Adhesion, % Removed Coatings, Biocides and Plastics
PVC 100 100 100 Additives businesses. Currently he is the Global Business
TPO 0 0 0 Manager for Luxate Aliphatic Isocyanates at Lyondell
Urethane 0 0 0 Chemical Company. Kiran has twenty -five U.S. Patents
ABS 10 0 0 and numerous publications in scientific journals.
SMC 90 0 10
Polycarbonate 0 0
Stephen L. Goldstein

Steve Goldstein received his B.S. in

for 30 minutes when these curatives are formulated with Chemistry from the University of
standard polyol resins. The resulting films have excellent Pennsylvania and his Ph.D. in
appearance, good mechanical performance and good Organic Chemistry from New York
chemical resistance. In addition, they show some University. After completing
advantages for use in coating plastics. postdoctoral work at Queens
As stated above, this is an interim report on a work in College, NY, he joined Olin, in
progress. We are going to more fully evaluate the effectof 1974, as a Sr. Research Chemist
modification o f the polyuretidione polyurethane backbone working on Urethane catalysts.
and to more fully explore the impact of catalyst loading Currently, he is a Manager at
and structure versus cure time and temperature on coating Lyondell Chemical Company,
performance. We will be attempting to optimize film leading the Aliphatic Isocyanate Product, Process and
properties as well as increase the solids content of the Applications Research and Development group. He holds
formulations. Further, we must complete and extend the more than twenty U.S. patents and has numerous
environmental testing on the coatings and compare them to publications in scientific journals.
current commercial products

REFERENCES

1. U.S. Patent 4,463,154; “Polyurethane Powder Enamels

Free o f Cleavage Products As Well As Enamel
Coatings Produced From Such Powders”; 07/31/1984

2. U.S. Patent 5,861,193; “Low Viscosity Polyuretidione

Polyurethanes and Their Use as Curatives For Solvent
and Water Borne Coatings”; 01/19/1999

Chandalia, Goldstein, Morgan, Wojcik / 371

Michael J. Morgan Ronald T. Wojcik

Michael Morgan received his Ronald Wojcik received his B.A. in

undergraduate degree from Chemistry from Ithaca College,
LeMoyne College. He joined Olin M.S. in Polymer Science from the
in 1981 and worked on a variety of Polytechnic Institute of New York,
projects including solar salts, M.S. and Ph.D. from the Polymer
photoactive color toners, energetic Science & Engineering Department
binders and conductive polymers. In at the University of Massachusetts
1991, he joined to the Aliphatic at Amherst. He has worked as a
Isocyanates coatings applications principal scientist in the chemical
group where he worked on urethane industry for 32 years on related
coatings formulations for solvent projects including urethane coatings,
borne, moisture cure, water borne and radiation cured structural adhesives, sealants, high performance composite
systems. Currently, he is continuing coatings application materials, gas permeable membranes, organic corrosion
development in aliphatic isocyanates at Lyondell Chemical inhibitors, and conductive materials. Currently, Ron is the
Company. Michael has more than fifteen U.S. Patents and Technical Service Manager for Luxate Aliphatic
numerous publications in scientific journals. Isocyanates at Lyondell Chemical Company. He is the
holder of fifteen U.S. patents and has given more than 50
presentations on various topics covering polymer
technology.

372 / Chandalia, Goldstein, Morgan, Wojcik

The Evaluation of Metal and Tertiary Amine Catalyst in
CASE Application

HIROYUKI KOMETANI a n d YUTAKA TAMANO ROGER VAN MARIS

TOSOH Corporation TOSOH EUROPE B.V
Chemical Research Laboratory Crown Building-South,
4560 Kaisei-cho ShinNanyo-city Hullenbergweg 359
Yamaguchi 746-8501 1101 CP Amsterdam Z.O.
Japan The Netherlands

KENNETH M. GAY
TOSOH USA, Inc.
1100 Circle 75 Parkway Suite 600
Atlanta, Georgia 30339-3097

ABSTRACT essential and metal catalysts, which exhibit the gelling

reaction only, are not useful.
This paper describes a newly developed catalyst system In CASE application,the gellingreaction is very important
based on metal and tertiary amine by fundamental and the formation of bubbles and foaming should be avoided
investigation in CASE, especially sealant application. In and therefore catalysts, which promote die blowing reaction
CASE application, the most important point is to get the can not be used. That is why metal catalysts are used
special step reaction profile: “Long pot life and with sharp normally. Basically, the catalytic activity of light metal
increase in viscosity”. catalysts is low and consequently the requested gelling
In the CASE application, amine catalysts are very seldom activity is not obtained. Heavy metal catalysts based on lead,
used because they exhibit a short pot life with a mild viscosity stannous or mercury catalysts do show a higher activity and
rise. In these applications, mercury, lead or stannous catalysts are therefore preferred.
are used mostly because they are strong gelling catalysts and There is another reason why such heavy metal catalysts are
exhibit long pot life with a sharp viscosity rise profile. used. In CASE application, a long pot life is required with a
However, some o f these heavy metal catalysts show high sharp viscosity rise profile. In other words, after mixing all
toxicity. Tin catalysts exhibit a lower toxicity compared to the ingredients, the mix should maintain a low viscosity for a
the mercury or lead catalysts, however, they may contain long time to allow a good handling and processing of the mix
endocrine disrupters such as tributyltin compound [1]. before it will cure thoroughly. Pot lives vary from 5 minutes
Nowadays, people even become more sensitive to the toxicity to lhr depended on the application. Heavy metal catalysts
of chemical substances as well as the effects to the especially mercury catalysts show long pot life with sharp
environment. O f course CASE producers or suppliers want to viscosity rise profile.
avoid the use of these toxic substances and look for Needless to say, mercury catalysts show very high toxicity
alternative, however, no catalysts exhibited the requested and will not be used on CASE producer’s own judgment in
viscosity rise profile. Japan although those catalysts are used in the European and
Tosoh Corporation has examined the effectiveness of American market. Lead catalysts also have a high toxicity
various metals and tertiary amine catalysts in several and has been phased out in some applications. Stannous
problems associated to the sealant application. We will catalysts show a lower toxicity, however, some reports
propose a novel catalyst system, which is based on an amine pointed out that they may contain tributyltin compounds
catalyst. This system is non toxic while maintaining a long which again show high toxicity. Tributyltin can’t be remove !
pot life with a sharp viscosity rise, as well as, excellent from stannous catalysts such as DBTDL
elastomer properties. These novel catalyst systems could also (Dibutyltindilaurate). In Europe, the USA and Japan, but also
be applied in other CASE applications. in European countries there is a tendency to avoid the use of
such hazardous catalysts. CASE producers or suppliers are
INTRODUCTION looking for a non-toxic alternative.
TOSOH Corporation evaluated a variety of amine catalysts
In polyurethane foam applications, amine catalysts are and metal catalysts from the point aiming at “long pot life
mainly used and metal catalysts are seldom. In polyurethane with sharp viscosity rise” and low toxicity. In addition, some
foam, besides the gelling reaction, the blowing reaction is physical properties of the elastomers were evaluated.

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 373

EXPERIMENTAL
Table 2. Metal catalysts
Catalysts and Formulations abbreviation
Catalyst

The catalysts evaluated and their abbreviations are listed in Mercury Neodecanoate Ha
Table land 2. A variety of metal catalysts are evaluated and Lead Octoate Pb
amine catalysts, which are classified as gelling side catalysts, Dibutyltindilaurate Sn
are evaluated. In the foam application, delayed action
catalysts, which are blocked by organic acid such as formic Iron 2-Ethylhexanoate Fe
acid, are used to obtain long cream time. TF is an acid formic Cobalt Naphthenate Co
acid blocked catalyst of TEDA (Triethylenediamine) and it Potassium 2-Ethylhexanoate K
was expected to obtain longer pot life compare to TEDA.
Manganese 2-Ethylhexanoate Mn
DBU (Diazabicycloundecene) exhibits strong gelling activity
however it will decompose easily in the presence of water, so Nickel 2-Ethylhexanoate Ni
basically it is used as a complex blocked by phenol or organic Titanium 2-Ethylhexanoate Ti
acids.
The formulations used are shown in Tables 3 and 4. Both Zirconium Naphthenate Oxide Zr
formulations use polyetherpolyol and MDI prepolymer.
Zeolite powder was used as molecular sieve to dry the polyol. Vanadium Naphthenate V
The isocyanate index was adjusted to 105 for both systems.
The effect o f water is discussed later.
Table 3. Formulation- A

Table 1. Amine catalysts Component Parts by weight

Catalyst abbreviation Polyol(1> 24

Triethylendiamine 33% in DPG L33 Polyol(2) 72

Formic acid blocked catalysts of Moleculair Sieve (3) 4

TF
Triethylendiamine Varied
Catalyst
1,2-Dimethylimidazole 70% in EG DM70
Isocyanate Prepolymer(4) Index = 105
1,8-Diazabicylo(5,4,0)undecene-7 DBU 1) Polyether polyol: OHV = 376mgKOH/g
Phenol blocked catalyst of DBU DB-Ph 2) Polyether polyol: OHV = 56mgKOH/g
3) Zeolum A3 powder ex Tosoh
Formic acid blocked catalyst of
DB-F 4) MDI base prepolymer: 26.0% NCO
DBU
2-Ethylhexanoic acid blocked Table 4. Formulation- B
DB-E
catalyst of DBU
Component Parts by weight
Acetic acid blocked catalyst of
DB-Ac Polyol(1) 75
DBU
Phtallic acid blocked catalyst of Polyol<2) 25
DB-Pht
DBU Chain-extender 5

Oxalic acid blocked catalyst of Anti Foaming Agent 0.1

DB-Ox
DBU Moleculair Sieve(3) 5

Special acid blocked catalyst of Catalyst Varied

DB41
DBU 50% in Dietyleneglycol Isocyanate Prepolymer(4) Index = 105
1) Polyether polyol: OHV = 56mgKOH/g
Special acid blocked catalyst of 2) Polyether polyol: OHV = 34mgKOH/g
DB42
DBU 50% in Dietyleneglycol 3) Zeolum A3 powder ex Tosoh
4) MDI base prepolymer: 23.1% NCO

374 / Kometani, Tamano, Van Maris, Gay

Reaction Profiles and Properties Evaluation of catalysts in formulation-A
Mixing Condition
A 20 g premix together with the catalyst were put into a In this formulation, the pot life requirement is more than 15
200ml-disposalable PE cup and were stirred well. The minutes. O f course, a sharp viscosity rise profile is required
mixture was degassed in a dessicator under vacuum at 5mbar. after 15 minutes. The required hardness is around 70 shore-
The isocyanate prepolymer was added and stirred well for A value.
another 30 seconds and degassed again. The system is left as
it is in case of evaluating the viscosity rise profile and METAL CATALYSTS
temperature rise profile. In case of evaluating the hardness, a Each metal catalyst is evaluated at different usage levels.
fixed weight of the reaction mixture was poured into another Typical viscosity rise profiles are shown in Figure 1.
PE cup to obtain samples with the same thickness. The As is expected, Hg catalysts exhibits long pot life with sharp
experiments were carried out with conditioned raw materials viscosity rise. Lead catalyst and stannous catalyst exhibit
at 20 degrees Centigrade. almost the same viscosity rise profile; however, it is milder
compared to Hg catalyst and the pot life o f both catalysts
Viscosity Profile becomes shorter. The shore-A hardness o f the Hg containing
Apparatus : Vibrational Viscometer VM-1G -MH (Yamco elastomer is 70 and the hardness of the Pb and Sn based
Corporation) elastomer is lower than Hg.
Thermocouple detector
The viscosity and the temperature values of the reaction Figure 1. Viscosity rise profile of metal catalysts
mixture were collected every second and the data were
collected by the PC automatically. To prevent the destruction
of the probe, the measurement was stopped whenever the
viscosity rose up to about 5000mPa.s.

Hardness
The hardness (after 1 hour and 24hours) were measured by
shore-A durometer. Basically, the elastomer sample
thickness was set to 6mm. However, to discuss the effect of
the reaction exotherm, the hardness of different thickness was
evaluated. The room where the preparation is made and the
testing is carried out, is 20 °C and 50% relative humidity. The
Figure 2 shows the viscosity data of Fe catalyst. In case of
samples are cured and post-cured at different temperatures
increased usage levels, it sharpens the viscosity rise profile,
depending on the formulation and catalysts. The hardness is
however the pot life is around 400 seconds not meeting the
measured at 20 °C after cooling the sample. The presented
target. To increase the pot life, the usage level was reduced to
hardness value is an average of four measured values.
one-third, however the viscosity rise becomes mild.
Pot Life, R -ratel, R-rate2, R-rate3, R-rate4, Hardness 1,
Hardeness2> H-Ratio are defined as follows. Figure 2. Viscosity rise profile of Fe-Cat

Pot Life : Time when the viscosity of the system rose up to

800mPa.s.
R -ratel : 1000/Tim e when the viscosity of the system
becomes 2000mPa.s from 1000mPa.s
R-rate2 : 2000 / Time when the viscosity of the system
becomes 3000mPa.s from 1000mPa.s
R-rate3 : 3000 / Time when the viscosity of the system
becomes 3000mPa.s from 1000mPa.s
R-rate4 : 2000 / Time when the viscosity of the system
becomes 4000mPa.s from 2000mPa.s
Hardness 1 : Hardness after 1hr (shore-A) As is shown in Figure 3 and 4, Cobalt, Potassium,
Hardness2 : Hardness after 24hr (shore-A) Manganese, Nickel, Zirconium and Vanadium catalyst do not
H-Ratio : Hardness 1 / Hardness2 (%) exhibit a long pot life with a sharp viscosity rise. Besides the
viscosity rise problem, most of the metal salts color the
reaction mixture such as green, blue, and purple. Therefore,
they can’t be used in CASE application.
RESULTS AND DISCUSSION
As is shown in Figure 5, only Titanium catalyst exhibits

Kometani, Tamano, Van Maris, Gay / 375

long pot life with sharp viscosity rise. The obtained hardness As is shown in Figure 7, when the usage level of DBU was
of Titanium catalyst based elastomer is acceptable. Several reduced to half to obtain a longer pot life, the viscosity of
kinds o f Titanium catalysts are investigated and they show system does not rise, i.c. no cure was obtained. It means that
equal properties. Titanium catalysts exhibit very low toxicity, DBU requires a minimum dosage level. However, it doesn’t
so it can be used in this application. However, the salts show exhibit a long pot life. Besides that, DBU is not stable in the
poor hydrolytic stability. presence of water, so basically it is used in a complex state:
blocked by phenol or organic acids. The pot life depends on
Figure 3. Viscosity Rise profile of metal catalysts
the blocking agent.

Table. 5 Usage level and properties of metal catalysts

Cat pbw Pot Life Viscosity Hardness 2 Color
(sec) Rise (shoreA)
Hg 0.500 710 Sharp 70 None
Pb 0.001 570 Mild 52 None
Sn 0.006 570 Mild 49 None
Fe 0.005 710 Mild 27 Black
Co 0.010 480 Mild 30 Purple
Zr 6.17 ______ _______ ____— None
Figure 4. Viscosity rise profile of metal catalysts V 3.607 830 Mild 15 None
Ni 0.817 780 Mild 23 Green
K 0.107 860 Mild 18 None
Mn 0.015 908 Mild 31 Purple
Ti 0.005 900 Sharp 68 None

Figure 6. Viscosity rise profile of amine catalysts

Figure 5. Viscosity rise profile of Ti(C8H170)4

Figure 7. Viscosity rise profile of DBU

AMINE CATALYST

Figure 6 shows the viscosity rise profile of typical amine

catalysts. Different from the metal catalysts, most of them
exhibit a short pot life and mild viscosity rise. TF, which is a
delayed action catalyst of TEDA, exhibit little bit longer pot
life but the viscosity rise is not sharp. Other amine catalysts
are also evaluated, however there is no catalyst which show
sharp viscosity rise. Only DBU catalyst shows extremely
sharp viscosity rise although pot life is around 600 seconds.

376 / Kometani, Tamano, Van Maris, Gay

THE EVALUATION OF BLOCKING AGENT 900 seconds) with a sharp viscosity rise. Toyocat DB41 and
DB42 use a special blocking agent resulting in moderate
Several blocking agents were evaluated to adjust the pot life. blocking. The system temperature will rise gradually due to
Typical viscosity rise profiles are shown in Figure 8-12. As is the reaction exotherm and when the temperature reaches
shown in Figure 8 and Table 6, phenol shows good blocking 50 °C, the deblocking takes place and the DBU molecule
and the pot life can be extended to 700 seconds. Formic acid recovers its catalytic activity and the urethane reaction will
and 2-Ethylhexanoic acid can also extend the pot life, start instantaneously. Toyocat DB41 and DB42 are non-toxic
however, they can’t extend the pot life to 900 seconds. materials and are good candidates as alternative for the
mercury catalyst.
Figure 8. Viscosity rise profile of DBU -Ph
Figure 11. Viscosity rise profile of DB41

Figure 9. Viscosity rise profile of amine catalysts

Figure 12. Viscosity rise profile of DB42

In case of using phtallic acid or oxalic acid blocked catalyst,

the viscosity of the system will not rise, i.c. showing a lack of Table. 6 Usage level and properties of amine catalysts
cure. It is supposed that such acid blocks the DBU molecule
Cat pbw Pot Viscosity Hardness 2 Color
so strong that the DBU can not exhibit its catalytic activity
anymore. Life Rise (shoreA)
(sec)
Figure 10. Viscosity rise profile of DB -Pht L33 0.500 220 mild 2 None
TF 0.558 240 mild 1 None
DM70 0.900 230 mild 2 None
DBU 1.026 530 sharp 65 None
DB-Ph 1.355 700 sharp 69 None
DB-F 1.165 660 sharp 66 None
DB-E 1.236 720 sharp 60 None
DB-Ac 1.815 730 sharp 61 None
DB-Pht 4.407 __ ___ None
DB-Ox 5.871 __ ___ ____ None

Figure 11 and 12 show the viscosity rise of Toyocat DB41 DB42 1.499 900 sharp 70 None
and DB42 respectively. They exhibit long pot life (more than DB41 1.235 900 sharp 70 None

Kometani, Tamano, Van Maris, Gay / 377

 Evaluation of catalysts in formulation-B rise profile. The elastomer sample surface, in case of
stannous catalysts, is sticky for a long time and the hardness
For this formulation, the pot life requirement is more than 5 can’t be measured. The lead catalyst also results in a sticky
minutes. Besides a sharp viscosity rise profile, fast curing is a elastomer sample, however the hardness after 24 hours is
very important factor. The completion of the cure is good. The sticky elastomer sample surface represents
expressed as the H-ratio, see the experimental section. The incomplete cure, which will continue for a long time and
requirement is the system should be cured for 80% after 1 therefore lead and stannous catalysts are difficult to use.
hour already. Teda L33 and Toyocat DM70 also exhibit a mild viscosity
rise and its sample surface is sticky for a long time. Both
VISCOSITY RISE AND HARDNESS catalysts can’t be used in this formulation.
Toyocat DB41, DB42 and the DB-Ph exhibit a sharp
Figure 13 shows the viscosity rise profile of each catalyst. viscosity rise profile. The elastomers also show good
Different from formulation-A, the mercury catalyst does not hardness values and cure has almost completed after 1 hour.
exhibit a sharp viscosity rise. Still, already 86% of the cure is It could be said that they are the good candidates for the
obtained after 1 hour. Lead and stannous catalysts have a alternative of mercury catalysts.
high catalytic activity; however, they exhibit a mild viscosity

Figure. 13. Viscosity Rise Profile of each catalysts

Table 7. Reaction data and properties

Catalyst pbw Pot Life VisRisel VisRise2 VisRise3 VisRise4 H1 H2 H-ratio
Hg 0.518 384 5.8 9.0 12.0 25.3 48 56 85.7
Pb 9.1*10 -4 256 9.4 13.1 16.0 25.0 ND 50
Sn 8.1*10 -4 354 6.21 8.1 9.4 12.6 ND ND
DB-Ph 0.1673 318 142.9 200.0 272.7 500.0 52 54 96.3
DB41 0.4407 309 142.9 200.0 272.7 500.0 51 53 96.2
DB42 0.3842 402 66.7 100.0 136.4 285.7 52 53 98.1
L33 0.078 239 6.5 8.9 10.5 15.2 ND ND
DM70 0.079 358 5.9 7.1 9.1 12.4 ND ND
Pot Life : Time when the viscosity of the system becomes 1000mPa.s
VisRiset : 1000 / (Time when the viscosity of the system becomes 2000mPa.s from 1000mPa.s)
VisRise2 : 1000 / (Time when the viscosity of the system becomes 3000mPa.s from 1000mPa.s)
VisRise3 : 1000 / (Time when the viscosity of the system becomes 4000mPa.s from 1000mPa.s)
VisRise4 : 1000 / (Time when the viscosity of the system becomes 4000mPa.s from 2000mPa.s)
H1 : Hardness after 1hr (shore-A)
H2 : Hardness after 24hr (shore-A)
H -ratio: H1/H2 (%)

378 / Kometani, Tamano, Van Maris, Gay

THE EFFECT OF WATER CONTENT thin sample, and so the use of Toyocat D B 41,42 and DB -Ph
becomes more critical in this formulation.
The water content of the system is controlled by the raw In case of the mercury catalyst, the surface is also still sticky
material supplier, however, there is a possibility that water after 1 hour for both 4 mm and 2 mm thickness, however,
content will increase when the elastomer producer uses the good hardness values are obtained even though the value
system. The effect of water content to the catalyst is itself is small. The lead catalyst has a same tendency as the
investigated. mercury catalyst.
As is shown in Table 8, the mercury catalyst will not cause If the sample is heated to some degree, different results will
the blowing reaction even at high water content. Stannous be obtained for thin thickness sample. More investigations
and Lead catalysts show the blowing reaction at high water will be done in near future.
content and the water content should be controlled under
0.35wt%. In case of DB41, DB42 and DB-Ph, the blowing CONCLUSIONS
reaction will occur more easily and therefore the water
content should be controlled, which is normally done. Metal and amine catalysts were evaluated by using two
sealant formulations:
1. In the evaluation of using formulation-A, mercury catalysts
Table. 8 The effect of water content in the system exhibit ‘long pot life with sharp viscosity rise’. Lead and
Water Blowing Reaction stannous catalysts have high catalytic activity and exhibit
content long pot life however the viscosity rise becomes mild.
Many metal catalysts like Cu, Ni, Co don’t show a sharp
(wt%) DB41.DB42, DB-Ph Hg Sn, Pb
viscosity rise profile, and only Titanium catalysts exhibit
0.65 Yes Slightly Yes sharp viscosity rise with long pot life.
0.35 Slightly No No 2. In case of amine catalysts, only DBU catalyst exhibits a
good performance, but DBU itself doesn’t result in a long
0.05 No No No
pot life. TOSOH Corporation has investigated several
kinds o f blocking agents and found that Toyocat DB41 and
DB42 which are specially blocked DBU catalysts, show a
THE EFFECT OF SAMPLE THICKNESS
long pot life with a sharp viscosity rise profile. These
It is thought that the reason why DB41, DB42 and DB-Ph
materials are non-toxic and can be used as an alternative to
exhibit a long pot life with a sharp viscosity rise profile is due
mercury catalyst.
to reaction exotherm of the system. If the thickness of the
3. In the evaluation of using formulation-B, Mercury catalyst
elastomer is small, not much heat built up is expected due to
does not exhibit a sharp viscosity rise. However it exhibits
heat loss. The effect of sample thickness to the hardness, i.c.
good curing. Toyocat DB41 and DB42 exhibited good
cure, was investigated.
curing and showed a sharp viscosity rise. However, at
As is shown in Table 9, Toyocat DB41, DB42 and DB-Ph
small thickness the use becomes more critical.
exhibit good hardness values in case of a 6 mm thick sample.
However, in case of 4 mm thick sample, the elastomer sample
REFERENCES
surface is still very sticky after 1 hour and the hardness after
24 hours resulted in low values. In case of a 2 mm thick
[1] Report o f the Federal Office for the Environment and the
sample, the surface is very sticky for a long time and the
Federal Institute for Health-related Consumer Protection
hardness can’t be measured.
and Veterinary Medicine , March 14, 2000
Table 9. The hardness of elastomer at different thickness
Catalyst Sample thickness
BIOGRAPHIES
6mm 4mm 2mm HIROYUKI KOMETANI
H1 H2 H1 H2 H1 H2 Hiroyuki Kometani joined TOSOH Corporation in 1990 after
Hg 48 56 ND 48 ND 40 receiving a B.S. and M.S. degree in Physical Chemistry from
Pb ND 50 ND 42 ND 35 Tsukuba University in 1988 and 1990 respectively. After
Sn ND ND ND ND ND ND development work in analytical chemistry, he has been
DB-Ph 52 54 ND 32 ND ND engaged in development of amine catalysts and other related
DB41 51 53 ND 33 ND ND materials for polyurethanes.
DB42 52 53 ND 31 ND ND
L33 ND ND ND ND ND ND YUTAKA TAMANO
DM70 ND ND ND ND ND ND Yutaka Tamano joined TOSOH Corporation in 1970. After
development work in organic chemical products, he has been
As is expected, not much heat built up occurs in case of a engaged in development of amine catalysts and other related

Kometani, Tamano, Van Maris, Gay / 379

materials for polyurethane. He is currently the group leader
of development and technical service of amine catalysts.

Roger van M aris

Roger van Maris obtained a MSc degree in Polymer
Technology from the Eindhoven University of Technology in
1987. In 1990 he completed a two year research course in
Polymer Technology at the same university. After working in
R&D and technical support in flame retardancy of mainly
flexible polyurethane foams, he joined TOSOH Europe B.V.
in 1998. He is the technical manager for polyurethane
catalysts in Europe but also in the near and Middle East &
African countries.

Kenneth M. Gay
Kenneth Gay joined TOSOH USA, Inc. in January 2000 as a
technical sales manager. He graduated from the University of
Southern Mississippi with a degree in Polymer Chemistry
and Engineering. After graduation, he worked with
Schering-Plough Health-Care Products as a new product
development chemist. Most recently, he worked with Isotec
International, a polyurethane systems developer, having
served as R&D director and sales manager. He will support
the customers not only in Northern USA but also in the
Southern American countries.

380 / Kometani, Tamano, Van Maris, Gay

RimSpray™: A Unique Method for Applying Polyurethane
Foam to Molds and Substrates

MICHAEL R BARTENSTEIN
Gusmer-Admiral, Inc.
305 West North Street
Akron, Ohio 44303

Due to the successful result of this

ABSTRACT
endeavor, several other applications where
Gusmer-Admiral, Inc. has spraying could be more efficient than
developed a unique spray method of pouring were discovered through sales and
applying polyurethane foams and coating to marketing efforts. Development of other
mold or substrate surfaces. The concept sizes soon followed based on these
combines technology from time-proven applications. As development progressed,
Gusmer Corporation spray and Gusmer- certain obstacles had to be overcome.
Admiral RIM processes into a mechanically PRODUCT DEFINITION
self-cleaning, recirculating RimSpray™
Characteristics: The mixheads are
system. This method allows covering large
mechanically self-cleaning, hydraulically-
areas of mold or substrate surfaces by
operated, double-acting, recirculating
dispersing the media into wide patterns
mixheads developed to spray polyurethane
directly related to tip orifice configuration
compounds by dispersing the mixed material
and height off the work surface. This
through a spray tip (nozzle) configured to
method has applications throughout the
create a fan or round pattern on a mold or
polyurethane industry from mold coating
substrate.
and filling to reinforced substrate
composition. (Figure 1) Mechanical Self-cleaning: To succeed in
expunging all the mixed material out the tip
of the mixhead upon closing, the male
DEVELOPMENT HISTORY spherical radius on the end of the mix piston
must fit exactly into the mating concave ball
Gusmer-Admiral was approached socket in the spray tip. To facilitate line-to-
by a large, first tier automotive supplier to line contact and not damage parts through
investigate the possibility of dispersing impact, a semi-resilient washer is
polyurethane foam in an open mold NVH compressed between the tip clamping
application. Drawing from forty years surface and the tip-retaining nut (Figure 3).
experience with spray-foam patterns at Basically, the rod and valve block
Gusmer Corporation, a prototype was dimensions are tooled to the same exacting
developed to fit the client’s specific needs of tolerances as Gusmer-Admiral standard
delivering around 150 grams per second mixheads. The thermoplastic bearing
with an approximate width of twelve inches. material embedded in the grooves to
The mixhead was required to fit on an separating the A and B chemicals and
existing robot and adapt to the daylight supporting the rod is a GAI standard and
allowed by the press and mold (Figure 2). also characteristic of the spray head design.

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 381

 Over Spray: Due to the dispersal of
Recirculation: The chemicals are routed uncured fluid from the tip, there is a
through grooves machined in the sides of the tendency to create over-spray at the
non-rotating mix piston allowing the extremities of the fan pattern. Selecting the
chemicals to recirculate back to their source proper tip based on the output being used
continuously for precise temperature can minimize this phenomenon. There will
control. When retracted (head open) the be a requirement for masking areas local to
piston covers and seals the return port the work, however, due to some over-spray
causing the two chemicals mix in the open and the ricochet of particles at the edge of
chamber and exit out the spray tip mounted the work.
on the end of the head. Both ingredients
flow through adjustable orifices used to SPRAY HEAD SPECIFICATIONS &
create pressures for impingement mixing. CAPABILITIES:
This is the case whether in recirculation or
dispensing. 6mm Spray head:
Mix piston diameter: 6mm (.236”)
Spray Tips (Nozzles): The spray tip profile Capacity: 302 grams/sec. (40
is based on the desired output of the lbs./min.) @1:1 ratio
mixhead, the resultant orifices required, and Maximum Viscosity: 1500 cps
the width of pattern desired. The criterion Pattern Width, max: 762mm (30”)
for tip development includes minimizing the Maximum pressure: 206 Bar (3000
cross-sectional tip angle to allow directing psi.)
the majority of material to the extremities of
the fan slot (Figure 4). 13mm Spray head:
After deciding the orifices required Mix piston diameter: 13mm (.511”)
to develop pressures needed for mixing, Capacity: 1664 grams/sec. (220
normally 1800-2200 psi, the area of the slot lbs./min.) @1:1 ratio
is determined. This is usually slightly less Maximum Viscosity: 1500 cps
than the orifice areas combined allowing for Pattern Width, max: 1524mm (60”)
sufficient backpressure to disperse the Maximum pressure: 206 Bar (3000
mixed material. Realizing the difficulty of psi.)
pinpointing all the tip combinations to be
used, certain sizes were determined using a All parameters are dependent on tip
15-degree tip angle cutter and various cutter configuration.
depths, which developed slot areas where
the cutter breaks through (Figure 5). The tip APPLICATIONS:
is then hardened and the concave ball
surface polished to fit the rod end. The following are actual and
possible applications for sprayed
Tip Cleaner (Figure 6): To insure polyurethane processing:
no distortion appears in the pattern, it is
imperative that the spray tip slot be kept Automotive: All NVH (Noise, Vibration,
clean. For this reason, a tip cleaner is and Heat) components such as carpeting,
available providing a mechanism that firewalls, trunk surfaces, door panels, and
employs two bronze or nylon wire wheels headliners (Figure 7) are excellent
spinning in opposite directions. The candidates for rapid mold or substrate
mixhead tip is moved parallel to the slot coating and filling.
over both wheels by manipulator or other
means providing foam cleaning in two Furniture: Moveable office panels, carpet
directions thus avoiding build-up in one slot underlay, seating, and table surfaces.
comer or the other.

382 / Bartenstein
Building Products: Modular roofing, spray coating applications employing
garage door laminar panel manufacturing, miniature components to mimic the RIM
moldings, modular housing and mobile devices used today.
homes, and slab insulation (Figure 8). These unique products and
processes will offer the industry new
Leisure Items: Pool float items, travel applications previously not accessible by
trailer or RV manufacture, and toys. stream pouring alone. Large surface coating
as well as small mold contours can be
Tanks and Pipes: Thermal insulation, covered with varying thickness of
barriers, external and internal coating, and polyurethane with standard RIM metering
condensation control. systems and the new Gusmer-Admiral spray
heads.
Aerospace and transportation: Fuel
system insulation, fuselage structure and
insulation, and refrigeration trailer
insulation.
BIOGRAPHY
SUMMARY:
Michael F. Bartenstein
In other unique applications,
Gusmer-Admiral, Japan has run tests in their Michael F.
laboratory coating firewalls and other Bartenstein has
components without upper mold halves been involved in
thereby gaining full coverage purely through the production and
spray-up. engineering of
The process allows more complete polyurethane
“wet-out” of porous substrates such as spray foam and
fiberglass mats or carpet swatches than RIM equipment
conventional pouring methods. The for 21 years at
recirculation feature allows improved Gusmer
temperature control over deadheaded, heated Corporation in
hose spray systems. No cold shots after Lakewood, New Jersey prior to becoming
dormant periods are evident as experienced president of Gusmer-Admiral, Akron, Ohio
with non-recirculating systems. in January 1998 where he also augments
Due to the ability to cover a wide R&D engineering and sales efforts. Mike
area rapidly, use of faster reacting chemical has been responsible for and assisted in the
systems are possible that allow de-molding development of improved CDC pumping
sooner than conventional RIM systems. systems, high-output plural component
spraying systems, innovative RIM metering
systems, associative dry-side components,
CONCLUSIONS: and complete turnkey projects during his
career. Mike is currently Director of Global
Developments at Gusmer-Admiral Engineering Services, Gusmer Machinery
are underway to use recirculating systems in Group, parent of Gusmer-Admiral, Inc.
the realm of super low-output elastomer

Bartenstein / 383
 Figure 1: 6mm Spray Head

Figure 2: 6mm spray head prototype

384 / Bartenstein
Figure 3: Tip Washer

Figure 4: Tip Angle

Bartenstein / 3 8 5
 Figure 5: Slot Area

Figure 6: Tip cleaner

386 / Bartenstein
 Figure 7: Headliner Demonstration

Figure 8: 48” wide fiberglass panel demonstration

Bartenstein / 387
Investigation of Polyurethane Adhesives for Use in the Apache
Aviator NBC Protective Mask

GERALD R. YOUNG, STEVEN CIBOROWSKI, DAVID MAY, LARRY OSWALD a n d STEVEN GROSS
United States Army Soldier Chemical-Biological Command
Edgewood Chemical-Biological Center
Berger Laboratory
Aberdeen Proving Ground, Maryland 21010-5424

A b s tr a c t

Polyurethane adhesives are currently used in Apache

aviator masks to secure polycarbonate lenses to a
bromobutyl-natural rubber face piece. This study examined
the rupture behavior of polycarbonate(pc)components
bonded to butyl-rubber with polyurethane adhesives.
Composites were fabricated and destructively tested to abraded with sand paper and treated with Lord 7701
reveal differences in failure mode and rupture strength as a adhesion promoter.
function of surface preparation, curing regime, and The rubber-adhesive-polycarbonate rupture strength
polyisocyanate-polyol formulations. results, loads reported in pounds are listed below(average
Alternative polyurethane adhesives packaged in pre ­ o f all specimens with error in measurement reported as
measured assemblies offer greater ease of application and 95% confidence limit).Loads were applied perpendicular
were investigated to determine if the replacement adhesive to the bond surface plane (1 inch diameter right circular
could be used in lieu of the original adhesive employed cylindrical surface) in an Instron Load cell. Grips o f the
during the manufacturing process. Data on the rupture load cell were moving apart at a constant rate o f 0.15
strength o f polycarbonate-adhesive-rubber joints was inches per minute. Load versus displacement results were
collected to compare LORD 7500 and Lord 7540(pre- recorded for seventy-five specimens and correlated with
packaged) adhesive. Rubber components of die prototypes failure mode. Distinct failure modes for heat-cured versus
were abraded with 80 grit ,100 grit, or no surface abrasion ambient-cured composites are evident in the
and bonded to polycarbonate with adhesive. Specimens specimens(photographs are included in the body o f the
were cured at temperatures o f 63 Celsius or ambient report.)
temperature(23 C). All polycarbonate surfaces were Load measurements indicate both adhesives are

Load to Induce Rupture

Adhesive Surface Curina Separation Mode Load at Failure

7500 100 grit 63 C rubber tearing 261+/-3tlbs

7540 100 grit 63 C adhesive tear from pc 239+/-41 bs
7500 100 grit 23 C adhesive peeled from rubber 195+/-21lbs
7540 80 grit 23 C adhesive peeled from rubber 177+1-36 lbs
7540 100 grit 23 C adhesive peeled from rubber 170+/-19 lbs
7500 80 grit 23 C adhesive peeled from rubber 149+/-15 lbs

Key Words: polyurethane, rupture strength, adhesives,

polycarbonate-polyurethane-bromobuytl rubber composite.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 389

adequate for securing optical lenses within the aviator (HO[CH2]6OH); alcohol-terminated polyethers such as
mask. To improve polyurethane adhesive bonding in the polyethylene oxide and polypropylene oxide (see Aliphatic
aviator mask application, curing at 63 degrees Celsius for polyethers); and flexible, alcohol-terminated polyesters
18 to 24 hours is recommended. such as poly-1,4-butylene adipate:

INTRODUCTION

Adhesives have a long history of use in human artifacts.

Studies by the Weizmann Institute of Science reveal that in
the stone age, humans were putting organic compounds to
use to join dissimilar materials together. As far as 8,000-
The alcohol-terminated polyethers and polyesters are
years-ago Neolithic cavemen who lived in the area of the
known as polyols. Isocyanates commonly used to prepare
Dead Sea were producing a form of collagen glue, as an
polyurethanes are toluene diisocyanate (TDI), methylene-
adhesive to hold together tools, utensils, and ceremonial
4,4 -diphenyl diisocyanate (MDI), and a polymeric
artifacts that included decorated skulls.1Natural glues and
isocyanate (PMDI). These isocyanates have the following
resins derived from animal and plant sources were used
extensively from antiquity up to the early 20th century. It structures:
is known that the ancient egyptians used collagen in its
gelatinous form as the basis for the "carpenter's glue" to
hold together fumiture.With the advent of the industrial
revolution and the establishment o f a chemical process
industry, adhesives began to be derived as adjuncts to a
wide range o f compounds derived from petrochemical
sources.
In the Apache Aviator Mask, Polyurethane is used as a
material to bond polycarbonate lenses to a face piece that is
composed o f a blend of natural and bromobutyl rubbers.
There is probably no more versatile material for so many During the late 1930s Otto Bayer, manager o f the IG
high-performance applications than polyurethane. With a Farben laboratories in Leverkusen, Ger., prepared many
molecular structure composed of four of the most common polyurethanes by condensation reaction of dihydric
elements - carbon, hydrogen, oxygen, and nitrogen - alcohols such as 1,4-butanediol with difunctional
polyurethane has a combination o f properties which are not diisocyanates. A major breakthrough in the commercial
available in any other thermoplastic material including application of polyurethane did not occur until 1941, when
excellent resistance to ozone, weak bases, fuels and a trace of moisture reacted with isocyanate to produce
oils.Urethane was first introduced in 1849 by Germany's carbon dioxide. The production of this gas resulted in
Wurtz and Hoffman, when they presented the chemical many small empty areas, or cells, in the product (which
reaction between Isocyanate and Hydroxyl.2 The was subsequently called "imitation Swiss cheese"). In 1953
precursors were initially limited to Diisocyanate and Bayer and the Monsanto Chemical Company (now
Polyester; however, the industry began looking for Monsanto Company) formed the Mobay Chemical
different Hydroxyl Polyols due to problems with Corporation to produce polyurethane in the United States.3
processability, cost, foaming physical properties, etc. By Urethane is a product of Hydroxyl compounds such as
1957, various Polyether Polyols were being produced Polyether Polyols and polymerized derivatives o f
,offering not only cost effectiveness but also anti-foaming Isocyanates. Urethane’s unique chemical structure is the
physical properties. Thereafter, the Urethane Industry was result of supplemental reactions between the
able to make leaping advances with the development of Polyisocyanate and Polyol(activation hydrides).
silicone surfactants. Polyurethanes are a class of The external characteristics of the polyurethane end-
extremely versatile polymers that are made into flexible products vary, resulting in diverse characteristics such as
and rigid foams, fibres, elastomers, and surface coatings. marshmellow-like softness to concrete-hardness, or rubber­
They are formed by reacting an isocyanate (a compound like stretch-ability that allows the product to stretch to 10-
having the functional group NCO) with an alcohol (having times its size to materials that offer almost no change in
the functional group OH).Polyurethane molecules can form. Additionally, Urethane, as a comfort and
adopt a linear or a network architecture. Linear convenience-providing material, is used in a wide range o f
polyurethanes are formed by reacting a dialcohol with a applications in automobiles, electronics, furnishings,
diisocyanate, whereas network polyurethanes are formed architecture, etc. because they offer unique qualities
relative to other general high molecular compounds.
from polyfimctional alcohols or isocyanates. Dialcohol
In the Apache Aviator Mask, polyurethane adhesives are
monomers include ethylene glycol (HOCH2CH2OH);
diethylene glycol (HOCH2CH2OCH2CH2OH); 1,4- critical materials for joining dissimilar materials together
butanediol (HOCH2CH2CH2CH2OH); 1,6-hexanediol and providing chemically resistant, mechanically tough

390 / Young, Ciborowski, May, Oswald, Gross

bonds and joints. In this study, two polyurethane polyurethane adhesive-polycarbonate plastic joints. The
adhesives cured at elevated and ambient temperature and rupture strength data will serve as a basis for the
with different surface preparations are compared. As a subsitution of adhesives in planned mask maintenance,
measure of the effectiveness of the adhesives, the bond upgrade, and repair operations where lens replacement is
rupture strengths were measured for the various rubber- required.

MATERIALS AND METHODS Approximately 0. 1 cubic centimeter of adhesive is

applied to each face o f the polycarbonate and then spread
Test items consisting of one inch diameter right circular between the two-polycarbonate mating surfaces.
cylindrical blocks o f polycarbonate, stamped circular Once both surfaces were covered with the adhesive the
rubber coins and polyurethane adhesive were assembled. prepared rubber coins were inserted between the
The polycarbonate substrates were machined from polycarbonate cylinders, compressed together by hand,
polycarbonate 1 -inch diameter rod with a l.Sinch stem excess adhesive removed and die joined components set in
(see figure A) using ASTM 429 as a general guide to the a 1 inch inner diameter plastic tube which acted as a jig to
configuration of test specimens. keep the parts from slipping apart or sliding out of
Once machined, the polycarbonate pieces were sanded alignment during curing. Ten Specimens prepared with
with either with 80 or 100 grit aluminum oxide sand paper, 7500 adhesive were then cured in a 63 degrees Celsius
and cleaned with isopropyl alcohol solution (99% by convection oven for 24 hours . Ten assemblies prepared
volume) to remove any residual polycarbonate dust or grit with 7540 adhesive were also cured in the convection oven
left on the surface from the sand paper. The surfaces were at 63 degrees Celsius for 24 hours. All the remaining
coated with LORD 7701 adhesion promotion solution assemblies were allowed to cure for a minimum o f 48
(consisting o f a mixture of ethyl acetate and alcohols), hours at room temperature.
followed by drying for several minutes while resting on
absorbent paper. The stamped rubber coins were also
sanded with 80 or 100 grit sand paper, followed by
cleaning with isopropyl alcohol solution and application of
the LORD 7701 adhesion promoter. The individual
components were then assembled in groups, which were
differentiated by the use o f either LORD 7500 adhesive or
LORD 7540 adhesive. In any assembly the polycarbonate
surfaces were all either 100 grit sanded or 80- grit sanded.
Some rubber coins were not sanded to determine the
impact o f no rubber surface preparation on joint rupture
strength.
Lord 7500 adhesive is a two part adhesive consisting o f a
black,,extraordinarily viscous polyol which must be mixed
with a cream colored isocyanate mixture in the proportion
of 1 part by weight polyol to 1.7 parts isocyanate. The
resulting mixture then appears a light gray color and
depending on degree of mixing will sometimes display
variegated strands of lighter and darker colors. Good
mixing o f the components results in a visually uniform
color without dark and light striations. The useful life of
this mixture is exceedingly short, on the order of 10
minutes or less. The reaction between polyol and
isocyanate quicly proceeds to the point where it is
impractical to apply the adhesive since it becomes
extraordinarily viscous and too difficult to apply.
Lord 7540 adhesive is also a two part polyurethane
system, however it is manufactured in pre -measured
applicator tubes, which are designed for use in a glue gun .
A static mixing element is attached to the outlet of the
cartridge to blend the two components. In this application
the ratio o f components appears to be 1 to 1 as both
preloaded tubes are of equal diameter. The 7540 adhesive,
when well mixed, is also a light gray color.

Young, Ciborowski, May, Oswald, Gross / 391

 Figure 1: Configuration of Prototypes

392 / Young, Ciborowski, May, Oswald, Gross

 RESULTS

The following rupture load results were obtained when the test specimens were placed under tensile stress in the Instron apparatus.

Load to Induce Rupture

A dhesive Surface Curina Separation M od e Load a t Failure

7500 100 grit 63 C rubber tearing 261+/-31!bs

7540 100 grit 63 C adhesive tear from pc 239+/-41bs
7500 100 grit 23 C adhesive peeled from rubber 195+/-21lbs
7540 80 grit 23 C adhesive peeled from rubber 177+/-36lbs
7540 100 grit 23 C adhesive peeled from rubber 170+/-19 lbs
7500 80 grit 23 C adhesive peeled from rubber 149+/-15 lbs

GROUP SURFACE CURING RUPTURE MODE Load at Rupture

(Lbs!
7500 100 Grit 63 Celsius Rubber Tearing 261+/-31

Young, Ciborowski, May, Oswald, Gross / 393

7540 100 Grit 63 Celsius Adhesive Tearing 2 3 9 + / - 4 1 lbs
from Polycarbonate

7500 100 Grit Ambient Rubber Delaminates 1 9 5 + / - 2 1 lbs

from Adhesive

394 / Young, Ciborowski, May, Oswald, Gross

7540 80 Grit Ambient Rubber Delaminates 1 7 7 + / - 3 6 lbs
from Adhesive

7540 100 Grit Ambient Rubber Delaminates 1 7 0 + /-1 9

from Adhesive

7500 80 Grit Ambient Rubber Delaminates 1 4 9 + /-1 5

Young, Ciborowski, May, Oswald, Gross / 395

 Rupture S trength Com parison
13 Feb 2001 Data

396 / Young, Ciborowski, May, Oswald, Gross

 Rupture Strength
Polycarbonate-Polyurethane Adhesive -Rubber Joints

Young, Ciborowski, May, OswaldGross / 397

 Rubber Delamination: Specimen in Instron Machine

C o n c lu s io n s surface preparations and curing regimes are equivalent.

Curing the specimens at 63 degrees Celsius for a period of
One-inch diameter right circular cylindrical Polycarbonate- 18 to 24 hours appears to strengthen the bonds. Surface
Rubber joints can be effectively bonded with polyurethane treatment and abrasion of the rubber in specimens cured at
adhesives. Preparation of the polycarbonate surface elevated temperature seems to improve the bond strength,
includes sanding with 100 or 80 grit aluminum oxide sand but only minimally from 3% to 6%.
paper, cleaning and removal of grit and abraded Surface treatment of rubber seems to have a stronger
polycarbonate with commercially available over-the- effect on improving bond strength in specimens that are
counter isopropyl alcohol followed by treatment with an not subjected to elevated temperature curing and are
ethyl acetate-based adhesion promoter such as Lord allowed to cure under ambient conditions for several days.
Product 7701. Rubber surfaces can also be pretreated with The average increase in bond strength for surface treated
100 or 80 grit papers or no sanding as long as adhesion and abraded rubber in ambient cured specimens ranges
promoter is applied to the rubber prior to application of from 38% to 58%. The overall bond strength o f ambient
uncured polyurethane adhesive. cured (20 degrees Celsius) specimens irrespective o f
The measurements show that joints of rubber and surface treatment is significantly less than those cured at
polycarbonate bonded with Lord Polyurethane Adhesive 63 degrees Celsius, on the order of 38% weaker than oven
7500 fail under loads that are approximately equivalent to cured samples.
the loads, which cause failure in rubber-polycarbonate
joints bonded with Lord Adhesive 7540, as long as the The following table summarizes the results obtained:

398 / Young, Ciborowski, May, Oswald, Gross

Summary of Results

R ubber P r e d o m in a n t
S u rfa c e C u re J o in t F a ilu r e S tre s s a p p lie d
A d h e s iv e P r e p a r a tio n C o n d itio n M ode a t jo in t fa ilu r e R a n g e
Rubber cleanly delaminates from
7540 none Ambient adhesive with little rubber tearing 146psi +/ - 18%
Ambient Rubber delaminates from adhesive with
7500 80Grit Small areas of Rubber Tearing 174psi +/ - 26%
7500 None Ambient Rubber delaminates from adhesive 182psi +/ - 9%
7540 100 grit Ambient Rubber delaminates from adhesive 217psi +/ - 11%
7540 80Grit Ambient Rubber delaminates from adhesive 231 psi +/ - 1 5%
7500 10Ogrit Ambient Rubber delaminates from adhesive 249psi +/ - 11%

7540 None 63C Cure Large area of rubber tearing 315psi +/ - 14%
7500 None 63C Cure Rubber tearing 333psi +/ - 12%

7540 100 grit 63C Cure Adhesive tearing from polycarbonate 333psi
7500 100 grit 63C Cure Rubber tearing 343psi +1 -8%

2 Organic Chemistry. Hendrickson J.B.; Cram

LITERATURE CITED D.J.;Hammond G.S., McGrow Hill Series in Chemistry;
1970.

1 These findings, reported in the Hebrew -language 3Encyclopaedia Brittanica Online. Polyurethane Entry.
Archeologia Umada'ei Hateva ("Archaeology and Natural
4. Modern Adhesives. McGraw-Hill Science Press, 1998
Sciences"), add a new chapter to the history of adhesives
and to humankind's technological history in general. (An
English translation of the paper is available, as are color
slides of the ancient decorated skulls).

B io g ra p h y highest civilian award, the Superior Civilian Service

Medal, for contributions to the development of
Gerald P. Young is an Engineer at the technologies to safely destroy the Nation’s obsolete
United States Army Chemical stockpiles of Mustard and Nerve Agents without the use of
Biological Center located at Aberdeen incineration. Jerry recently traveled the country providing
Proving Ground, Maryland. With a chemical and biological terrorism awareness training under
Chemical Engineering Degree from the auspices of the Nunn-Lugar-Domenici Domestic
the University o f Delaware, Jerry has Preparedness Act, providing realistic consequence
contributed to the design, management exercises for Firefighters, Police, and
development and production of Emergency Medical Workers in 20 major U.S. Cities.
Pyrotechnics, Explosives, Chemical
Detection and recently, Individual
Respiratory Protection Equipment.
Jerry was awarded the Army’s second

Young, Ciborowski, May, Oswald, Gross / 399

 SESSION K
FURNISHINGS

TECHNICAL SESSION ORGANIZERS

Ray Thomas, Chair
GOLDSCHMIDT CHEMICAL COMPANY

Robert Lockwood
HUNTSMAN POLYURETHANES

Linda Jeng
DOW CHEMICAL CANADA INC

TECHNICAL SESSION MODERATOR

Dimitri Dounis
HICKORY SPRINGS MANUFACTURING COMPANY

TECHNICAL SESSION VICE-MODERATOR

Tony Duley
WOODBRIDGE FOAM CORPORATION
Less Temperature-Sensitive LR Slabstock Foam

MASAKAZU KAGEOKA, HIDEAKI OOTSUKA, YASUHIKO SATO a n d TAKASHI KUMAKI

Mitsui Takeda Chemicals, Inc.
2-17-85 Jusohonmachi
Yodogawa-ku, Osaka 532-8686
Japan

ABSTRACT the production of the LR foams applicable to various

purposes, however the obtained foams show the drastic
The Low Resilience (LR) foam market is now rapidly increases of the foam hardness in cooling down to 0 degree
expanding and has got attention over the past few C[4]. In order to improve the temperature sensitivity of
years[l,2]. The shape conformance, pressure relaxation, the hardness for LR foams, an extensive study on the
vibration and sound dampening, and dynamic energy relationships between the foam morphologies and the
absorption are the key characteristics of LR foam. Those physical properties for the LR foams was carried out.
unique properties make the LR foam versatile in various Mitsui Takeda Chemicals, Inc. has developed novel
applications: pillows, mattress pads, chair seats, etc. polyol for LR foam with reduced temperature sensitivity[3].
In general, the LR foam is difficult to be manufactured; The developed LR foams possessed unique morphology
the cell opening is not an easy task. Moreover there is a showing two glass transition temperatures. The
critical issue, H temperature sensitivity”; the LR foam morphology of the resin matrix is the important factor
becomes glassy hard at low temperature. determining the physical properties of the foam[7].
Mitsui Takeda Chemicals, Inc. recently succeeded in This paper reports the features of novel polyol,
developing a novel LR polyol based on the concept ACTCOL LR-00, as compared with the conventional LR
presented in the preceding paper[3]. The polyol foam based on ACTCOL LR-03.
developed provides a wider process latitude and a reduced
temperature sensitivity of the foam hardness. EXPERIMENTAL
The resilience and the hardness of the LR foam can be
designed by adjusting the formulations. A super soft foam Raw Materials
and rigid foam like high-load-bearing LR foam can also be
obtained. The following raw materials were used in this study:

INTRODUCTION POLYOLS

The Low Resilience (LR) foams, or visco-elastic The conventional LR polyol, ACTCOL LR-03, and the
energy absorbing foams, have been developed for many novel LR polyol, ACTCOL LR-00, were examined in this
applications, e.g., vibration damping, impact energy study. The typical data of those polyol are summarized in
absorption, sound insulation, and comfort cushionings[l-5]. Table 1.
The shape of the glass transition plays a critical role in the
engineering performance of polymer[6]. The Table 1. Typical data for ACTCOL LR polyol
viscoelasticity of the polyurethane foams can be mainly
controlled by designing the chemical structure of the LR-00 LR-03
reactants: the molecular weight distributions and the OH value (mgKOH/g) 107 177
functionalities of the polyols and the polyisocyanates. Viscos ity (m Pa • sec) 560 930
The ACTCOL LR-03 polyol has been widely used for Water content (wt%) 0.03 0.03

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 403

SILICONES Table 2. Storage stability ofpolyol
F-242T, from Shin-Etsu Chemical Co., Ltd.
B -8017, from Th. Goldschmidt AG. ACTCOL LR-03 Storage period: day(s)
Temperature (°c) 1 7 30 90
AMINE CATALYSTS 60 T T T T
DABCO 33LV, from Air Products and Chemicals, Inc. 25 T T T T
15 T T T T
TIN CATALYST 10 T A B B
DABCO T-9, from Air Products and Chemicals, Inc.
0 S S S S
FLAME RETARDANT
ACTCOL LR-00 Storage period: day(s)
Fyrol 780, from Akzo kashima LTD.
Temperature (°c) 1 7 30 90
60 T T T T
ISOCYANATE
25 T T T T
Takenate-80, 80/20 blend of 2,4- and 2,6-toluene
diisocyanate (Takeda Chemical Industries, Ltd.). 15 T T T T
10 T T T T
Blowing Procedure 0 T T T T
T, transparent homogenious liquid; A, hazy liquid;
The foams were prepared by low-pressure machine B, hazy liquid with transparent supernatant; S, solid
blowing procedure. The reactant temperature is maintained
at 25± 1 degree C for polyol component and at 20± 1 LR-03 polyol showed poor storage stability at the
degree C for isocyanate. temperature lower than 10 degree C as shown in Table 2.
The reacting mixture was dispensed into a 500 mm x Those solidified polyol or hazy liquid can be in the original
500 mm open wooden box. The foaming reaction was state, transparent liquid, by heating the system to 25 degree
allowed to take place in this box. One minute after the rise C.
time, the resulted foam was taken from the box and was put The novel polyol, LR-00, were consistently transparent
into the oven for curing at 70 degree C for 1 hour. liquid at any temperature for 3 months.

Physical Properties Test Blowing Characteristics

The general foam properties were tested by JIS 6400. For many years, an LR foam have been difficult to be
The temperature sensitivity tests were carried out at various manufactured: easy to shrink, narrow process latitude,
temperature after 24 hours holding of the specimen. limited grade of ancillary chemicals, difficult to tailor the
In this study, 25%CLD value is measured without hardness and density, etc. ACTCOL LR-00 had been
precompression; the data is read after 20 seconds relaxation developed based on a novel concept to improve those
at 25% compression. The ordinary precompression process above mentioned issues.
may raise a problem; the 75% compressed foam can not go ACTCOL LR-00 provides improved cell opening; Tin
back to 25% deformed position during the measurement level range is wider and the shrinkage can be prevented.
especially at lower temperature. The density range is wide ranging from 120kg/m3 down to
The recovery time is an important factor for LR foam. 20kg/m3. The cell structure is uniform and fine even at low
Some in-house method to assess this property has been index and even at less Tin level. Any type of silicone for
reported[2]. In this study, the recovery time is defined as flexible slabstock can be used. The foam properties, e.g.
the time the finger-compressed foam deforms back to the hardness, resilience, air flow, etc., can be tailored for
original shape: foam specimen configuration, 100 x 100 x diverse market requirements.
50mm thickness; indentation, compressed by a fingertip to
the thickness of ca. 5mm and retained 20 seconds to
equilibrium. Table 3. Typical LR-00 formulations

RESULTS AND DISCUSSION (pbw)

ACTCOL LR-00 100
Storage Stability of Polyol H 20 1,1.2,1.5, 2.0
Glycerin 0 ,1 ,2
In many case, an LR polyol is not stable in terms of Silicone B-8017 1
homogeneity over a certain period. The storage stability of Fyrol 780 10
polyol is examined for ACTCOL LR-03 and LR-00. One DABCO 33LV 0.4
hundred gram of polyol is charged in a sealed 200mL glass T-9 0.1
container, and stored at various temperatures. TDI-80 index 7 0 ,7 5 ,8 0 ,8 5 ,9 0 ,1 0 0 ,1 1 0

404 / Kageoka, Ootsuka, Sato, Kumaki

 The typical formulation for LR-00 is shown in Table 3. and conventional LR, the hardness increases with Tin level.
Glycerin is one of the cross linkers utilized for tailoring the This somewhat strange trend seems coming from the
foam properties. unique morphology of the LR -00 foam.
The hardness changes for the formulations with 2pbw
glycerin are plotted in Figure 1. Although the density
decreased at higher index, the hardness increased for any
water level. The effect of water level on the hardness for
LR-00 was different with that for LR-03: The higher the
water level, the lower the hardness for LR-00 foam.

Figure 2. Effect of Tin level

Temperature Sensitivity

HARDNESS

The temperature dependencies of the hardness of

Figure 1. CLD variation for LR -00 foam LR-00 and LR-03 foam were examined. Table 5 is the data
of the foam for the testing. Figure 3 indicates the distinct
Another unique behavior in LR-00 system is the effect difference in the hardness sensitivity. The LR-03 foam was
of Tin level. The example is described in Table 4 and vitrified at the temperature below 10 degree C, whereas
Figure 2. The resilience and the hardness decreased with an LR-00 foam retained the flexibility even below 0 degree C.
increase of Tin; in the case of the ordinary slabstock foam Although the temperature sensitivity of hardness changes
with formulation factors to some extent, inherently LR-00
Table 4. Effects o f Tin catalyst
provides less temperature sensitive foam comparing to that
based on LR-03 and other conventional LR polyols.
Formulation (pbw) 50M -2 50M-3 50M-4 50M-8
ACTCOL LR -00 100 100 100 100 RESILIENCE
H 20 2 2 2 2
Silicone B-8017 1 1 1 1 The temperature sensitivity of resilience is also
DABC033LV 0.4 0.4 0.4 0.4 important issue as well as hardness; e.g., the impact
Fyrol 780 10 10 10 10 absorbing foam should also absorb the energy even in
T-9 0.1 0,2 0.25 0.3 winter. The sensitivity data for the foam described in Table
TDI-80 31.4 31.4 31.4 31.4 5 is shown in Figure 4. The LR-00 foam shows lower ball
NCO-Index 85 85 85 85 rebound at lower temperature where the LR-03 foam
CT (sec) 15 13 14 14 becomes glassy and rebounds the ball more.
RT (sec) 155 127 118 110
Density (k g /m 3 ) 56.3 57.2 56.8 56.6 Foam Property Tailoring
Ball rebound (%) 9 9 8 7
Recovery time (sec) 1 1.5 1.8 2 ACTCOL LR-00 is a versatile polyol for the production of
25% CLD (N/100cm2)* 20.0 16.8 15.9 14.6 wide spectrum of grades: from super soft to rigid foam like
*without precompression grade. The extreme data is presented in Table 6. The
hardness can be ranged from 12N ILD to rigid like

Kageoka, Ootsuka, Sato, Rumaki / 405

hardness; 25%CLD, 620N.
ACTCOL LR-00, based on novel concept, will be
utilized for a new class of foam manufacturing in near
future.
Table 5. Physical properties for LR foam

#50 -10 #JLR-3

ACTCOL LR-00 100 -

ACTCOL LR-03 - 100

H 20 2 1.5
Glycerin - 3
Silicone B-8017 1 -

Silicone F-242T - 1
Fyrol 780 10 -

DABC033LV 0.4 0.3

T-9 0.3 0.02
TDI-80 29.8 40.4
NCO-Index 80 80
Cream time (sec) 10 18
Rise time (sec) 107 172
Figure 4. Temperature sensitivity of Resilience
Density (k g /m 3 ) 56.0 62.5
Ball rebound (%) 7 2
Recovery time (sec.) 4 5
Table 6. Design o f hardness
25% ILD (N/314cm2) 34 48
Compression set (%) 1.2 0.2
Soft Hard
Tensile strength (kPa) 35 98
ACTCOL LR-00 100 100
Elongation (%) 170 280
H 20 2 1.5
Tear strength (N/m) 200 590
Glycerin - 5
Airflow (cc/cm2 • s) 45.3 0.5
Silicone B-8017 - 1
Silicone F-242T 1 -

DABC033LV 0.2 0.4

Methylene chloride 20 -
T-9 0.2 0.1
TDI-80 28.9 45.6
NCO-Index 80 100
Cream time (sec) 16 11
Rise time (sec) 240 53
Density (k g /m 3 ) 28.6 70.9
Ball rebound (%) 9 13
25% ILD (N/314cm2) 12 -
25% CLD (N/100cm2)* - 620
*without precompression

CONCLUSIONS

Mitsui Takeda Chemicals, Inc. recently succeeded in

developing a novel LR polyol based on the concept
presented in the preceding paper[3]. The polyol
developed provides wider process latitude and reduced
temperature sensitivity of the foam hardness.
The resilience and the hardness of the LR foam can be
designed by adjusting the formulations. A super soft
foam and rigid foam like high-load-bearing LR foam can
Figure 3. Temperature sensitivity of hardness also be obtained.

406 / Kageoka, Ootsuka, Sato, Kumaki

ACKNOWLEDGEMENTS Hideaki Ootsuka

The authors would like to thank the members of the Hideaki Ootsuka joined Takeda
urethane raw materials laboratory, Mitsui Takeda Chemical Industries, ltd. in 1998
Chemicals, Inc., for their continuous interest, after receiving his M.S. degree in
encouragement, and valuable discussions during this work. Organic Chemistry from Chiba
University in 1998. He has been
REFERENCES working on the research and the
development for polyol synthesis.
1. Kageoka, M., M. Kuwagaki, H. Ootsuka, Y. Sato and He is currently working on the
T. Kumaki. 2000, "Polyols for TDI Slabstock Foams R& D for polyol products for the
over Wide Ranges of Hardness and Resiliency," in flexible slabstock / mold foam for
Mandarin, PU China 2000, Shanghai. MTCI.
2. Kintrup, S., J. P. Treboux and H. Mispreuve. 2000.
“Low Resilience-High Performance Recent Advances Yasuhiko Sato
in Viscoelastic Flexible Slabstock Foam,” presented at
POLYURETHANES CONFERENCE 2000, Oct. 8-11, Yasuhiko Sato joined Takeda
2000 Chemical Industries, Ltd. in 1984
3. Kageoka, M., K. Inaoka, T. Kumaki and Y. Tairaka. after receiving his M.S. degree in
1998. “Low Resilience Polyurethane Slabstock Foam polymer chemistry from Kyoto
with Microphase Separated Morphology,” presented at University. He has been
POLYURETHANES EXPO ’98, Sep. 17 -20,1998. working on research and
4. Kageoka, M., K. Inaoka, T. Kumaki and Y. Tairaka. development for polyols and
1998. “Low Resilience Polyurethane Slabstock Foam polyurethane technology. He is
with Reduced Temperature Sensitivity,” presented at currently responsible to the
PU China 98, Shanghai. Environment & Safety
5. Sperling, L. H. and R. Corsaro (Chairmen), Department, MTCI.
symposium held under the auspices of the American
Chemical Society, Dallas, Texas, April 1989; Takashi Kum aki
published as Sound and Vibration Damping with
Polymers, ACS Symposium Series No. 424, Takashi Kumaki joined Takeda
Washington, D. C , 1990. Chemical Industries Ltd. in 1982
6. Sperling, L. H. and J. J. Fay, 1991. “Factors Which after receiving his M.S. degree in
Affect the Glass Transition and Damping Capability of polymer science from Kyoto
Polymer,” Polymers for Advanced Technologies, 2(1) University. He has been
pp. 49-56. working on research and
7. Kageoka, M., Y. Tairaka and K. Kodama. 1995. development of the latex
“Effects of Melamine Particle Size on Flexible technology and the polyol
Polyurethane Foam Properties,” SPI 95, Chicago. synthesis. He is currently the
research manager of the urethane
BIOGRAPHIES raw materials laboratory, MTCI.
M asakazu Kageoka

Masakazu Kageoka joined

Takeda Chemical Industries Ltd.
in 1992 after receiving his B.S.
degree in Industrial Chemistry
from Kobe University in 1990,
and M.S. degree in Applied
Chemistry from Osaka City
University in 1992. He has
been working on the research and
the development for polyols and
polyurethane technology. He is
currently working on the flexible slabstock foam for
MTCI.

Kageoka, Ootsuka, Sato, Kumaki / 407

Flame-Retardant and Physical Characteristics of Flexible
Hybrid Polyurethane Foam Materials Prepared Utilizing
Foam-One Technology

CHANDRA JAYAKODY, DAN MYERS, MALCOLM CROCKER, KELLY BURES, JE F F BRIDGE

a n d MARK BROWN
Chestnut Ridge Foam, Inc.
Route 981 North
RO. Box 781
Latrobe, Pennsylvania 15650

ABSTRACT Appendix F, Part II, oil burner test and FAR 25.853 (a),
Appendix F, Part I, (a), (1), (ii), 12 second vertical Bunsen
A new generation of highly flame-resistant, open-cell, burner test for aircraft seat cushions, FRA performance
flexible polyurethane-polychloroprene hybrid foam criteria for flexible cellular foams used in passenger cars &
materials has been developed utilizing Foam-One locomotive cabs, CAL 133 and ASTM E l 537 for contract
technology. Foam-One is a batch process of furniture, CAL 129, ASTM E 1590, BFD IX - 11, NAVSEA
manufacturing polyurethane flexible slabstock foams under PD-1-00, NFPA 267-98 (modified) for mattress
variable pressure settings varied from 1.0 psi (7 kPa applications.
absolute) to about 27.5 psia (189.6 kPa absolute). This
equipment helped enable us to make a revolutionary family Many commercial and regional aircraft seat cushion
of fire-resistant foams utilizing entirely new technology. assemblies are constructed in such a way that in the case of
The new foam materials will easily meet fire-testing emergency, when an aircraft is forced to land on water, the
standards required by the federal agencies.* In addition to seat bottom cushion could be used as a floatation device.
fire-resistant properties, these new low-density foam In order to be used as a floatation device, the seat bottom
grades show excellent comfort and physical properties. The cushion must to be sufficiently buoyant to support the
applications of this new technology include foam products weight of the passenger in water. Since aircraft seat
such as seat cushions for airline, ground transportation and cushions are made of open-cell flexible foam materials
contract furniture applications and, mattresses for which absorb water easily, this requirement is made by
dormitories, military, marine and naval ships, hospitals, fabricating a piece of polyethylene or polypropylene closed
nursing homes, youth/crib bedding, and psychiatric cell floatation foam within the lower depth of the seat
facilities. In this paper, the variations of physical properties cushion. Even though FR foam materials alone meet the
of the base foams with respect to variable pressure settings oil burner requirement, once this closed cell foam is
and the fire performance characteristics and physical incorporated into the seat cushion, the cushion assembly
properties of these polyurethane-polychloroprene hybrid now may fail the oil burner requirement as flammable
foam materials will be discussed. closed cell foam material may contribute to weight loss of
more than 10 percent when subjected to the oil burner test.
INTRODUCTION The normal procedure to protect this floatation foam is to
cover it with an expensive fire blocking fabric materials
In order to use in aircraft, ground transportation, contract such as Kevlar© and Nomex® blended fabrics.
furniture, and mattress industries, the flexible foam
materials must meet the highly stringent federal fire The cushioning technologies currently available to meet
performance test standards. These include FAR 25.853 (c), the above mentioned oil burner test include incorporation
of heat expandable graphite, impregnation o f flexible
polyurethane foams and use of heat expandable graphite-
modified ethylene-vinyl acetate closed cell flexible foams.
*
Fire-resistant results are based on small- scale laboratory tests for
purpose o f relative comparison and are not intended to reflect the hazards
presented by this or any other material under actual fire conditions.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 409

 In this paper, we describe a new process of making low to the farthest point where damage to the test specimen due
density, fire-resistant polyurethane-polychloroprene hybrid to that area’s combustion, including partial or complete
foam products for aircraft seat cushions, contract furniture, consumption, charring, or embrittlement, but does not
mattresses, ground transportation and other fire-retardant include areas sooted, stained, warped, or discolored. The
foam applications utilizing Foam One Technology. Foam dimension of the seat bottom cushion is 18"x20"x4" while
One technology is a batch process of making polyurethane the seat back cushion is 18"x25"x2". The requirements of
slabstock foams under vacuum and under pressure varied the bum length are; no bum length shall exceed 17 inch on
from 1.0 psi (7 kPa absolute) to about 27.5 psia(189.6 kPa at least 2/3 of the total number of specimen-sets tested.
absolute) [1,2]. This technology enables the making of
low-density foams without using auxiliary blowing agents The percent weight loss for a specimen-set is the pre-test
such as methylene chloride. We have utilized this process weight of the specimen-set less the post-test weight of the
to incorporate a specifically designed fire-retardant specimen-set expressed as the percentage of the pre-test
package into the base polyurethane foams so that the post- weight. All droppings falling from the specimens and
treated foams easily meet all federal requirements at a very mounting stand are to be discarded prior to determining the
low impregnation level. This has significantly lowered the post-test weight. The requirement for passing the percent
density and has improved the physical properties of the weight loss is the average percentage weight loss shall not
foam materials. exceed 10 percent and the weight loss of at least 2/3 of the
total number of specimen-sets tested shall not exceed 10
The flame retardant behavior of six new aircraft foam percent [4].
grades was evaluated using the vertical Bunsen burner test
for cabin and cargo component materials (FAR 25.853 (a), The other fire performance characteristics of these foam
Appendix F, Part I, (a)(l)(i) and (a)(l)(ii)), the oil burner grades were measured according to CAL 133, ASTM E-
test for seat cushions (FAR 25.853 (c), Appendix F, Part II, 1537 for contract furniture, CAL 129, ASTM E 1590, City
Cone Calorimeter (ASTM E 1354), Buoyancy and Extreme of Boston BFD IX - 11 for mattresses, FRA performance
Temperature tests (FAA Technical Standard Order No. criteria [radiant panel flame spread index (ASTM D 3675)
TSO-C72c) and other ASTM test standards. and smoke density (ASTM E 662)] for flexible cellular
foams used in passenger cars and locomotive cabs and
Vertical Bunsen Burner Test NAVSEA PD 1-00, NFPA 267-98 (modified) standard test
for military and shipboard mattresses. The physical
The vertical Bunsen burner test method is intended for properties were measured according to the ASTM D 3574
use in determining the fire resistance of cabin and cargo test methods.
compartment materials when tested according to the 60-sec
and 12-sec vertical Bunsen burner tests specified in FAR EXPERIMENTAL
25.853(a), Appendix F, Part I, (a)(l)(i) and (a)(l)(ii)
respectively. The requirements of the three test parameters For aircraft seat cushions, the low density and low IFD
(flame time, drip flame time and bum length) are: the foam grades were prepared under vacuum settings while
average flame time for all of the specimens tested (at least high IFD foam grades were prepared under positive
three) shall not exceed 15 sec for either 12-sec or the 60- pressure settings. A specifically designed fire-retardant
sec vertical test, the average drip extinguishing time for all package was utilized for each of the foam grades. The
of the specimens tested shall not exceed 3 sec for the 60- block length and the width of the base foam buns were 100
sec vertical test or 5 sec for the 12-sec vertical test, and the inches and 83 inches respectively, while bun heights were
average bum length for all of the specimens tested shall not in the range of 35 to 45 inches depending on the foam
exceed 6 in (152 mm) for the 60-sec test or 8 in (203 mm) grades. A representative foam formulation is shown
for the 12-sec test [3]. below.

Oil Burner Test for Seat Cushions Ingredient php

The oil burner test method (FAR 25.853 (c), Appendix F, Polyether polyol 100
Part II) evaluates the bum resistance and weight loss Water 3.60-6.00
characteristics of aircraft seat composites when exposed to Amine catalyst 0.17-0.20
a high-intensity open flame. Tin Catalyst 0.17-0.22
Surfactant 0.70-1.20
The four principal bum lengths are measured along the TDI Index 105-120
topside of the horizontal seat cushion, bottomside of the Pressure 1.0 p si - 2 7 .5 psi.
horizontal seat cushion, frontside of the vertical seat
cushion, and the backside of the vertical seat cushion. The The base foam grades were post-treated with a
four bum lengths are defined as the distance measured, in polychloroprene compound to obtain the hybrid foam
inches, from the edge of the seat frame nearest the burner grades with a variety of densities and firmness choices.

410 / Jayakody, Myers, Crocker, Bures, Bridge, Brown

RESULTS AND DISCUSSION for aircraft applications such as carpet underlay and
miscellaneous cushioning. All density and IFD values are
Effect of Variable Pressure on Base Foam Properties average values and have ± 10% ranges.

The effects of varying levels of pressure settings on foam FAR 25.853 (a), Appendix F, Part I - Vertical Burn
properties were studied using a fire -retardant formulation Test Results
containing 3.60 parts of water per 100 parts of polyol. For
this study, a mini-scale Foam-One equipment was utilized. All six foam grades easily met both the 12-second and
The isocyanate index of this formulation was 110. The 60-second vertical Bunsen burner test requirements. The
results are shown in Table 1. The pressure variation has a results, which are summarized in Table 2, represent an
significant effect on foam density and the firmness of the average of more than 20 test trials.
foams. The density, firmness and the tensile strength of the
foam increase with increasing pressure levels. Compared The flame time for both 12-second and 60-second bum
to the foam prepared under atmospheric pressure (14.5 tests showed zero seconds. These foam materials did not
psi), formulations run under vacuum (below 14.5 psi) gave bum with a flame and showed no drippings after the
low density, low IFD foams while formulations run under removal of the Bunsen burner. The bum length test results
positive pressure (above 14.5 psi) gave high density, high are also below the required bum length standards under
IFD foam grades. Elongation and support factors are both 12-sec and 60-sec tests.
largely unaffected.
FAR 25.853 (c), Appendix F, Part II - Oil Burner Test
APPLICATIONS Results

Due to their excellent flame resistance and physical FAR 25.853 (c) is a composite test performed using all
properties, these hybrid foam materials are ideal candidate seat components. Individual foam grades were also tested
materials for many other foam applications. Such with and without upholstery covering materials. Both
applications include use of these foam materials in aircraft, individual foam grades as well as composite seat cushions
transportation, contract furniture, and mattress industries. easily passed the FAA oil burner requirement for aircraft
seat cushions. The weight losses were in the range of
Aircraft Seat Cushion Applications 4.0% to 8.5%. and the flames self extinguished within 1-3
minutes after the removal of the oil burner flame. One of
Using this technology, six foam grades, trade named the great advantages of Airflex® foam grades is that they
Airflex®, have been specifically developed for aircraft helped to obtain consistent compliance; even with
seat cushion applications. The average densities and the upholstery materials considered most difficult to pass this
indentation force deflection ranges (IFD) (at 25% test such as treated leather materials. With Airflex® foam,
deflection) measured at 2" thickness of all six Airflex® these upholstery materials have shown a greater tendency
foam grades are presented in the Figure 1. to pass the oil burner test. The test also produced low
smoke generation and this was further evident from the
The six Airflex® foam grades are color coded with six ASTM E 662 smoke density test results. All these foam
different colors. The grades are named after their grades also passed the Airbus Industrie ATS 1000.001 and
respective IFD ranges (lbs) at 2-inch thickness. For Boeing BSS 7239 toxicity test standards.
example, Airflex® 15-25 represents IFD in the range of 15
to 25 lbs. Two low-density foam grades, Airflex® 15-25 Cone Calorimeter Test Results
(red) and Airflex® 30-40 (light green) showed an average
density of 2.70 lbs/ft3 (43.2 kg/m3) and are designed for The fire performance characteristics of some Airflex®
seat back cushion applications. The average IFD values of foam grades was also studied using the Cone Calorimeter
these two grades at 4-inch thickness are 30.8 lbs (137 N) (ASTM E-1354), the instrument that has emerged as an
and 40.0 lbs (178 N) respectively. Airflex® 20-30 important tool for the analytical testing of materials for
(lavender), 40-50 (yellow) and 55-65 (orange) grades have their fire properties. The cone data of 4"x 4"xl" foam
an average density of 3.60 lbs/ft3 (57.7 kg/m3) and are specimens were conducted at 35 and 50 kW/m2 exposure
designed for aircraft seat bottom cushion applications. The in a horizontal orientation using a Stanton Redcroft cone
average IFD values of these three grades at 4-inch calorimeter. Airflex® 20-30 gave a peak heat release of
thickness are 38.3 lbs (170 N) 53.8 lbs (239 N) and 74.9 98.9 kW/m2 and smoke generation (average specific
lbs (333 N) respectively. Airflex® 90-110 (blue in color) extinction area) of 171.9 m2/kg at 35 kW/m2 exposure.
is a high density, high IFD foam grade with an average These values represent a reduction of peak HRR by 82 %
density of 4.50 lbs/ft3 (72.1 kg/m3) and IFD of 90-110 lbs and a reduction of smoke by 65% compared to the base
at 2" thicknesses. The average IFD of this foam grade at foam material which has an average peak heat release rate
4-inch thickness is 110 lbs (489 N). This grade is designed

Jayakody, Myers, Crocker, Bures, Bridge, Brown / 411

T able 1. The E ffect o f P re ssu re on the P h y sic a l P ro p e rtie s o f W ater - B low n F lex ib le P o ly u re th a n e F o a m s (W ater: 3 .6 0 ph p,
Index: 110)

Variable Pressure (psi)a

7.50 10.5 12.0 14.5 17.0 19.5

Density (pcf) 1.04 1.37 1.77 2.11 2.40 2.59

IFD at 25%, lbs 10.76 17.76 22.04 35.05 45.70 58.32
IFD at 65%, lbs 20.24 32.94 44.26 66.68 88.32 115.04
Support factor 1.9 1.9 2.0 1.9 1.9 1.8
Compression set, % 8.90 6.82 7.34 3.19 5.02 3.89
Tensile Strength ,lbs/in 6.69 8.96 10.78 13.66 16.12 19.86
Elongation, % 184 162 181 175 192 192
Tear strength, psi 1.80 1.90 3.90 2.60 2.60 3.00
Resilience, % 40 43 40 40 38 40

F ig u re 1. D e n sity a n d IF D ra n g es f o r A irfle x ® fo a m G ra d es

4 1 2 / Jayakody, Myers, Crocker, Bures, Bridge, Brown

 T able 2. V ertical B unsen B u rn er T est R esu lts o f A irfle x ® F oam G ra d e s

Foam Grade 12-sec Vertical Bum Test 60-sec Vertical Bum Test
Flame Drip Bum Flame Drip flame Bum Length,
time, sec Flame time Length, time, sec time (sec) in (mm)
(sec) in (mm)
Airflex® 15-25 0 0, No drips 4.2 (107) 0 0, No drips 5.0 (127)
Airflex® 30-40 0 0, No drips 4.0(102) 0 0, No drips 5.1 (130)
Airflex® 20-30 0 0, No drips 4.6(117) 0 0, No drips 5.3(135)
Airflex® 40-50 0 0, No drips 5.0(127) 0 0, No drips 5.5 (140)
Airflex® 55-65 0 0, No drips 5.0(127) 0 0, No drips 5.6(142)
Airflex® 90-110 0 0, No drips 4.5(114) 0 0, No drips 5.4(137)

T able 3. P h y sic a l P ro p e rtie s o f A irfle x ® F oam G ra d es

Property Test Method Airflex® Grade Result
Compression set (50% ASTM D3574, Test D, All grades 10% max.
deflection) Percent of original thickness
Tensile Strength ASTM D3574, Test E All grades 10 psi (69.0 kPa) min
(except 90-110)
90-110 20 psi (137.9 kPa) min
Elongation ASTM D3574, Test E All grades 125 % min
(except 90-110)
90-110 80 % min
Tear Strength ASTM D3574, Test E All grades 2.0 lbs/in (350.2 N/m) min.
Dynamic Fatigue test by ASTM D3574, Test I3-IFD loss at AF 20-30 18% max.
constant force pounding 40% deflection (80,000 cycles, AF 40-50 18% max.
measured 24 hrs after test completion AF 55-65 18% max.
Flex Fatigue ASTM D1055, Suffix H-250,000 All grades 5% max.
cycles (thickness loss) (except 90-110)
Resilience ASTM D3574, Test H, Ball rebound All grades 30% min
test)
Support factor 65% lFD/25% IFD All grades 2.1 min.

of 550 kW/m2 and specific extinction area of 493 at this 35 are subjected to both mental and physical stresses when
kW/m2 exposure level [5]. The peak heat release rate and exposed to vibrations, noise and different weather
smoke generation of Airflex® 20-30 at 50 kW/m2 conditions. Polyurethane foams are instrumental in
exposure were 117.6 kW/m2 and 86.5 m2/kg respectively. reducing those stresses. To meet these goals, Airflex®
Airflex® 55-65 grade also reduced both peak heat release seat cushions are designed to achieve longer service life
rate and smoke generation significantly when tested at 35 with better firmness retention, less thickness loss over
kW/m2 exposure. At 35 kW/m2 exposure, peak HRR of time, higher resiliency, good support factors, and improved
this material was 107.4 kW/m2, a 81% reduction and aging characteristics. These new grades also showed
Average SEA of 135.5 m2/kg, a reduction of 73% approximately 12-27% weight savings over previous
compared to the base foam material at this exposure. Cone polyurethane/polychloroprene technology. The physical
data also showed very low values of average heat release properties of these Airflex® foam grades are summarized
rate, total heat release, average effective heat of in Table 3.
combustion and average mass loss rate for these foam
grades when compared to base foam material. Contract Furniture Applications

Physical Properties Some Airflex® grades such as Airflex® 20-30 (89-133

N), which has a great comfortability, are currently being
In addition to stringent fire testing standards, used to make fire resistant contract furniture that often
improvements in passenger comfort continue to be one of needs to meet the California Technical Bulletin No. 133
the key needs of the global airline industry. Comfort and ASTM E 1537 requirements.
experience is a combination of many different factors,
including aesthetics. During a journey, aircraft passengers

Jayakody, Myers, Crocker, Bures, Bridge, Brown / 413

Fire-Retardant Mattress Applications These requirements are a) Net peak heat release rate
shall not exceed 150 kW b) Burning droplets shall not be
Fires and fire deaths in which a bed was listed as the first formed and dropped and c) average specific extinction area
item to ignite are major contributors to the US fire (Ave. SEA) shall not exceed 300 m2/kg. The foam
statistics [6]. These foam grades can be used to make mattresses prepared using this technology with 4 to 6 inch
traditional flat surfaces or convoluted with “waffle design” full thickness, will easily meet these requirements. The
or similar convoluted surfaces. Airflex®, used to results are shown in Table 4.
encapsulate fire-retardant polyurethane, demonstrates
excellent performance when fire-tested in composite to The net peak heat release rate obtained for our mattress
California Technical Bulletin No. 129. The use of was 22.92 kW, a 85% reduction compared to the
Airflex® as an exterior cushioning barrier provides facility requirement. Also, this foam mattress did not produce any
occupants and consumers additional escape time in the burning droplets. The average specific extinction area was
event of a mattress being exposed to cigarettes or open 65.7 m2/kg, a reduction of 78 % compared to the
flame ignition sources. In addition, the application as a requirement.
“barrier” eliminates potential “wear and tear” changes in
long term testing performance which are possible with Flexible Cellular Foam Applications for Passenger
many fabric type barriers. This concept provides a Cars and Locomotive Cabs
beneficial and economical choice of fire-resistant
mattresses for hospitals, nursing homes, universities and Variations of this technology can also be utilized to
residential sectors of youth and crib bedding. For make flexible foams that meet Federal Railroad
increased fire-performance, Airflex® can be used to make Administration (FRA) requirements for flexible cellular
mattresses in a 4-6 inch full thickness for psychiatric and foam materials used in passenger cars and locomotive cabs
military applications to achieve high levels of such as Radiant Panel flame spread index (Is < 25) (ASTM
comfortability as well as durability. D 3675), and ASTM E 662 smoke density requirement [Ds
(1.5 min) < 100, Ds (4.0 min) <175)] thus finding seat
Military and Shipboard Mattress Applications cushion applications in the public surface transportation
industry. We are currently in the process of developing
Military and shipboard mattress/bedding assemblies special soft and medium foam grades with low-density for
must meet the NAVSEA PD-1-00, NFPA 267-98
(modified) standard test requirements.

T able 4. N AVSEA P D 1-00, N FPA 2 6 7 -9 8 (M odified) S ta n d a rd T est R esu lts f o r 4-inch F u ll Thickness M ilita ry
a n d S h ip b o a rd M a ttre sse s (D im en sion : 76 '*x26 " x4 " F a b ric: 1 0 0 % FR C otton).

Requirement Result
Net Peak HRR Shall not exceed 150 kW 22.92 kW
Burning Droplets Shall not be formed & dropped No burning droplets
Ave. SEA Shall not exceed 300 m2/kg 65.7 m2/kg

T able 5. P h y sic a l a n d F la m e R e sista n t C h a ra c teristic s o f th e S u rface T ran sp o rta tio n F oam G rades.

Property Test method Result

Density ASTM D3574, Test A 4.00-5.00 lb/ft3
(64.1-80.1 kg/m3)
Indentation force deflection (25% deflection) ASTM D3574, T estB l 20- 30 lbs
Measured at 2” thickness (89-133 N)
Tensile strength ASTM D3574, Test E 12.0 psi (82.7 kPa) min
Elongation ASTM D3574, Test E 188 % m in
Tear strength ASTM D3574, Test E 2.5 lb/in (4.38 N/cm) min
Compression set (50% deflection) ASTM D3574, Test D, 10 % max
Percent of original thickness
Resilience Ball rebound test 30% min
Flame spread index ASTM D 3675, Radiant Panel* 10.0 max
Smoke density ASTM E662 50-113-150 (flaming)
(D s 1.5 min " Ds 4min“D s M ax (corr) 75-150-200 (non­flaming)
# Instrument constant, p = 1.08 °F/Btu.min

414 / Jayakody, Myers, Crocker, Bures, Bridge, Brown

surface transit applications with radiant panel flame spread contribution to introduce this new foam grades. Thanks are
index as low as 10, and smoke density well below the FRA also due to Mr. Allen Hendrix of J.M. Huber Corporation
maximum requirements. These foam grades will meet all for running cone calorimeter tests.
FRA requirements at the density as low as 4.0 lbs/ft3 (64
kg/m3) and the IFD (25% deflection) range of 20-30 lbs REFERENCES
(89-133 N) at 2-inch thickness. These foam grades, trade
named not yet assigned, will be introduced to the transit 1. Carson, S. 1993. “Method and apparatus for forming
market very shortly. Radiant panel and smoke data articles made of polyurethane” US Patent 5,182,313.
together with some physical properties of these foam 2. Carson, S. 1995. “Apparatus for manufacturing
materials are shown in Table 5. articles made of polyurethane” US Patent 5,456,586.
3. Aircraft Material Fire Test Handbook, 1990.
CONCLUSIONS DOT/FAA/CT-89/15, FAA Technical Center, Atlantic
City International Airport, NJ 08405, pp 1-1 to 1-10.
A new generation of low-density, flexible, hybrid 4. Aircraft Material Fire Test Handbook, 1990.
polyurethane foam material has been developed utilizing DOT/FAA/CT-89/15, FAA Technical Center, Atlantic
Foam-One technology. This new equipment enabled us to City International Airport, NJ 08405, pp 7-1 to 7-14.
make a series of substantial improvements in our fire- 5. Jayakody, C., D. O. Myers, U. Sorathia, and G. L.
retardant cushioning technology. In addition to Nelson. 2000. “Fire-retardant Characteristics of
lightweight, comfort, and excellent physical properties, the Water-Blown Molded Flexible Polyurethane Foam
new Airflex foam grades consistently and easily meet all Materials,” J. o f Fire Sciences, 18(6): 430-455.
fire-testing standards required by the Federal Aviation 6. Ahrens, M. 1999. “Risky Behavior,” NFPA Journal,
Administration including full-scale FAR 25.853 (c), Jan/Feb:43-47.
Appendix F, Part II, oil burner test with weight loss
ranging from 4.0% to 8.5% depending on the seat BIOGRAPHIES
construction. Since these foam grades are highly flame
resistant, it is not necessary to cover them with expensive C handra Jayakody
fire blocking fabric materials (except the base of the seat Chandra Jayakody is a R&D
bottom cushion with polyethylene floatation foam Scientist at Chestnut Ridge
fabrication) to meet the oil burner test. These foam grades Foam, Inc. He received his
can also be used to make healthcare, psychiatric, military, B.Sc, Postgraduate degree in
university and crib/youth mattresses that need to pass fire Applied Statistics, and M. Phil
performance standards such as CAL. TB No. 129, City of degree in organic chemistry
Boston BFD IX-11, ASTM E-1590 and the US Navy from the University of
(NAVSEA PD 1-00, NFPA 267-98 modified). The Colombo, Sri Lanka in 1981,
contract furniture applications include Cal TB No. 133 and 1986 and 1990 respectively.
ASTM E-1537. He received his Ph.D. degree
in Organic/Polymer chemistry
Cone calorimeter test results of these foam grades, even from Marquette University,
at 50 kW/m2 heat exposure showed very low peak heat Milwaukee, WI, in 1994 and postdoctoral training at
release rates, average heat release rates, total heat release, Florida Institute of Technology. Dr. Jayakody’s current
mass loss rates and smoke generation compared to the base research interests are development of new fire-resistant
polyurethane foam materials. foam materials. He has published 30 papers including 3
book chapters in the field of fire -retardant polymers. Dr.
Specially designed surface transportation foam grades Jayakody is a member of the ASTM Committee E-05 on
using this technology easily passed the radiant panel flame Fire Standards and also a member of the American
spread index and smoke density requirements for Chemical Society.
transportation foams (ASTM D3675 and ASTM E662
respectively) with flame spread index below 10 and smoke Dan Myers
density values below the FRA standards. Dan Myers is the Technical
Director of Chestnut Ridge
ACKNOWLEDGEMENTS Foam, Inc., Latrobe, PA. He is
a graduate of Thiel College at
We wish to thank Mr. Scott Baton, Chairman of Greenville, PA with thirty
Chestnut Ridge Foam, Inc., for giving permission to years of experience in the latex
present these results. We also would like to thank Foam and polyurethane foam
One operation crew, Mr. Jamie Albert, Mr. Ken Spillar, industries. Mr. Myers has
Mr. Tony Rich, Mr. Richard Baker, Mr. Joe Jamison, and extensive knowledge of
Mr. Terry Rolla, Quality Control Manager for their industry physical and

Jayakody, Myers, Crocker, Bures, Bridge, Brown / 415

flammability test standards. He is a member of the Mark Brown
American Society for Testing Materials.
Mark Brown received his BS degree in Chemistry from the
Malcolm Crocker University of Pittsburgh, Pennsylvania in 1994. He was
Malcolm Crocker is West instrumental in implementing the Foam-One equipment
Plant Manager of Chestnut and running foam formulations initially.
Ridge Foam whose
responsibilities also include
Foam One operation
management. Mr. Crocker, a
graduate of the College of
Aeronautics at Cranfield,
England, has over 25 years
experience in processing
polyurethane foams. Before
joining Chestnut Ridge Foam
in 1999, he was Corporate Technical Director for Future
Foam. Mr. Crocker’s background also includes working
for Dunlop in Australia, B.F. Goodrich in British
Columbia and Calgary in Canada and Leggett & Platt in
USA. Mr. Crocker is a member of the American Institute
of Chemical Engineers and also a member of the American
Chemical Society.

Kelly Bures

Kelly Bures joined Chestnut

Ridge Foam in 1994 and is
currently Senior Laboratory
Technician whose responsi­
bilities include Production
Quality Assurance and
physical testing for R&D. Ms.
Bures has extensive
experience of ASTM test
standards on flammability and
physical characteristics of
flexible foam products.

Jeff Bridge

Jeff Bridge joined Chestnut

Ridge Foam inc., in 1987 and
is currently Laboratory
Manager whose responsi­
bilities also include R&D and
Quality Assurance. Mr.
Bridge has been instrumental
in combining the poly-
chloroprene and poly ­
urethane to achieve a low
density fire retardant
cushioning product, which integrates the best qualities of
both technologies to create economical foam with
applications in many areas.

4 1 6 / Jayakody, Myers, Crocker, Bures, Bridge, Brown

 A New Generation of Silicone Surfactants Optimized for
Liquid Carbon Dioxide Blown Slabstock Foams

O. EYRISCH a n d G. BURKHART R. BORGOGELLI

Goldschmidt AG Goldschmidt Chemical Corporation
Goldschmidtstrafie 100 914 E. Randolph Road
45127 Essen Hopewell, Virginia
Germany

ABSTRACT

In the last decade, the Liquid C 02 Technology for opportunities for cost savings by optimizing use level and
manufacturing Flexible Polyurethane Foam (FPF) has reducing top-to-bottom density variations.
evolved from a novel experimental technique to an
established widely used process. During this development, INTRODUCTION
both the equipment manufacturers and foam producers
refined their knowledge to successfully implement liquid How can a fine cell structure be achieved?
C 02 blowing in low density and soft high-density foam In conventional foaming, all o f the C 02 is generated
grades. In parallel, additive manufacturers also met the chemically by the blowing reaction. In contrast, a certain
specific challenges required o f silicone surfactants in portion o f the total C 02 in Liquid C 02 foaming - usually
support o f this technology [1-4]. Regardless o f this varying from 2 to 5 pphp (parts per hundred parts polyol) -
progress, some drawbacks still remain. Achieving an is solubilized in the polyol stream under pressure. As the
optimal cell structure and minimizing the distribution o f reaction mixture passes from the mixhead to the foaming
physical properties within the block are still areas that need conveyor, the pressure release at the diffuser liberates the
some improvement. In response to these targets, solubilized C 02, causing it to immediately fill the gas
Goldschmidt has developed two new silicone surfactants, bubbles and mechanically froth the liquid.
namely TEGOSTAB® B 8050 and TEGOSTAB® B 8250, It has been suggested that under pressure, solubilized
that exhibit enhanced properties as compared to established C 02 may reduce the viscosity o f the polyol under industrial
products like TEGOSTAB® BF 2370 and TEGOSTAB® B conditions by more than 50% [5]. Furthermore, the surface
8240. This development addresses the full latitude o f tension o f the reaction mixture is significantly reduced by
processing and efficiency characteristics that are needed to pressurized C 02. For example, pressurizing a standard
cover the growing variety o f foam grades made with the polyol with 15 bar C 02 at room temperature reduces the
Liquid C 02Process. surface tension from 32 to 27 mN/m [5]. Both o f these
The performance o f these new silicones has been effects, the lower viscosity and lower surface tension o f the
extensively studied on the liquid C 02 pilot line at polyol-C02 blend, tend to promote shear-force-induced
Goldschmidt, as well as on industrial Novaflex®, CarDio® nucleation in Liquid-C02-Technology. This phenomenon
and Beamech CO-2™ machines, to prove their broad may explain the fine-celled froths that are obtainable in
applicability in all processes. principle. Furthermore, the comparably lower amounts o f
With enhanced nucleation properties, these surfactants nitrogen necessary to nucleate foam formulations with
allow to make foams that are consistently finer and more higher C 02 levels could also be a consequence o f the lower
regular in cell structure by compensating for disturbing viscosity, rather than the self-nucleating-effect caused by
influences caused by variations in production conditions solubilized C 02[6].
(machine and climate parameters), as well as differences in Although nucleation seems to be facilitated, the final cell
raw material quality and sources. structure o f the resulting foam is challenged by coalescence
Furthermore, the outstanding efficiency and processing o f initially frothed fine cells, particularly in the case o f
latitude o f the new silicone surfactants provides liquid C 02 foam processes.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 417

 The initial mechanical froth height at time to is similar to R E SU L T S
the foam height in conventional processes after a certain
time ti when the comparable amount o f C 0 2 has been C ell Structure
chemically generated by the blowing reaction (Figure 1).
Since at to there is no chemical stability contribution from To optimize nucleation, an ideal balance must be reached
the polymerization reaction, the physical stability o f the among various machine parameters, including mixing head
froth is challenged, and the cells initially formed at pressure and nitrogen load, to achieve the desired froth and
nucleation may begin to coalesce. Unless physically foam cell structure.
stabilized by the proper surfactant, the large number o f fine Figure 2 shows the dependency o f cell structure on the
cells may become a fewer number o f coarser cells. At ti nitrogen level using a formulation with 1.5 pphp o f C 0 2.
however, a certain amount o f viscosity - building polyurea Generally, as nitrogen is increased, the resulting foam
has already been formed, thereby providing some chemical becom es finer in cell structure. Exceeding a certain
stability. The physical stabilization provided by surfactants nitrogen level (in this experiment 20 N L) however, does
in this case is not as essential for maintaining a fine cell not further improve the cell structure. Remarkable also is
structure. the increase in irregularity o f the cell structure at higher
nitrogen levels. Although the average cell count increases,
it appears that larger cell diameters cannot be avoided by
applying elevated nitrogen levels.

Figure 1. Conventional and Liquid-COr-foaming

Given the differences between conventional and liquid
C 0 2 foaming as outlined above, a silicone surfactant must
support the follow ing properties to achieve a fine and
regular cell structure: Figure 2. Influence of nitrogen on cell structure
The need to apply higher nitrogen levels to achieve a finer
1) Nucleation cell structure is also dependent on the choice o f the
supporting conditions for the formation o f the nuclei surfactant. Figure 3 shows the benefit o f the new
created by the mechanical shearing process; development TEGOSTAB® B 8050, indicating a higher
2) Frothing nucleation efficiency compared to TEGOSTAB® BF 2370,
rapidly moving to the gas/liquid interface to
since a fine cell structure can be achieved with lower
facilitate the homogeneous release o f C 0 2 into the
nitrogen addition. This in turn reduces the possibility o f
nuclei and ensure that the C 0 2 does not boil out o f
pinholes, which often result when excessive levels o f
the liquid froth, nitrogen are used in an effort to increase nucleation. The
3) Coalescence Minimization foamer can profit by the pronounced high “nucleation
stabilizing the growing bubbles in the rising foam to
latitude” o f TEGOSTAB® B 8050.
ensure a uniform expansion o f the foam cells without
agglomeration.

In the follow ing section, results are presented from

evaluations on the Novaflex® pilot line at Goldschmidt, as
w ell as customer plant trials on industrial Novaflex® and
CarDio® equipment.
Silicone surfactants were evaluated both for
improvements to cell structure (related to the tasks o f
nucleation, frothing and coalescence minimization), as well
as stability parameters (improvements to foam yield, block
height and density distribution) which are also applicable
for conventional foaming.
Figure 3. Average cell counts depending on nitrogen load
and silicone surfactant

4 1 8 / Eyrisch, Burkhart, Borgogelli

 The proper choice o f surfactant also allows foams to be • 6 bar M ixing Head Pressure
produced with improved cell regularity, even at relatively • 20 Normal Litres o f Nitrogen
high nitrogen levels. Figure 4 shows microscope The centers o f the circles in figure 5 correspond to the
photographs o f two liquid C 0 2 blown foams, each average cell count o f the foam and the circle diameters
produced with 20 N litres o f nitrogen. It is obvious that the reflect the regularity o f the resulting cell structure.
foam made with TEGOSTAB® B 8050 has a finer cell Both new products, TEGOSTAB® B 8050 and
structure. But it is also true that this foam has a more TEGOSTAB® B 8250, provide the finest cell structures in
uniform cell structure, indicating that TEGOSTAB® B this comparative study.
8050 better resists the coalescence o f cells that result in a
more irregular cell structure, which can often occur at
elevated nitrogen levels.

Figure 5. Benchmarking of different silicone surfactants

according to celt structure achieved on Novaflex9
In order to compare the frothing efficiency o f a silicone
surfactant, w e first established classifications o f different
froth qualities according to the stiffness (viscosity) and
fineness (cell size) o f the froth (table 1).

table 1. Classification of different froth qualities

category viscosity Cell - structure Estimated characterization
cell diameter
[mm]
A liquid 0 Coarse >2 Like bathing foam

Froth easily 1 Medium 2 - 0 ,5 Like low viscous

migrates on beer - foam
the bottom 2 Fine <0,5 Like liquid cream
of the box
Figure 4. Microscope photograph of foams made using 2
B medium 0 Coarse 2 - 0 ,5 Like bathing foam
pphpofC 02
“waves - on 1 Medium <2 Like beer ­ foam
the surface
B y utilizing this procedure, various silicone surfactants
2 fine <0,5 Like liquid cream
were evaluated on the Novaflex® pilot line to benchmark
products according to their nucleation and frothing 1
to

C viscous
CM

o*

medium C oarse shaving

l

cream
efficiency. With the exception o f TEGOSTAB® B 8229,
2 fine <0,5 Shaving cream
which is not recommended for use in liquid C 0 2processes,
all other candidates resisted observable cell coalescence,
D high 1 medium 2 - 0 ,5 Like cream
which can easily be detected by a “ spraying phenomenon ” whipped too long
viscous
on the surface o f the rising foam. Froth forms 2 fine <0,5
The comparison shown in figure 5 is based on the “sa u ssag e
layers*
follow ing parameters:
• 5 pphp Water Different froth viscosities can easily be obtained by
• 2 pphp C 0 2 varying nitrogen levels in formulations using low amounts
• 1.1 pphp Surfactant o f liquid C 0 2. A lack o f nitrogen often leads to very liquid
froths (category A ) that are not able to capture all the

Eyrisch, Burkhart, Borgogelli / 419

formulated C 0 2. An excess o f nitrogen sometimes yields Reduced shear combined with a smaller open area
froths that are too viscous (category D ) and form layers that content (1%) resulted in a 15 bar mixing head pressure.
are hindered in migration. The disturbances at the interface The resulting froths were lower in viscosity than those
o f these layers result in cell coalescence, since the froth is produced with the tw o - sieve setup, which yielded a mixing
too stiff to “knit” together. A fine cell structure combined head pressure o f only 5 to 6 bar.
with striations across the bun might be the consequence Under conditions o f higher shear and lower mixhead
(figure 6). pressure, froth quality between surfactants could not be
differentiated. However, foam produced with the new
TEGOSTAB® B 8250 was finer in cell structure than that
produced with TEGOSTAB® B 8240, when evaluated at
the lower surfactant level (run #7 versus #8). When
surfactant concentration is increased, both products result
in a finer cell structure, and differentiation is no longer
possible (run #5 versus #6).
Under conditions o f lower shear and higher mixhead
pressures, generally poorer foam results are obtained since
nucleation and froth development becom es more critical
(runs #2-3 versus runs #5-6). However, foam quality with
both TEGOSTAB® B 8250 and B 8240 was better than
with a competitive surfactant (run #4). The very fine - celled
viscous froth produced by the competitive surfactant, in
combination with the poorer foam cell structure is an
indication that cell coalescence has occurred. The best
results under these most sensitive high -pressure conditions
were obtained using TEGOSTAB® B 8050 (run #1).
Figure 6. Striations caused by too stiff froth layers in a
discontinuous process table 2. trials on Novaflex^-pilot-plant
Average Cell
Hoad Cell
laval Sieva Froth Regularity
A medium viscosity froth (category B ) was generally Run Surfactant
[pphp] Package*
Pressure
Quality Count poor-good
P»rl [1-5]
found to be the ideal froth condition for providing the [cm'1]

optimum foam cell structure. #1 B 8050 1.5 30 mm 1x1% 15 02 10 4

However, froth quality does not necessarily correlate to #2 B 8250 1.5 30 mm - 1x 1% 15 B/1 9 3
the cell size characteristic o f the resulting foam. In some
#3 B 8240 1,5 30 mm - 1x1% 15 B/1 9 3
cases, a stiff viscous froth caused by an initial excellent
nucleation can result in a coarse cell structure. This occurs #4 Comp. Surf. 1,5 30 mm- 1x 1% 15 C-D/2 8 3
when the original fine cells coalesce so quickly that the #5 B 8250 1,5 40 mm 2x 2% 6 D/2 11 5
mechanical froth breaks down immediately into a coarse
#6 B 8240 1,5 40 mm -2x 2% 6 D/2 11 5
structure.
Besides the influence o f nitrogen levels, surfactant type #7 B 8250 1,0 40 mm -2x 2% 5,5 D/2 10 5
also affects froth quality. Table 2 shows a comparison o f #8 B 8240 1,0 40 mm 2x 2% 5,5 D/2 9 5
several surfactants, including the best candidates identified
in the initial benchmarking trial (figure 5) under varying
machine conditions. The follow ing formulation and
Stability Parameters
conditions were used to stress the surfactants for nucleation
and froth development:
B esides achieving a fine cell structure, optimizing
• 4.5 pph water physical foam parameters like gas yield, density and
• 3.8 pphp C 0 2 hardness gradients across the foam bun are as important in
• a blend o f two all PO polyols liquid C 0 2 processes as they are in conventional foaming.
• a TDI index o f 115 Goldschmidt offers a full range o f surfactants to meet
• no added nitrogen for nucleation these various requirements (Figure 7). Surfactants which
Two different surfactant concentrations were used with provide a medium efficiency and a very pronounced wide
two different creamer setups; one using a single sieve with processing latitude are available to foamers who desire a
a smaller diameter, and the other using two sieves at a product for the lower density C 0 2 blown grades, as w ell as
larger diameter. The tw o - sieve setup was used to promote the higher density conventional foam grades. A lso,
nucleation by increasing mechanical shear. universal and even true FR surfactants are available to

420 / Eyrisch, Burkhart, Borgogelli

foamers who prefer to make both their flame retarded and gatebar machine in a back-to-back trial with TEGOSTAB®
non - FR foam grades with the same liquid C 0 2 surfactant. B 8240, using two different formulations, both containing
The follow ing section reports results from industrial graft polyol. Table 3 summarizes the results o f this plant
trials comparing the newly developed surfactants, trial, which show that TEGOSTAB® B 8250 produced
TEGOSTAB® B 8250 and TEGOSTAB® B 8050, to other foams with a superior cell structure, and also offered the
benefit o f improved block yield. The higher efficiency o f
TEGOSTAB® B 8250 resulted in foams with a lower
average density and airflow, as w ell as narrower top-to-
bottom distributions.
TEGOSTAB® B 8050 is a hydrolyzable (SiOC type)
conventional surfactant designed for optimized processing
latitude. It was evaluated on a Novaflex® machine in a
back-to-back trial with a competitive higher efficiency
surfactant. Table 4 summarizes the results o f this plant
trial, which show that TEGOSTAB® B 8050 produced
foam with a finer cell structure. Due to the wider
processing latitude, TEGOSTAB® B 8050 yielded a more
open - celled foam, resulting in a higher average density at
similar use levels. The higher processing latitude offers the
opportunity to increase surfactant and/or tin levels without
tightening the foam. With consecutive increases in both
Figure 7. Efficiency and processing iatitude of silicone surfactant and tin, foam was produced that was still more
surfactants suitable for the Liquid-COr-Technology open than the standard quality, but also exhibited improved
density and airflow distribution. D ue to the higher
table 3. Industrial trial on CarDio® applicable tin levels, foams with higher hardness could also

Water
r~ 4,5A
I
B
4,3
be obtained.
This example shows that surfactants with lower
Formulation C02 3,8 3,5 efficiency can provide desired quality improvements if it is
[pphp] Index 112 114 combined with an outstanding processing latitude.
TEGOSTAB B 8240 1.76 - 1,65
table 4. Industrial trial on Novaflex®
TEGOSTAB B 8250 - 1,75 - 1,65
+ ++ Water 4,8
Cell Structure + ++
Formulation C02 4,7
Foam Height +1inch +0.5inch [pphp] Tin 0,19 0,19 0,19 0,21
Top 0 ,9 4 0 ,9 3 1,09 1,09 Competitive Surf. 1,6
^ ^ Middle 0 ,9 5 0 ,9 3 1,11 1,10 TEGOSTAB B 8050 1.6 1,8 1,8
Density [pcf]
Bottom 0 ,9 8 0 ,9 6 1,14 1,13 Cell Structure + ++ ++ ++
Average 0,96 0,94 1,12 1,11 Top 13,6 13,9 13,9 14,0

max. Variation 4% 3% 5% 4% . Middle 14,1 14,5 14,4 14,5

Density [kg/mn
Top 8 ,0 4 ,5 5 ,6 3 ,6 Bottom 14,4 14,9 14,7 14,6

_ Middle 5,3 5,7 4 ,4 Average 14,0 14,4 14.3 14,4

Airflow [scfm] 5,7
Bottom | 4 ,5 5,4 4 ,5 max. Variation 6% 7% 6% 4%
5,7
Average | 6,5 4,8 5,6 4,2 Top 2,0 1,0 1,0 2,0

max. Variation 40% 18% 6% 2% „ Middle 3,5 1,5 1,5 2,0

Airflow [mm H20]
Top 3 2 ,3
Bottom 4 ,0 2,0 1,5 2,0
3 0 ,0 2 8 ,8 3 0 ,3
Average 3,2 1,5 1.3 2,0
„ ^„ Middle 33,2 3 2 ,3 30,1 3 0 ,8
25% IFD [lbs] max. Variation AZ A1 AO,5 A0
Bottom 3 4 ,3 3 3 ,3 3 0 ,9 3 0 ,0
Top 1,9 1,9 1,9 2,0
Average 33,3 31,9 29,9 30,4
2,2 2,1 2,0 2,3
max. Variation 6% 11% 7% 3% CLD 40% [kPa] Middl®
Bottom 2,2 2,2 2,1 2,3
Average 2,1 2,1 2.0 2,2
TEGOSTAB® B 8250 is a non - hydrolyzable (SiC type) max. Variation 16% 16% 11% 15%
universal surfactant designed with an outstanding
efficiency profile. It was evaluated on a CarDio® Airless

Eyrisch, Burkhart, Borgogelli / 421

Although not presented in this paper, industrial trials using Foaming Mixtures ” , Polyurethanes Conference 2000,
the Beamech CO-2™ process were also conducted, again October 8 -1 1 ,2 0 0 0
confirming the superior performance o f both surfactants, 6. Kanner, B. and T.G. Decker. 1969. “ Urethane foam
TEGOSTAB® B 8250 and TEGOSTAB® B 8050. formation role o f the silicone surfactant ”, Journal o f
Cellular Plastics, January.
CONCLUSION

Two new surfactants optimized for the Liquid C 0 2 BIOGRAPHIES

technology - TEGOSTAB® B 8250 and TEGOSTAB® B
8050 - were developed according to a protocol that defines Oliver Eyrisch
cell structure as a consequence o f optimized nucleation,
frothing support and coalescence prevention. Under Dr. Oliver Eyrisch, bom in
practical pilot-plant conditions these two candidates were 1966, studied Chemistry at the
found to give an optimum state-of-the-art cell structure University o f Freiburg. After
following an experimentally developed froth categorization receiving his Ph.D. in Bioorganic
scheme. The superior performance was proven under Chemistry he joined the Hoechst
industrial conditions using all three different machine AG in 1994, later Clariant AG,
technologies that are currently in use. working as a R&D -manager in
Both surfactants exhibit an improved nucleation profile the field o f cationic surfactants
but differ according to their stabilization characteristics in for detergent and cosmetic
order to meet the different requirements o f foamers. applications. In 1998 he joined
Accordingly, the PU - labs o f Th. Goldschmidt AG. Today he is
responsible for the development and technical service o f
TEGOSTAB® B 8250 polyurethane additives to the flexible slabstock and molded
• is non hydrolyzable (SiC technology), foam industry in Europe.
• universal (also suitable for FR grades) and
• highly efficient (higher foam yields achievable).
Rob Borgogelli

Rob Borgogelli is a Senior

TEGOSTAB® B 8050
Technical Service Representative
• is hydrolyzable (SiOC technology),
for Goldschmidt Chemical
• bears a wide processing latitude (also suitable for
Corporation in the Hopewell, VA.
high density conventional foaming),
After graduating from Queen’s
• minimizes top - to bottom variations and University in Kingston Ontario
• allows for a fine tuned hardness design. Canada with a Bachelor’s degree
in Chemical Engineering, he
REFERENCES joined W oodbridge Foam
Corporation in 1985, gaining
1. Dounis, D.V. and B.L. Hilker. “The Influence o f experience in flexible slabstock
Silicone Surfactants on Frothing Using D issolved foam production and product development. He began his
Carbon D ioxide ” , Polyurethanes EXPO *98 career with Goldschmidt Chemical in 1994, and is
2. Mell6, A., M.- C. Desnier, B.L. Hilker, L.F. Lawler and responsible for technical support o f polyurethane additives
S.B. M cVey. “N ew Silicone Stabilizers for Liquid to the flexible slabstock industry in North America.
Carbon D ioxide Blow n Flexible Foam ” ,
Polyurethanes World Congress '97, September 29 Georg Burkhart
October 1, 1997.
3. Ridgway, D.H., J.G. Kniss, L.A. Mercando and M. Dr. Georg Burkhart joined Th.
Huhtasaari. “N ew Surfactants for U se in C 0 2 Goldschmidt AG in 1981 after
Technology ” , Polyurethanes EXPO '99, September receiving his Ph.D. in chemistry
12 - 15, 1999. at the University o f Bochum,
4. Burkhart, G. and A. Weier. “Liquid Carbon Dioxide working mainly at the Max -
Foaming in Flexible Slabstock Foam Production - A Planck-Institut o f Coal research.
Challenge to be Met with N ew Structural Building He has been involved in both
Blocks ” , Journal o f Cellular Plastics, Vol. 35, polyurethane research and
July/August 1999 technical service, and since 1986
5. Koot, W. and W.P.M. Maas. ’’Physical Properties o f he is the leader o f Goldschmidt
Flexible Polyurethane “ Liquid C 0 2” Slabstock A G ’s polyurethane laboratories.

422 / Eyrisch, Burkhart, Borgogelli

DBTDL Replacement in High Resilience Slabstock Foams

E . - J. GERARD, H. VERSTRAETE B. SCHLENTER

a n d W. MAAS Shell Chemicals
Shell Chemicals P.O. Box 8610
Monnet Centre International Laboratory 3009 AP Rotterdam
Avenue Jean Monnet 1 The Netherlands
B-1348 Louvain-La-Neuve
Belgium

ABSTRACT INTRODUCTION

Over the last two years, serious concerns have been For several years, the polyurethane industry has
expressed in the press about human exposures to trace critically examined its products to meet present and
amounts o f alkyl tin moieties such as tributyltin (TBT). possible future environmental regulations. It is clearly
Organostannic compounds are commonly used as the role o f each major polyurethane feedstock supplier,
catalysts in polyurethane chemistry. Whereas stannous together with their customers, to remain aware o f all
octoate (SNOCT) exhibits a TBT content below the potential health and environment hazards associated with
detection limit o f most analysis equipment (< 0.01%), the use o f its products to ensure its long term
TBT levels as high as 0.6% have been measured in sustainability.
dibutyltin dilaurate (DBTDL). Although DBTDL As a result o f the recent publications related to the
manufacturers are currently optimising their process to potential hazards that TBT could have on human health
lower the TBT content in DBTDL, the (chemo)pressure [1], Shell Chemicals have embarked on a research
exerted on polyurethane foamers is such that DBTDL - programme the aim o f which was to present the foamer
free formulations are being actively looked for. with HR formulation guidelines that do not use DBTDL
High resilience polyurethane foams based on various as the gel catalyst.
polymer polyols - particularly polyols made with pure As a consequence, POSTech-based HR formulations
polystyrene (POSTech*) or poly(styrene/acrylonitrile) - that had been traditionally developed using DBTDL had
have been developed using DBTDL as the gel catalyst. to be adapted to comply wit a TBT -free criteria in line
The direct replacement o f DBTDL is not straightforward with the wishes o f both the foamers and the public.
as foam block dimensional stability issues (often referred An obvious step toward the elimination o f DBTDL from
to as “cold flow ”) are encountered when other catalyst HR polyurethane foam formulations consists in the
(i.e. SNOCT) are used. replacement o f the latter by another tin catalyst such as
This “cold flow ” phenomenon is very much dependent SNOCT. However, such a straightforward replacement
on the size and the weight o f the foam block produced, without reformulation efforts results in a block stability
making all small scale laboratory experiments unsuitable problem (cold flow, bulging, ballooning...).
for developing novel stable HR foam formulations. The present paper is first concerned with the definition
The present research has shown that properties derived o f a series o f laboratory tests to assess the “cold
from small scale experiments can nevertheless be related flowability ” o f a given HR formulation. Factors which
to the “cold flowability ” o f HR formulations. Amongst influence the foam block stability have been investigated
the various tests investigated, the variation in sag factor in the light o f these tests and are presented in the second
from top to bottom o f the foam block was found to be part o f the paper.
the most relevant test method to assess the dimensional
stability. :Sag factor is here defined as the force exerted COLD FLOW PHENOMENON
by the foam compressed to 65% o f the original thickness
divided by that exerted at 25% compression during the HR slabstock foams find interest in upholstered
compression load deflection (CLD) test. furniture and mattresses where quality consistency is a
The test method based on this variation in sag factor key objective. As a result, any heterogeneity in a foam
through the foam block (vSag) made possible the block severely impairs its use. Block stability issues such
laboratory examination o f formulation parameters for as densification, settling or shrinkage are related to block
their impact on foam block stability. It also allowed us to heterogeneity as they induce property changes across the
develop a wide density range o f POSTech -based foam block.
DBTDL -free HR slabstock foams that are dimensionally The block instability o f concern here is o f another
stable. nature as it affects not only the foam block homogeneity

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 423

 but also the dimensions (and the visual aspect) of the commercial size (typically 2 meters wide by 0.8-1.2
foam resulting in an unacceptable foam waste meter high) are needed to make it occur. Consequently,
generation. Although it may take place in the foam block our primary research effort was directed towards the
already after gelation, its occurrence becomes visible a definition of a test to assess the “cold flowability” of any
few hours after production. Hence the name, cold flow, HR formulation that does not require the use of a
as one notices the instability long after the foaming commercial continuous production line.
reaction exotherm has passed its maximum. It is also
sometimes referred to as “foam bulging” or
“ballooning”.
The following example illustrates the extent to which a
block shape can be affected by the cold flow
phenomenon. Both HR foams were produced using the
same formulation (Table 1) with the exception of the gel
catalyst which was either SNOCT or DBTDL. In both
cases, the foam density targeted was 40 kg/m3.

Table 1. HR formulations for Block Comparison.

C o m p o n e n ts p hp p hp

C A R A D O L M D 3 0 -0 2 100 100
W ATER 2 .3 2 .3
DEoA 1 .2 1 .2
Dabco X 2 9 0 0.1 0.1
D a b c o T 9 (S N O C T ) 0 .2 5 -

D a b c o T 1 2 (D B T D L ) - 0 .1 7 Figure 2. SNOCT-catalysed HR Slabstock Foam.

O si L 21 0 0 0 .3 0 .3
C A R A D A T E 8 0 , In d ex 105 105 Basic Considerations and Assumptions

Comparison of Figure 1 and Figure 2 clearly illustrates It is a common understanding within polyurethane
the impact of the cold flow phenomenon on the block circles that the dimensional stability of a foam slab in the
foam dimensions. It also shows that simple replacement several hours that follow its production depends on
of DBTDL by SNOCT does not constitute a valid option numerous factors.
to develop TBT-ffee HR foams. •The one that first comes to mind is the polymer
strength. The latter is determined by the type and extent
of the chemical reactions that have occurred during the
foaming reaction. In brief, the polymer strength is
influenced by the isocyanate conversion, the cross-link
density, the height of the exotherm peak and the
presence of urea-type hard domains or fillers.
•A second factor, the cell-opening event has been
reviewed in the literature [2]. It has been shown that the
opening of the cells occurs to a large extent as the foam
reaches full rise. At that moment (i.e. extent of reaction),
the polymer matrix would be at gel point. This means
that the system would have developed a very high
viscosity with a low level of elasticity. The high
viscosity would prevent the matrix from flowing fast
enough to expand and relieve the still-increasing gas
pressure. Under such conditions, the cell membranes
burst, giving birth to an interconnected “open” cell strut
network. Dimensional stability problems arise either
when a large population of foam cells remains intact at
Figurel. DBTDL-Catalysed HR Slabstock Foam.
full isocyanate conversion (shrinkage) or when the struts
have not attained sufficient strength at the time of cell
TEST DEFINITION
window rupture (splits, collapse). In both cases, tuning
of the catalyst concentrations help remedy these
Unfortunately for those whose aim is to develop novel
problems.
HR formulations, the cold flow phenomenon cannot be
observed in the laboratory as large foam slabs of

424 / Gerard, Verstraete, Maas, Schlenter

 •The plasticisation of the polymer matrix has also an 0)
effect on the foam block stability as it reduces the
polymer strength (i.e. lowering of the viscosity before (2)
gel point and lowering of the elasticity thereafter).
Ancillary chemicals (like fire retardents) used in where 5 is the phase angle, i.e. the angular
polyurethane foam manufacturing can act as plasticisers. displacement between stress and strain. Two dynamic
One may also place in this plasticisation heading the moduli are commonly defined, G’ and G” such that:
presence in the polyurethane gel of large
(3)
macromolecules that have not reached full conversion.
These large “dangling” chains / blocks, although
chemically linked to the rest of the polyurethane where G’ is in phase with the deformation and G” is
network, remain present in the foam block until full n/2 out of phase with the deformation. G’ is called the
conversion. Due to their size and their lack of mobility, elastic modulus and G” the viscous modulus. Tan 5 is
their reactivity becomes wholly controlled by diffusion referred to as the loss factor, i.e. it shows a minimum
phenomena. Both their size and the diffusion limitation when nearly all the energy stored in deforming the
increase whilst approaching full isocyanate conversion. sample is recovered when the stress is removed. It shows
One may state that upon establishing their last chemical a maximum when most of the energy needed to deform
links, the “dangling” blocks can inherently induce in-situ the sample has been converted into heat
a stress that the foam slab cannot accommodate due to its (macromolecular motions), hence its name.In this study,
high degree of cross-linking and subsequent lack of DMS spectra have been measured throughout the
freedom. isocyanate conversion on the three different systems
Quantifying the extent to which the factors described displayed in Table 2.
above influence the block stability of a given HR
formulation remains complex and fastidious. On the Table 2. Formulations for DMS analysis.
C o m p o n e n ts p hp p hp p hp
other hand, one may state that the simple exchange of
DBTDL by SNOCT cannot induce fundamental changes
C A R A D O L M D 3 0 -0 2 100 100 100
in those factors. Therefore, the approach we chose was DEoA 1 .2 1 .2 1 .2
more pragmatic. The major assumption made in this D abco T 9 (S N O C T ) 0 .2 2 - 0 .2 2
work was that the cold flow induced dimensional D a b c o T 1 2 (D B T D L ) - 0 .2 2 -

instability of foam slabs was the extreme manifestation G o ld s c h m id t O rte g o l 2 0 4 - - 2

of the phenomenon. This assertion implies that, although O si L 2 1 0 0 0 .3 0 .3 0 .3
not always visible, cold flow : C A R A D A T E 8 0 , In d e x 100 100 100
• may have taken place in foams visually acceptable
• can also occur on a smaller scale One should note that these formulations correspond to
• should have an influence on some foam properties polyurethane elastomer formulations. In fact, due to the
change in volume, monitoring the change of rheological
In fact, we tried to establish a correlation between properties during foaming remains a difficult task.
different laboratory tests and facts arising from the It has been postulated [3] that the gel point can be
market. In particular, it was known that the catalyst defined as the time (or the conversion) at which a cross-
combination amine / SNOCT gave the worst results in linking system shows a viscous contribution that equals
terms of block instability induced by cold flow, whereas the elastic one (G” = G’). Before the gel point, the
the combination amine / DBTDL would yield the best system is a viscoelastic liquid with G” larger than G \
stability. Between these two extremes, some Beyond the gel point G’ is greater than G”. Figure 3
improvements had been achieved by using urea in compares the results obtained with the three catalytic
combination with SNOCT. Furthermore, additives such systems investigated.
as cross-linkers or Goldschmidt “Ortegol 204” had also
been put forward as block stabilisers.

Dynamic Mechanical Spectroscopy

The first laboratory test studied, the dynamic

mechanical spectroscopy (DMS), measures the ability of
a viscoelastic material to store and dissipate mechanical
energy. An oscillating sinusoidal load (stress) is applied
to a specimen at a particular frequency which induces a
strain (deformation) in the specimen. For a viscoelastic
material, the stress a and strain e vary in a similar
sinusoidal manner, but they are out of phase: Figure 3. Gel Point (G’ = G”) versus catalytic systems.

Gerard, Verstraete, Maas, Schlenter / 425

 Figure 3 shows that SNOCT yields a faster system than
DBTDL (i.e. approximately 100 seconds versus 1000 to (5)
reach gel point). It is also quite remarkable that one of
the effects of “Ortegol 204” is to delay the gel point. After integration and substitution the following
From a purely rheological viewpoint, “Ortegol 204” expression can be derived:
makes a SNOCT-based system resemble a DBTDL-
based one. This effect is furthermore illustrated in Figure
4 which shows the pace at which the polyurethane (6)
chemical network develops elastic strength at the gel
point. This strength development property is defined as
dG’/dt. Figures 5 and 6 refer to the SNOCT- and DBTDL-
based formulations, respectively. They show both the
isocyanate conversion as well as the “linearity” factor
(1/1 -p) as a function of the time.

F igure 4. Strength increase (dGVdt) versu s catalytic

system s.
F igure 5. Kinetics d erived from S N O C T - b ased
Again, it is clear that SNOCT yields a system whose s y s te m s .
cross-linking at the gel point is much faster than that of
DBTDL. Interestingly, Figure 4 corroborates the
“slowing down” action of “Ortegol 204”. The use of the
latter, when combined with SNOCT, results in cross-
linking system that “smoothly” develops in the vicinity
of the gel point.

Fourier Transform Infrared Spectroscopy

The second test used, the FTIR, is a useful tool that is

commonly used in polyurethane chemistry to detect and
monitor the evolution of characteristic peaks. The
chemical reactions of the formulations based on SNOCT
and DBTDL displayed in Table 2 have also been
F igure 6. Kinetics derived from D BTD L - based s y s te m s .
followed by means of FTIR. The disappearance of the
isocyanate peak in the 2250-2280 cm*1 wave number Figures 5 and 6 corroborate the fact that SNOCT
region has been monitored throughout the conversion. imparts more reactivity to the polyurethane system than
Although the complete description of the polyurethane DBTDL. Interestingly, one can see that the SNOCT-
chemistry is rather complex [4], one can to a first based system encounters very rapidly the onset of
approximation accept an overall second-order stepwise diffusion limitation, i.e. the point at which the plot of
polymerisation reaction. The rate of disappearance of the 1/(1-p) versus the time deviates from linearity. This is
isocyanate functions can thus be written as: not the case for DBTDL. One may argue that this non­
linearity occurs after approximately 80% conversion but
if one considers the time instead of the conversion one
can postulate that the diffusion limitation occurs faster
with SNOCT than with DBTDL. This also implies that it
If p represents the fraction of NCO groups that have occurs when the foam is at a higher temperature with
reacted at time t i.e.: SNOCT than with DBTDL. As a result, one may state
that the combination of these two factors (faster onset of

426 / Gerard, Verstraete, Maas, Schlenter

mass transfer limitation occurring in a warmer foam) can variation in sag factor (vSag) was correlated best with
possibly explain why SNOCT-catalysed HR slabstock visual observations o f “cold flow ” . The sag factor
formulations are more sensitive to cold flow than variation defined in this study was based on compression
DBTDL -based ones. A higher temperature during the hardness measurements (CLD). One should note that
last stages o f foam cure would allow dangling blocks to similar results are obtained when indentation hardness
“flow ” and react, hence inducing stress in the foam and measurements (ILD) are performed. The sag (or
subsequent block deformation. comfort) factor is commonly defined as:

Sag Factor Variation (vSag)

(7)
In order to define a third laboratory test differing from
simple visual inspection o f the foam block shape, a
series o f HR formulations has been produced with a The variation in Sag factor across the foam is given by:
commercial continuous slabstock machine. Two
densities were aimed at, namely 28 and 40 kg/m3 (Table ( 8)
3 and 4, respectively).

Table 3. HR formulations for 28 density. Note that for the determination o f vSag, two
C o m p o n e n ts 2 8 -1 2 8 -2 2 8 -3 centimetres have been cut away from the foam at the top
and the bottom to avoid the influence o f the skin effect.
C A R A D O L M D 3 0 -0 2 60 4 7 .3 4 7 .3 Figure 7 illustrates the variation in sag factor measured
C A R A D O L M C 2 7 -0 3 40 5 2 .7 5 2 .7 on the foams derived from the formulations displayed in
C A R A D O L S A 3 6 -0 2 1 1 .5 2 .5 Tables 4 and 5.
S o lid s C o n te n t, % 1 0 .5 8 .3 8 .3
W a te r 3 .3 5 3 .3 5 3 .3 5
DEoA 1 .6 1 .6 1 .6
A m in e C a ta ly s t 0 .1 5 - -

U re a - 0 .1 5 0 .1 5
D abco T 9 (S N O C T ) - 0 .2 5 0 .2 5
D a b c o T 1 2 (D B T D L ) 0 .2 - -

O si L 2 1 0 0 0 .2 2 0 .2 2 0 .2 2
C A R A D A T E 8 0 , In d e x 106 106 106

Table 4. HR formulations for 40 density.

C o m p o n e n ts 4 0 -1 4 0 -2 4 0 -3

C A R A D O L M D 3 0 -0 2 100 100 100

S o lid s C o n t e n t ,% 1 7 .5 1 7 .5 1 7 .5 Figure 7. VSag as a function of HR formulations.
W a te r 2 .3 5 2 .3 5 2 .3 5
DEoA 0 .8 1 .0 1 .2 Meter Cube Experiments
A m in e C a ta ly s t 0 .3 - -

U rea - 0 .2 0 .2
Assuming that this vSag - cold flowability relation is
D a b c o T 9 (S N O C T ) - 0 .2 5 0 .2 5
D a b c o T 1 2 (D B T D L )
valid regardless o f the size o f foam block produced, one
0 .1 6 - -

O si L 2 1 0 0 0 .3 0 .3 0 .3
may be able to use it even when the foam block does not
C A R A D A T E 8 0 , In d e x 105 105 107 show dimensional instability.

Upon visual inspection o f the foams some 24 hours Table 5. HR formulations for the meter cube scale
after production, the following “stability ranking” could C o m p o n e n ts 5 5 -1 5 5 -2 5 5 -3 55 4

be done (from best to worse):

C A R A D O L M D 2 5 -0 3 50 50 50 50
28-1 > 2 8 -2 > 28-3
C A R A D O L M D 3 6 -1 3 50 50 50 50
40-1 > 40-2 > 40-3
S o lid s C o n te n t, % 14 14 14 14
Again, the DBTDL-based formulations (28-1 and 40-1) W a te r 1 .6 1 .6 1 .6 1 .6
did not show dimensional instability. This instability DEoA 1 1 1 0 .5
could be seen for all other formulations, in particular in Dabco X 290 0 .1 8 0 .1 8 _ _
the case o f the 40 density foams. All the blocks produced U rea - - 0 .2 2 0 .2 2
have been thoroughly characterised from top to bottom D a b c o T 9 (S N O C T ) - 0 .2 5 0 .2 5 0 .2 5
by taking a series o f samples across the foam slab D a b c o T 1 2 (D B T D L ) 0 .2 - - _

O rte g o l 2 0 4 - - - 1 .5
thickness. From all the series o f tests carried out
O si L 21 0 0 0 .4 0 .4 0 .4 0 .4
(density, hardness, porosity, tensile...), we found that the
C A R A D A T E 8 0 , In d e x 105 105 105 105

Gerard, Verstraete, Maas, Schlenter / 427

 Table 5 displays the HR formulations run in our FORMULATION PARAM ETERS INFLUENCING
laboratory meter cube discontinuous production THE COLD FLOW ABILITY
machine. The density targeted was 55 kg/m3. Figure 8
shows the variation in sag factor measured for the In the light of the results described above, a series of
various formulations. It is remarkable to see that, runs has been conducted using our laboratory conveying
although all the blocks produced were apparently stable, equipment during which various formulation parameters
a significant difference in sag factor variation was have been varied. It has to be noted that the size of the
measured. slabs produced using this equipment remains limited,
typically to 1 meter wide and 0.8 meter high. As a
consequence, the visual dimensional block instability
could not be seen. The following formulation parameters
have been investigated for their effect on the sag factor
variation, thus on the cold flowability of the resulting
foam blocks:
• The SNOCT concentration
• The DEoA level
• The use of urea as the amine / blow catalyst
• The use of Ortegol 204 as the foam “stabiliser”
• The solids content
• The reactivity of the diluent polyol

The corresponding formulations are too numerous to

Figure 8. VSag as a function of HR fonnulations be presented here. Both CARADOL MD30-02 and
CARADOL MD25-03 were used as the polymer polyols
It can also be seen that, as observed during the and the T-80 index was kept constant at 105. The
continuous commercial production trials, the use of urea targeted density was 45 kg/m3. Figure 10 compares the
as the amine catalyst tends to limit the sag factor influence of the gel catalyst (SNOCT or DBTDL) and
variation. This is also true for “Ortegol 204”. Another the DEoA level on the variation of sag factor.
test was performed during these meter cube experiments.
A piece of hard foam was placed at the bottom of the
meter cube tool before the formulations were cast into it.
After de-moulding of the foam block (some 5 minutes
later), the rigid foam insert was removed. Then the depth
of the cavity created was measured 24 hours after
production. Figure 9 shows the depth of this cavity for
the four formulations produced.

Figure 10: VSag versus gel catalyst and DEoA level.

The better results obtained with DBTDL are

confirmed, regardless of the DEoA level. The influence
of the latter on vSag does exist but seems to be
somewhat more complex. We believe this is due to the
double role of DEoA in an HR foam: that of a cross
Figure 9: Cavity depth versus HR fonnulations. linker associated with a “weakening” action on the hard
domains.
Figure 9 tends to confirm the results obtained with the
Figure 11 depicts the influence of the SNOCT
variation in sag factor, i.e. DBTDL-based formulations
concentration on vSag for both blow catalysts X290 and
withstand their weight whereas SNOCT-based ones
urea. It can be seen that lowering the SNOCT
show some sign of settling down. However, the use of
concentration from 0.25 to 0.15 php reduces vSag. This
both urea and “Ortegol 204” reduces this phenomenon.
is observed for both the X290 and urea blow catalysts.
However, best (lowest) vSag values are consistently
obtained using urea as blow catalyst.

428 / Gerard, Verstraete, Maas, Schlenter

 Figure 11. Vsag versus blow catalyst and SNOCT level. Figure 13. VSag versus solids content and SNOCT.

Figure 12 displays the influence o f both Ortegol 204

and DEoA on vSag.

Figure 14. VSag versus polyol reactivity and Ortegol 20

level..

Figure 12. Vsag versus Ortegol 204 and DEoA levels. Upon looking at Figure 14, it becomes clear that the
use o f Ortegol 204 permits the attainment o f a very low
First, one can notice that the levels o f vSag attained level o f sag factor variation. In the system studied, best
with Ortegol 204 are much lower than those obtained so results are obtained with low levels (1 php) o f Ortegol
far using SNOCT. Second, the influence o f DEoA is 204. Furthermore, no significant change in vSag is
again visible but limited. Figure 13 displays the observed through the use o f a higher reactivity polyol
dependence o f vSag upon the solids content. One must phase.
add here that lowering the solids content requires the use Overall, it can therefore be seen that the variation in
o f a larger amount o f diluent polyol (CARADOL MD36 - sag factor within the foam block can be significantly
13). This implies that two variables are changed: the reduced by tuning the various formulation parameters.
solids content and the polyol reactivity. It is confirmed Accepting that this variation in sag factor is related the
that, regardless o f the solids content, lowering the foam slab stability, the results obtained so far allow
SNOCT level reduces vSag. Figure 13 also shows that SNOCT-based HR systems to compete with DBTDL-
reducing the solids content and / or increasing the based ones.
reactivity o f the polyol matrix induces a smaller
variation o f vSag. To assess the effect o f the only polyol CONCLUSIONS
reactivity in isolation on vSag, experiments have been
carried out using polymer polyols differing in their solids Research has been carried out to develop DBTDL -free
content (CARADOL M D30 - 02 (17.5% solids) and HR polyurethane foams that has led to the following
CARADOL MD25 -03 (28% solids)). This implies that, conclusions:
in the case o f CARADOL M D25 - 03 more diluent polyol • The variation in sag factor through the foam is the best
is used to achieve a similar formulation solids content. indicator to assess dimensional instability (cold flow) o f
Figure 14 shows how vSag is influenced by both the a block foam on the laboratory scale.
concentration o f Ortegol 204 and the reactivity o f the • This test can be applied even when blocks are
polyol phase. produced that do not exhibit visual deformation.
• Best results are obtained through lowering the SNOCT
level (0.15 php) combined with the use o f urea as the
blow catalyst while keeping the HR formulation solids

GerardVerstraete, Maas, Schlenter / 429

 content as low as possible. The use of Ortegol 204
undoubtedly helps achieve excellent block stability. No Hilde Verstraete
clear correlation could be established between the sag
factor variation and the DEoA concentration or the Hilde Verstraete joined Shell
reactivity of the polyol phase. Chemicals in 1994 after studying
• Combining the formulation parameters described in chemistry at Leuven, Belgium. Since
this paper allows the achievement of sag factor then she has been a member of the
variations that can compete with those measured on flexible polyurethane group and is
DBTDL-catalysed HR foams. currently working on research and
development aspects and technical
The research results gathered so far give us confidence service in flexible slabstock
that DBTDL-free HR formulations can be produced in a applications.
commercial continuous mode across a wide density
range without suffering from the cold flow phenomenon.
This will be the last stage of this development route. Wilfried Maas

Wilfried Maas is currently

ACKNOWLEDGEMENTS department manager of the
polyurethane slabstock foam
The authors would like to thank Gary Biebaut and product application R&D and TS
Christelle Bar from the Monnet Centre analytical group team in Shell Chemicals. Wilfried
for their valuable help and fruitful discussions. They also joined the company in 1989 after
would like to acknowledge their urethane colleagues earning a Ph.D. in gas phase ion
Allan Batt, Chris Dam, Marc Esselens, Johan Clerinx chemistry. Prior to work in the foam
and Laurence van Herzeele for their contribution to this product application area he worked
work. in process development, process design and plant
technology.
REFERENCES
Bernadette Schlenter
1. “Jotun launches tin-free antifouling system”, PCI
July 2000, Vol.XVI-7:80. One could also consult the Bernadette Schlenter is a Market
following sites: Development Manager for
http://www.atrp.com/tbt.htm Urethanes in Shell Chemicals. She
http://www.marisec.org/icskeyissues2000/tbts.htm studied Chemistry at the University
http://www.greenpeace.org/-toxics/reports/tbtfact of Utrecht and obtained a Ph.D. in
sheet.html Technical Sciences from the
2. R.M. Herrington and R.B. Turner, “The formation, University of Eindhoven (The
cell-opening and resultant morphology of flexible Netherlands). She joined Shell in
polyurethane foams”, Advance in Urethane Science 1985 and held several positions in the area of equipment
and Technology, Vol. 11:1-67. engineering research, plant technology and flexible and
3. F. Chambon et al. 1986, Macromolecules, 19:2146. rigid urethane product application research. She has been
4. M. Sato 1962, J. Org. Chem., 27:819. active in urethane market development since 1997.

BIOGRAPHIES

Eric - Jack G erard

Eric Gerard joined Shell Chemicals

in 1989 and is currently member of
the flexible polyurethane group
working on research and
development aspects in flexible
slabstock applications.

430 / Gerard, Verstraete, Maas, Schlenter

Catalysts and Silicone Surfactants for Reduced VOC Emissions
of Polyester Slabstock Foam

MATTI S. HUHTASAARI, RENO PLAUMANN JANE G. KNISS a n d FRANK D. WOMACK

a n d JO B ST GRIMMINGER Air Products and Chemicals, Inc.
Chemicals Division Europe 7201 Hamilton Boulevard
c/o Air Products GmbH Allentown, Pennsylvania 18195
Robert Koch Strasse 27
22851 Norderstedt
Germany

ABSTRACT Silicone surfactants can also contribute to VOC and

FOG emissions from foam. This paper also reports on
Polyester-based polyurethane slabstock foams are recent work to develop new silicone surfactants for
mainly used in the automotive industry in the form o f polyester foam applications, which result in reduced
textile laminates. In recent years the general awareness emissions.
o f environmental, health and safety issues has grown This paper will present a total polyurethane additive
rapidly, resulting in more stringent requirements solution that completely eliminates amine emissions and
regarding the emissions o f volatile compounds from the significantly reduces silicone surfactant emissions, while
foam. Reduction o f such emissions leads to lower improving emulsification and overall foam physical
fogging values and to potentially decreased health risks properties.
during the manufacture and handling o f fresh foams.
The automotive industry has recently placed increased INTRODUCTION
pressure on the foam producers to significantly decrease
emissions o f volatile organic compounds (VOC ’s) and, Polyester foams were the first continuous slabstock
in some cases, to eliminate the amine emissions from foams to be produced commercially. Although polyester
polyurethane foam. based flexible polyurethane has widely replaced
Low fogging polyester polyols and careful selection of polyester in furniture applications, polyester has
antioxidants (BHT free) can subsequently reduce the maintained approximately 15% market share o f all
contribution o f the polyol towards VOC emissions and flexible slabstock polyurethane produced.
also reduce fogging. There have been recent advances Polyester foams possess certain unique physical
in reducing odor and emissions from polyester slabstock properties that give distinct advantages. Some o f these
foam by introducing low odor, low emission catalysts. properties are summarized in Table 1.
However, these intermediate advances may not be
sufficient for future requirements. This paper will Table 1. Important Property Application Relatbnships of
describe brand new innovations that enable the Polyester Foam
elimination o f amine catalyst emissions. These new, Property Application
non -fugitive catalysts chemically bind into the Fine regular cell -Textile industry - need for fine
polyurethane foam matrix rendering them incapable o f structure, free of ceil structure to enable cutting of
pinholes thin sheets (1 mm).
migrating back out o f the foam after the reaction is
-Flame lamination process
complete. requires pinhole free structure to
This paper will compare results o f these new non ­ avoid burn through.
fugitive catalysts with standard industry amine catalysts
Solvent resistance -Gaskets
and the intermediate solutions adopted by some
polyurethane slabstock foam manufacturers in recent Low resilience -Packaging
years. Coarse reticulated -Filters
foam ceils

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 431

 Silicone surfactants and amine catalysts are key pressure and odor, assisted in reducing plant emissions
ingredients that control these important physical and worker exposure to such volatile amines [1].
properties o f polyester foam, especially cell structure. Ultimately, organic emission reductions that approach
The myriad o f performance requirements silicone zero amine emissions can only be achieved by utilizing
surfactants provide include: reactive, low vapor pressure catalysts.
■ Fine, regular cell structure, Therefore, in the early 1990’s, Air Products and
■ Bulk stabilization across a broad range o f foam Chemicals, Inc. initiated a research and development
densities, program for non-fugitive catalysts to address this
■ Reduced density gradient throughout the foam anticipated step change in the polyurethane industry.
block, As a result of this work, in 2000 we introduced non­
■ Pinhole reduction, fugitive catalyst packages for flexible molded
■ Enhanced water solubility for pre-blended applications, Dabco NE1060 and Dabco NE200, and for
“activator” streams (surfactant, catalyst and water flexible polyether slabstock applications, Dabco NE500
solutions, for machines not equipped with separate and Dabco NE600. With consideration for the special
additive streams), requirements of polyester slabstock applications, Dabco
■ And, minimized contribution to total VOC. NE400 was recently developed. As a well-balanced
catalyst, Dabco NE400 is suitable for a wide variety of
Amine catalyst performance requirements include: formulations ranging in density from 22 - 90 Kg/m3
■ Strong blow capabilities for 80% 2,4- toluene (1.4 - 5.6 lb/ft3). Until now, this has not been possible
diisocyanate /20% 2,6 - toluene diisocyanate with a single catalyst as the multi-catalyst package had
(80/20 TDI) formulations where N- to be optimized for every foam density.
ethylmorpholine or N -methylmorpholine are
historically used, Surfactant Development
■ Well balanced blow/gelation characteristics for
formulations based on mixtures o f 80/20 TDI and Continuing the drive for reduced emissions, a series of
65% 2,4- toluene diisocyanate /35% 2,6- toluene surfactants designed specifically for the emerging
diisocyanate (65/35 TDI), polyester market needs were developed. Dabco DC5526
■ Stronger gelation characteristics for formulations was developed to provide a fine, regular, cell structure
based on TDI 65/35, and excellent processing performance. Subsequently,
■ Low odor in combination with low toxicity to Dabco DC4000 was developed for further reduction of
minimize health & safety risks, such as offensive the organic emissions of the foam, to provide fine, open
odors, eye irritability (blue haze effect), etc., celled foam structure and to increase water solubility.
applicable to raw material handling, foam Dabco DC4020, is the latest development in polyester
production, plant fabrication areas, and in the slabstock surfactant technology and combines all
final use, required performance attributes of today’s polyester
■ Zero amine emissions, recently requested by slabstock market:
European automobile manufacturers to minimize ■ Fine, open, regular cell structure,
or even eliminate negative interaction with other ■ Excellent emulsification,
components and to improve interior air quality, ■ Wide processing latitude,
and minimizing passenger chemical exposure, ■ Water solubility,
■ And provide synergy with silicone surfactants for ■ And, emission optimized.
good processing latitude, resulting in smooth rise
profiles, minimal recession, reduced density EXPERIMENTAL WORK
■ gradient, sufficient cell openness, etc. Formulations and Additives

DEVELOPMENT A typical polyurethane foam formulation (see Table 2)

Catalyst Development used in the polyester slabstock industry was selected for
laboratory development work.
Due to the greater reactivity and selectivity o f polyester
polyols towards gelation compared with conventional Table 2. Formulation Used for Laboratory Work
polyether polyols in slabstock applications, catalyst Component Use Level
packages are designed to promote the isocyanate-water
Desmodur® VP PU 60WB01 100.00 parts
or blow reaction. Blow catalysts such as morpholines,
piperazines or benzyl dimethylamine are traditionally Water 3.05 pphp
used for this application. All o f these catalysts exhibit Silicone Surfactant 0.80 pphp
relatively high vapor pressures and consequently can be Catalyst Varied
characterized by a relatively strong, fish -like smell. The Index : 100 TDI 80/20 23.40 pphp
commercial development o f Dabco 2039 in 1991, a 15.60 pphp
moderate blow catalyst possessing a relatively low vapor TDI 65/35
Desmodur® Trade name of Bayer AG

432 / Huhtasaari, Plautnann, Grimminger, Kniss, Womack

 Table 3. Catalysts and Their Vapor Pressures For This Study
Catalyst | Composition I Vapor pressure
Dabco NEM N-Ethylmorpholine (Industry Standard) 5.03 mmHg @ 20°C
Dabco NMM N-Methylmorpholine 13.20 mmHg @ 25°C
Dabco BDMA Benzyldlmethylamine 0.58 mmHg @ 25°C
Dabco 2039 Low odor Catalyst 1.00 mmHg @ 21 °C
Catalyst A 1,4-Dimethylpjperazine/ N.N-Dimethyl-hexadecanamine combination < 5.22 mmHg @ 25 °C
Catalyst B 1,4 -Dlmethylplperazine 5.22 mmHg @ 25 °C
1,4 -Dlmethylpiperazlne / Bis (2 -Dlmethylamlnoethyl) -ether < 5.22 mmHg @ 25 °C
Catalyst C
combination
New non-fugitive catalyst package developed for polyester slabstock < 1.50 mmHg @ 21 °C
Dabco NE400
applications

For all handmix evaluations, rise profiles were

Table 4. Surfactants of This Study recorded using an ultrasonic Rate o f Rise apparatus
Dabco DC5526 First generation polyester from Format Messtechnik. Software capabilities o f this
slabstock surfactant equipment also allowed measurements o f maximum
rise height, recession and shrinkage to be recorded.
Dabco DC4000 BHT-free water soluble polyester
slabstock surfactant Immediately following completion o f the Rate o f Rise
test, foams were cured in an oven at 70 °C for 24 hours
Dabco DC4020 Emission optimized polyester
surfactant before additional evaluation. After oven curing, the
foams were then allowed to cool down to room
Competitive General purpose, first generation
temperature for 2.5 hours. For emission testing, foam
Surfactant 1 polyester silicone surfactant
samples were cut from the bulk foam with a razor
Competitive Low fogging version of knife, wrapped in aluminium foil and enclosed in a
Surfactant 2 Competitive Surfactant 1
polyethylene bag to avoid loss o f volatiles or cross
Competitive Low fogging standard polyester contamination o f the samples during shipment for
Surfactant 3 silicone surfactant testing.
Competitive Low fogging, optimized for highly
Surfactant 4 branched polyol containing Machine Trials
formulations
In addition to the first handmix lab evaluation, the
Two main sets o f evaluations were conducted to following catalysts in Table 5 were used to conduct
compare the performance o f silicone surfactants and frill-scale industrial line trials. Dabco DC4020 was
catalysts. The catalysts listed in Table 3 were used in used as the silicone surfactant in all three runs.
the first set o f handmix evaluations. N -ethylmorpholine For these production line trials we chose a 38 Kg/m3
at 1.70 pphp was chosen as the standard catalyst and density (2.37 pcf) foam formulation based on the low
use level. All o f the other amine catalysts ’ use levels
fogging polyester polyol, Lupraphen® 8148, from
were adjusted to meet the reaction profile o f this
BASF.
industry standard catalyst.
All three catalysts provided acceptable processing
The Silicone Surfactants evaluated and tested in this
performance based on the customer’s broad experience.
study are summarized in Table 4.
Typically, use o f reactive amines in polyester
formulations can cause scorching and/or shrinkage.
Laboratory Handmix Technique
However, the processing performance o f the non­
fugitive catalyst Dabco NE400 was identical with the
Due to the inherently high viscosity o f the polyester
standard fugitive and emission reduced catalyst
polyols traditionally used in polyester slabstock foams
package. N o shrinkage or scorching was observed with
and the relatively low viscosity o f the TDI, laboratory
foams produced with Dabco NE400.
handmix can be difficult. The method below describes
a handmix method that minimizes any potential
Table 5. Use Levels o f Catalyst Machine Trials
inconsistency.
1. Weigh polyester resin, silicone surfactant, amine Amine catalyst | Use level [pphp]
catalyst and water into a 1.2 liter plastic cup. Dabco NMM 1.20
2. Mix for 30 seconds at 4,000 rpms using a round Dabco 2039 / Blow 1.10
mixing blade 6 centimeters in diameter. catalyst combination
3. Add isocyanate. Dabco NE400 0.65
4. Mix for 6 seconds as described in step 2. Rapidly
pour into a 20 x 20 x 15 centimeter plastic or paper box.

Huhtasaari, Plaumann, Grimminger, Kniss, Womack / 433

Emission and Odor Test Description Table 6. Odor Rating Scale for VDA 270B3 Test
Perception of odor Score
The Daimler Chrysler Method PB VWT 709 [2] was
Not noticeable 1.0
used to determine the gaseous condensable emissions
by utilizing a thermo desorption technique. This Noticeable, not annoying 2.0
method represents one o f the most stringent analytical Noticeable, unpleasant 3.0
methods currently employed in the automotive industry Annoying 4.0
to determine organic emissions from plastic materials. Strongly annoying 5.0
This method distinguishes between gaseous emissions
Intolerable 6.0
(VOC) and condensable emissions (FOG).
To determine the VOC, the foam sample was
enclosed in an empty glass tube and heated for 30 A second type o f odor test was conducted to evaluate the
minutes at 90°C. The headspace above the foam was odor propagation according to specifications in GME
swept with a defined amount o f nitrogen and the 60276. A second set o f handmix foams were prepared to
volatile components o f the foam were condensed at - compare three catalysts, Dabco NEM (1.6 pphp),
150°C in a cold trap in the injection system o f a gas piperazine (1.2 pphp), Dabco 2039 (0.7 pphp) and Dabco
chromatograph (GC). These condensed volatiles were NE400 (0.6 pphp). This method required foam samples
then evaporated at 280 ° C, separated by the GC column with 150 cm3 o f volume to be sealed in 3 liter capacity
and identified through mass spectroscopy. preservative jars. These jars were then heated to 70 °C for
Following the VOC analysis, the same foam sample 22 to 24 hours. The samples were then removed from the
was used for the FOG testing. The sample was heated oven and allowed to equilibrate to room temperature for 2
for 60 minutes to 120°C and the volatile components hours. The jars were then opened and the contents
were again cold trapped and analyzed for FOG using assessed for odor by a panel o f six people, utilizing the
GC-MS as described above. odor rating scale shown in Table 7.
Foam samples for all o f the emission testing were
prepared with the aforementioned handmix technique. Table 7. Odor Scale for the GME 60276 Test
Machine trial foams were also evaluated for VOC and Perception of odor Score
FOG in the same manner. Odorless 10.0
These emission tests were conducted at: Just noticeable 9.0
Institut Fresenius Angewandte Festkorperanalytik
Noticeable 8.0
Hauert 9
D - 44227 Dortmund Slight 7.0
Germany Tolerable 6.0
Prior to additional odor testing, the following Borderline tolerable 5.0
European industry standard conditioning o f the foam Objectionable 4.0
samples was performed. The foams were cured in an
Annoying 3.0
oven at 70°C for 24 hours. After curing, the foams
were stored at room temperature for 48 hours and then Very annoying 2.0
the test specimens were cut out o f the center o f the Intolerable 1.0
foam bun. These cut samples were stored at room
temperature in an isolated environment to avoid cross Physical Properties Testing
contamination for an additional 48 hours. Each sample
was then individually wrapped with aluminum foil and Physical testing o f the line trial foams from Table 5
sealed in a polyethylene bag. was conducted at Air Products’ Allentown laboratories
The first test, VDA -Verfahren 270B3 odor test was [3] utilizing various automated test equipment. Foam
used to quantify the odor o f each amine catalyst in this samples were tested after both ambient
study, using foam from Handmix Set 1. temperature/humidity conditioning and humid aged
In this test method, specimens were conditioned in a conditioning, as humid aged performance is a very
defined glass tube for 2 hours @ 80 °C. The samples important criterion for hydrolytically sensitive
were allowed to cool down to 60°C. At least three test polyester foams. The use o f reactive catalysts may
subjects evaluated the smell o f the samples according sometimes lead to unacceptable deterioration o f
to the scores listed in Table 6. If the results o f the test physical properties after humid ageing [4] depending
subjects differ by more than 2 points the entire test was upon intensity and duration o f the ageing conditions.
repeated for that sample. The standard hydrolysis conditions used for molded
This odor testing was performed at: automotive polyurethane’s (200 hours @ 90°C and
IMAT GmbH 100% Relative Humidity (RH)) are much too stringent
Institute fur Mefi- und Analysentechnik for hydrolytically sensitive polyester slabstock foams.
Prof. Dr. Schlegelmilch GmbH Therefore, in accordance with the method currently
Pollerbaumchen 1 used by European automotive producers, 200 hours @
D - 41169 Monchengladbach Germany 70°C and 100% RH was chosen.

434 / Huhtasaari, Plaumann, Grimminger, Kniss, Womack

RESULTS
Handmix Foams: Set 1 Amine Catalysts

Figure 3. Reactivity Profile Comparison N-

Ethylmorpholine versus Catalyst A

Figure 1. Reactivity Profile Comparison N-

Ethyimorpholine versus Dabco 2039

Figure 4. Reactivity Profile Comparison N-

Ethyimorphoiine versus Catalyst C
Figure 2. Reactivity ProfSe Comparison A/-
Ethylmorpholine versus Dabco NE400 The odor test results according to GME 60276 Test
specifications for foams prepared in Handmix Set 2 are
The use levels o f the catalysts evaluated were adjusted presented in Table 9.
until the Rate o f Rise profiles matched the standard
profile (1.70 pphp NEM). The results o f the use level Table 9. Odor Test GME 60276 For Handmix Set 2
adjustments for each catalyst are presented in Figures 1 Catalyst Odor rating
through 4. D abco N E M 3 .0
1 ,4 - D im e th y lp ip e ra z in e 4 .0
The results o f the use level determination, emissions
Dabco 2039 5 .5
and odor testing (VDA 270B3) for the first set o f
Dabco N E 400 8 .5
handmix foams are summarized in Table 8.

Table 8. Handmix Set 1 Catalyst Study, Use Levels, Emissions and Odor
Catalyst Use Level Total VOC Amine VOC FOG Odor VDA 270B3
[pphp] [ppm ] [ppm ] [ppm ]
D abco N EM 1 .7 0 148 77 110 2 .5 0

D abco N M M 1 .2 0 87 26 108 3 .0 0
D abco B D M A 1 .1 0 199 147 111 2 .0 0
D abco 2039 0 .7 0 533 480 106 2 .5 0

D abco N E 400 0 .5 0 56 0 118 1 .5 0

C a ta ly s t A 1 .0 0 245 192 202 2 .0 0
C a ta ly s t B 1 .2 0 103 38 113 2 .0 0
C a ta ly s t C 0 .6 5 159 109 107 1 .5 0

Huhtasaari, Plaumann, Grimminger, Kniss, Womack / 435

 Silicone Surfactants Machine Foams

The results of the emission tests for the Handmix Set The emission and odor data of the full-scale line trials
1 silicone surfactant evaluation are listed in Table 10. is summarized in Table 11.

Table 10. Surfactant Emissions Handmix Set 1 Table 11. Emissbn and Odor Data of Line Trials
Surfactant Total VOC FOG Total Amine Odor
FOG
Catalyst VOC VOC VDA
[PPm] [PPm] [PPm]
[ppm] .iP P m ] 270B3
Dabco D C 5526 255 207 Dabco N E 400 254 0 698 3 .0

Dabco D C 4 0 0 0 61 206 D abco 2 0 3 9 /

B lo w C a ta ly s t 605 340 746 3 .0
D abco D C 4020 46 178 C o m b in a tio n

C o m p e titiv e Dabco N M M 257 5 723 5 .0

S u rfa c ta n t 1 242 515

C o m p e titiv e
S u rfa c ta n t 2 69 203
Samples of each foam trial were cut to expose the cell
structure in the foam when viewed along the foam’s
C o m p e titiv e rise direction (top view) and perpendicular to the rise
S u rfa c ta n t 3 58 158 direction or foam’s face (side view). Each sample was
C o m p e titiv e
photographed at a 25X magnification to show the cell
S u rfa c ta n t 4 72 206 structure in detail using a Scanning Electron
Microscope (SEM). These comparisons are shown in
Figures 5 through 7.

Figure 5. Catalyst A Top View (left)/Side View (right) .

Figure 6: Catalyst B Top View (left)ZSide View (right)

436 / Huhtasaari, Plaumann, Grimminger, Kniss, Womack

 Figure 7. Catalyst C Top View (left)/Side View (right)

Physical Testing
Table 12 lists the standard physical properties for the machine trial flexible slabstock foam after conditioning a
minimum o f 48 hours at 25°C with 50% RH.

Table 12. Physical Properties Ambient Conditbning

Catalyst 0.65 pphp Dabco NE400 0.26 pphp Dabco 2039 1.2 pphp Dabco NMM Test Method
0.83 pphp Com petitive
Blow Catalyst
Top Middle Bottom Top Middle Bottom Top ... Middle Bottom
Density
(kg/m3) 37.4 37.0 38.1 37.0 37.4 37.3 38.9 40.2 39.6 ASTM
[pcf] [2.3] [2.3] [2.4] [2.3] [2.3] [2.3] [2.4] [2.5] [2.5] D-3574
A irflow
(slm) 14.4 12.4 12.1 18.3 14.3 15.2 33.5 19.7 25.3 ASTM
rscfml [0.51] [0.44] [0.43] [0.65] [0.51] [0.54] [1,181 [0.70] [0.89] D-3574
IFD:
25% (N) 202 224 195 183 202 184 239 265 239 ASTM
Dbfl [45] [50] [44] [41] [45] [42] [54] [60] [54] D-3574
65% (N) 379 426 380 331 374 340 456 513 472 Test B
m [85] [96] [86] [74] [84] [76] [103] [115] [106]
25% R (N) 142 158 143 133 147 137 165 187 173
m [32] [36] [32] [30] [33] [31] [37] [42] [39]
40% Hardness
(kPa) 3.6 4.1 4.1 3.6 4.1 4.2 4.0 5.1 5.2 DIN 53577
Dbfl [8.0] [9.2] [9.2] [8.01 [9.2] [9.3] [8.9] _ [11.5] r n .7 i
Tensile
(kPa) 155 165 165 170 174 171 192 179 195 ASTM D-
[PSI] [22.5] [24.0] [23.9] [24.6] [25.3] [24.8] [27.8] [25.9] [28.3] 3574 Test E
Elongation ASTM D-
(%) 286 273 253 281 289 265 285 270 261 3574 Test E
Tear
(N/m) 574 567 508 517 574 503 590 546 493 ASTM
m [3.3] [3.2] [2.9] [3.0] [3.3] [2.9] [3.4] 13.11 [2.8] D-3574
ASTM D-
90% C. Set
(%) 4.3 4.2 4.6 5.1 4.6 4.5 4.4 4.5 4.9 3574 Test D
50% C. Set
_[%}__________ 3.5 3.0 3.7 3.1 3.0 3.3 3.0 3.1 3.4 ISO 1856

All three machine produced foams were then subjected by conditioning for 200 hours at 95% RH / 70 °C, then
to an extensive heat aging process before testing their dried for 24 hours at 70 °C pries' to testing for physical
physical properties a second time. Initially, each fresh properties. These humid aged physical properties are
foam sample was heated for 24 hours at 70 °C, followed listed in Table 13.

Huhtasaari, Plaumann, Grimminger, Kniss, Womack / 437

 Table 13. Physical Properties After Humid A led Conditbning
Catalyst 0.65 pphp Dabco 0.26 pphp Dabco 2039 1.2 pphp Dabco NMM Test
NE400 0.83 pphp Com petitive Method
Blow Catalyst
Top M id d le B o ttom Top M id d le B o ttom Top M id d le B o tto m
IFD:
OCOL /k l\ 160 199 156 166 159 174 208 228 200 ASTM
Z 0 7 0 (IM) n v j i iv i
[36 ] [45] [35] [37 ] [36 ] [39] [47] [51] [45 ]
[Ibf] D -3 5 7 4
Test B
65% (N ) 326 393 320 321 307 334 414 466 416
[ibf] [73] [88 ] [72] [69] [69 ] [75] [93] [1 0 5 ] [93]

2 5 % R (N ) 113 138 112 119 117 126 147 159 141

[Ibf] [25 ] [31] [25 ] [27 ] [26] [28 ] [33] [36] [32]
40%
Hardness
(k P a ) 3 .2 3 .6 3 .7 3 .2 3 .7 3 .7 3 .6 4 .3 4 .4 D IN 5 3 5 7 7
Bbfl [7 .1 ] _ [8 .2 ] [8,3 ] [7.3 ] [8.3 ] [8 .2 ] [8.2 ] [9 .8 ] [9 .9 ]
Tensile
(k P a ) 153 183 180 166 177 186 183 177 201 ASTM D-
[P S I] [22] [27] [26] [24] [26] [27] [27] [26] [29] 3574 Test E
Elongation A STM D-
(% ) 275 276 223 268 261 244 148 124 156 3574 Test E
Tear
( N /m ) 578 582 537 566 592 564 621 600 528 ASTM
m [3 - 3] [3-3]_ [3 .1 ] [3.2] I3.41 [3.2 ] [3.5] [3 - 4] [3 .0 ] D 3574
90% C. Set A STM D-
(% ) 11 20 48 41 30 34 52 30 16 357 4 Test D
50% C. Set
(%) 4 .0 3 .4 3 .5 4 .6 4 .8 3 .8 3 .2 2 .7 3.1 IS O 1 8 5 6

DISCUSSIONS
Emissions evaporates undo* the stringent conditions of this
emission test. Dabco NE400 does not contribute to the
As presented in Figure 8 the low odor catalyst, Dabco emissions, as it cannot be detected in the VOC or FOG
2039 and non-fugitive catalyst, Dabco NE400 require condensate analysis, proving that the catalyst is
the lowest use levels of all catalysts evaluated. completely chemically bonded into the polymer matrix.
The conclusion can be made that the vapor pressures With the use of this study’s polyester formulation (see
of these catalysts, as they do not vaporize as readily Table 2), the remainder of the total VOC is quite
during the exothermic foaming process, to make them constant (50 to 71 ppm) and it is more than 95%
less likely to volatilize during foam production. The diethyleneglycoladipate, a byproduct of the polyester
total VOC and amine emissions of the highly volatile polyol.
catalysts, Dabco NEM, Dabco NMM and the Only catalyst A (1,4-dimethylpiperazine / N,N-
piperazine based catalysts A, B and C are relatively dimethyl-hexadecanamine combination) contributes to
low (Figure 9). Since most of these catalysts rapidly the FOG value, with almost two times the emissions
volatilized during foaming, the result is significant (202 ppm; see Table 8) compared to all other amine
higher plant emissions and worker exposure. catalysts used. The high molecular weight of Catalyst
The low odor, non-reactive catalyst, Dabco 2039 A with consequently low vapor pressure in conjunction
scores the highest VOC and amine emissions as it with non-reactivity makes this catalyst volatilize only

Figure 8. Handmix Evaluation Catalyst Use Levels

Comparison Figure 9. Amine Catalyst Emissions

438 / Huhtasaari, Plaumann, Grimminger, Kniss, Womack

 Figure 12. Surfactant FOG
Figure 10. Emissions Machine Produced Foam
The older generation competitive silicone surfactant 1
under the FOG conditions of 120°C. Comparing all the has more than double the FOG emission value
other catalysts, the FOG value varies between 106 and
compared to all the other silicone surfactants including
118 ppm. Closer analysis of the individual components
the first generation product Dabco DC5526. Again,
in the FOG emissions show more than 95% of these
the silicone surfactant contribution to the FOG
FOG emissions are again diethyleneglycoladipate and
emissions is mainly caused by the carrier/emulsifier
the rest are mainly traces of antioxidants, present in the
typically used in these older products.
polyester polyol. With the new generation silicone surfactants (Dabco
These laboratory handmix results of Set 1 were
DC4000 series), no contribution to VOC and FOG
confirmed in the full-scale production runs. Again,
emissions could be detected.
there are absolutely no amine emissions from the non­
fugitive Dabco NE400 catalyst in the VOC/FOG
Odor
condensate as demonstrated in Figure 10 and Table 11.
Flame retardants, polyester polyol components and
The odor test results of the Handmix foams Set 1
antioxidant ingredients are shown to be the main cause
according to VDA 270B3 do not show great differences
of the FOG emissions with no influence from the
in odor rating (Table 8 and Figure 13). As already
amine catalysts being detected.
discussed, most of the high vapor pressure catalysts
In the silicone surfactant portion of the Handmix Set
evaporate during foam formation and do not contribute
I emission study significant differences in terms of
to emissions or to the odor of the final foam.
VOC and FOG emission contribution of these additives
Furthermore, the small foam sample (handmix foam)
were found. Figure 11 shows the total VOC values of
and the significant conditioning steps prevent the
the silicone surfactants evaluated.
highly volatile catalysts from being detected or from
The first generation silicone surfactants, Dabco DC5526
contributing to the odor. Dabco NE400, however,
and competitive surfactant 1 for polyester applications,
scored the lowest odor contribution of all catalysts
contribute to the total VOC by about 200 ppm, consisting
tested.
mainly of emissions from the carrier and/or emulsifier.
The odor differences are significantly larger when
The new generation silicone surfactants provide for
using the GME 60276 method (see Table 9) on the
significantly lower emissions as demonstrated in Figure
Handmix Set 2 foams. In the GME 60276 test, the
II and 12. Dabco DC4020 provides for the lowest VOC
extensive conditioning steps before odor testing were
emissions of all silicone surfactants tested in this study.
not used. Dabco NE400 scored a rating of 8.5, which
The remainder of the low emission value for Dabco
corresponds to between noticeable and just noticeable
DC4020 (between 46 and 72 ppm) consists to a large
(see Table 7). Even Dabco 2039 provides for
extent, 98%, of diethyleneglycoladipate, with traces of
significantly lower odor compared to NEM and DMP
antioxidant from the polyester polyol.
with a rating of 5.5 (tolerable) vs. annoying and
objectionable, respectively.

Figure 11. VOC Surfactants Comparison

Huhtasaari, Plaumann, Grimminger, Kniss, Womack / 439

 third foam made with Dabco NE400 are once more very
similar throughout the physical property range after
humid aged conditioning.

CONCLUSIONS

The newly developed additive package consisting of

Dabco NE400 non-fugitive catalyst and Dabco DC4020
silicone surfactant provides for favorable standard
industry performance and low emission contributions
from the additive package in the final foam product.
Figure 13. Handmix Foams Odor Scoring VDA 270B3 Dabco NE400 meets the latest automotive industry
requirements for zero amine emissions, as well as,
The machine made foam performance in the VDA minimizes negative interactions with other plastic
270B3 test after five days of conditioning is as follows. materials and reduces human exposure to chemicals.
For the low odor catalyst package, Dabco 2039 and Physical properties, humid ageing and processing
commercial blow catalyst combination, as well as for the performance are not negatively impacted for the
non-fugitive Dabco NE400, a rating of 3.0 (noticeable, manufacturer, and therefore, absolutely comparable to
not yet annoying) was obtained. For the morpholine- standard industry catalysts. Dabco DC4020 is a high
based catalyst, the score was 5.0 (strongly annoying). performance silicone surfactant suitable for many
These results support the previous findings of the polyester slabstock grades, providing excellent
handmix foam odor evaluation. This validates the processing and reduced emissions.
conclusion that the non-fugitive catalyst Dabco NE400
has the lowest contribution to the odor of the polyester REFERENCES
foam.
[1] Hulme, S. P., Muha, K., and Harakal, M. E.,
Physical Properties “New Silicone Surfactants and Amine Catalysts
for Flexible Polyester Slabstock Foam”,
All three foams, shown in Figures 5 through 7, Proceedings of the SPI 34th Annual
produced at a customer manufacturing facility have a Technical/Marketing Conference, page 679.
fine regular cell structure along the rise direction (top [2] Daimler-Chrysler Test Method PB VWT 709:
view) and only slight elongation of cells when viewed Work Specification for Determination of Gas
perpendicular to the rise direction or foam face (side Forming and Condensation Emissions from
view). Automotive Interior Components by Thermo­
Table 12 shows equivalent airflows and densities with desorption.
the exception of Dabco NMM that resulted in higher [3] Balabhadra, V., Eckert, M. A., Hubble, D. R.,
densities and overall higher airflows, in spite of Robbins, S. E., and Rusak, M. E., Yoon, S. R.,
equivalent rise profiles when compared to the other two “Solving the Automation Puzzle - A
foam runs. Higher IFD’s, 40% Hardness, and Tensile Comprehensive Approach For Polyurethane
Strength values are also noted. It is believed that the Physical Testing”, proceedings of the API
unexpected physical properties are due to a lack of cell Technical Conference, October 2000.
wall drainage stability, leading to sigh back after [4] Wendel, S. H., Mercando, L. A., Tobias, J. D.,
fullrise, which increased the density, 40% Hardness, “Reduced Emanations Using Non-Fugitive
Tensile Strength and airflows. Catalysts For Flexible Polyurethane Foams”,
Foams made with Dabco NE400 compared to foams proceedings of the UTECH 2000 Conference,
with the customer’s current standard catalyst March 2000.
combination, Dabco 2039/Commercial Blow Catalyst
have very similar physical properties throughout the ACKNOWLEDGMENTS
range of tests. Dabco NE400 provides a more uniform
top to bottom gradient across the foam bun for the The authors would like to thank the efforts of the
majority of properties. Little or no shrinkage was also following people at Air Products & Chemicals, Inc. for
noted for both of these foam runs. their contributions to this work; Ilean Ruhe, Steve
After the humid aged conditioning sequence of the Robbins and Renae Bennett. We would also like to
foam was complete, similar trends in physical properties thank Mr. Herbert Fischer of Eurofoam, Linz, Austria
were observed again between the three foam runs in for his assistance with the line trial foam and his
Table 13. The foam produced with NMM maintained permission to publish the results in this study.
higher IFD’s, 40% Hardness, Tear and Tensile
Strengths. Elongations were reduced.
The customer standard catalyst package foam made
with Dabco 2039 / commercial blow catalyst and the

440 / Huhtasaari, Plaumann, Grimminger, Kniss, Womack

BIOGRAPHIES

Matti S. Huhtasaari Jane G. Kniss

Matti Huhtasaari is currently Jane G. Kniss is currently an
an Application Development Applications Chemist in the
Chemist in the Solutions Solutions Technology Group
Technology Group at Air at Air Products and
Products and Chemicals, Inc. Chemicals, Inc., located at the
located at the Technology corporate headquarters in
Centre in Norderstedt, Allentown, Pennsylvania. She
Germany. He received his received her B.S. degree in
Ph.D. from Westfahlische Polymer Science from
Wilhelms Universitat Munster, Pennsylvania State University,
Germany in 1984. University Park, PA in 1986. She is currently
He is currently responsible for applications responsible for applications development and technical
development and technical service for catalysts and service for catalysts and surfactants with a dual
surfactants with emphasis on supporting the flexible emphasis on supporting both the flexible slabstock and
slabstock market segment. flexible molded market segments.

Reno Plaumann Frank D. Womack

Reno is currently working as Frank is currently the Global
an Application Technician Market Manager for Flexible
for AD/TS in the area of Foam Additives at Air
flexible slabstock foams. He Products and Chemicals,
received his B.S. degree in Inc., located at the
1997 from Fachhochschule Corporate Headquarters in
Hamburg, Germany. After Allentown, Pennsylvania.
he finished his studies, he He received his B.S. Degree
joined Air Products and in 1983 from the University
Chemicals GmbH in of Tennessee. He has held
Norderstedt, Germany. He is responsible for manufacturing,
applications development and technical service support technology, sales and marketing positions in the
for the flexible slabstock market segment. polyurethane industry over the past 18 years. His
current responsibilities include the global development
Jobst Grimminger and marketing of new products for flexible
Jobst is currently the polyurethane foam applications.
Technical Manager of
Solutions Technology group
iji Europe. He received his
B.S. degree in 1983 from
Fachhochschule Aaelen,
Germany. After 8 years of
experience in development
and technical service
support at ZELU Chemie, a
system house in Germany, Jobst joined Air Products
and Chemicals GmbH in Norderstedt, Germany in
1991. There he held the position of Senior Application
Chemist with responsibilities for rigid polyurethane
foam additives and additives for high density molding
applications for 7 years prior to his current expanded
responsibilities for both the flexible and the rigid foam
disciplines.

Huhtasaari, Plaumann, Grimminger, Kniss, Womack / 441

Novel Polymer Polyol with High Polymer Content and
Low Viscosity

T. TOMOSADA, H. OHTA, S. KONO, T. FURUTA a n d H. HORIGUCHI

Sanyo Chemical Industries, Ltd.
Research Division
11-1 Ikkyo Nomoto-cho, Higashiyama-ku
Kyoto P.C. 605
Japan

ABSTRUCT INTRODUCTION
High solids polymer polyols are widely used in the Polymer polyols have been used in polyurethane
polyurethane industry for the purpose of enhancing the applications for a long time. Polymer polyol is dispersion
load bearing capacity. The most widely used polymer of solid particles consist from styrene, acrylonitrile and
polyols consist of stable dispersions of styrene and other polymers in polyether polyol and useful to increase
acrylonitrile copolymer particles. The technology for the load bearing capacity of polyurethane foam. In the
producing these polymer polyols continues to advance to eighties, the content of polymer polyols were limited to
meet the needs of new polyurethane foam applications. solid levels of less than 30 weight %. In the nineties, the
content increased to about 40 weight % by using effective
Recent polyurethane foam industry requires high dispersant. But recently, the requirement for high polymer
solids polymer polyol to enhance the load bearing content polymer polyol has become much more stronger.
capacity of a wide variety of foam types (slabstock, HR, High polymer content polymer polyol can increase the
RIM, semi-flexible, etc.). High polymer content (over load bearing capacity of polyurethane foam further more.
45wt%), good filterability and low viscosity are the key But it causes the rise in viscosity of polymer polyol to
requirements. increase the polymer content. So it is difficult to handle
such a viscous sample. The most challenging target of
When polymer content is high, polymer polyol’s polymer polyols has been the compatibility of high
viscosity is high too. When polymer content is 50 polymer content such as 50 weight % and low viscosity
weight %, polymer polyol’s viscosity is over 10,000mPav such as 5,000mPa •s.
s. But polyurethane foam industry requires high solids The relationship of polymer content and viscosity in
polymer polyol with low viscosity. conventional polymer polyols is shown in Figure 1.

We found the process that can reduce the viscosity of

high solids polymer polyol. In this paper we present
Sanyo’s new polymer polyol that has high polymer
content such as 50 weight % and low viscosity such as
4,800mPa •s at 25*C.

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 443

 We found the new process for producing high
polymer content and low viscosity polymer polyols. This
process is the way to prevent generation of the soluble
oligomer to polyether polyol in addition to the
coalescence of polymer particle by using special
monomers. Our newly developed polymer polyol has 50
weight % of polymer content and 4,800mPa -s viscosity at
2 5 V ,. So it is useful to increase the load bearing capacity
of polyurethane foam without the difficulty for handling.
Also it is possible to decrease the amount of polymer
Figure 1. The relationship o f polym er content and polyol for obtaining the equal load bearing capacity
viscosity in conventional polymer polyols polyurethane foams. The relationship of polymer content
and viscosity of polymer polyol by using our newly
The factor of rise in viscosity has been considered as
developed process is shown in figure 1.
the coalescence of polymer particles that include the
DISCUSSION
polyether polyol in them. It means the decrease of
The concept of "Sanyo’s new polymer polyol”
polyether polyol as dispersion solvent and apparent
The high content polymer polyols tend to have much
polymer content grows higher than the real polymer
higher viscosity than the conventional polymer polyols.
content. It causes rise in viscosity of polymer polyols.
As mentioned before, the factor of high viscosity has
In addition to the coalescence of polymer particles,
been considered to the coalescence of polymer particles.
we found out that the soluble oligomer to base polyether
But we found out that the factor was not only the
polyol is another large factor of rise in viscosity. Soluble
coalescence of polymer particles but also soluble
oligomer is produced from polymerization of monomers
oligomer to base polyether polyols. We studied those
with insoluble polymers at the same time. The soluble
factors and considered that it would be possible to
oligomer has high viscosity and causes rise in viscosity of
produce the polymer polyols that are compatible of high
polymer polyol. We examined the relationship of soluble
polymer content and low viscosity by preventing
oligomer content and viscosity of polymer polyol by dispersing
generation of the soluble oligomer to polyether polyol in
the polymer separated from conventional 50wt% content
addition to the coalescence of polymer particles.
polymer polyol to base polyether polyol adjusted to fixed
We developed the new polymerization process of
soluble oligomer content. The relationship of soluble
polymer polyol in order to remove those factors by the
oligomer content and viscosity of polymer polyol is shown in
following methods;
figure 2.
©W e polymerized the special monomer A that has high
affinity to polyether polyol with acrylonitrile, styrene
and other monomers to prevent the coalescence of
polymer particles in polymer polyol. So we obtained
the low viscosity polymer polyol with stable polymer
particles. Special monomer A has excellent function
as dispersant.
©W e polymerized the special monomer B that has large
molecular weight with acrylonitrile, styrene and other
monomers to prevent the generation of soluble
oligomer to polyether polyol. So we obtained the low
Figure 2. The relationship of soluble oligomer content viscosity polymer polyol without soluble oligomer.
and viscosity o f polymer polyol with 50wt% content This paper refers to the result of such study by the
(1 .3 u m polym er particle) application of Sanyo’s new polymer polyols. Here we

444 / Tomosada, Ohta, Kono, Furuta, Horiguchi

focus on the special monomers to obtain polymer polyols The reason of the difference between figure 1 and
with high polymer content and low viscosity. Sanyo’s figure 4 is the difference of apparent polymer content as
new polymer polyols can be applied for wide variety of mentioned before.
slabstock foam.
And we observed the high molecular weight oligomer
Viscosity o f high polymer content polymer polyol that dissolved in base polyether polyol separated from
Viscosity of polymer polyol becomes higher polymer polyol. It is another factor to increase polymer
acceleratively as polymer content increases. The polyol’s viscosity. This observed soluble oligomer’s
relationship of polymer content and viscosity in molecular weight is 10,000 to 30,000 and viscosity is
conventional polymer polyols is shown in Figure 1. When over 100,000mPa • s. We investigated the amount of
polymer content changes, polymer particle size varies as soluble oligomer in polymer polyols with four levels of
shown in figure 3. polymer content. The relationship of polymer content and
amount of soluble oligomer is shown in figure 5.

Figure 3. The relationship o f polymer content and

polymer particle size
Figure 5. The relationship o f polym er content and amount

But we estimated that the relationship of polymer o f soluble oligomer to polyether polyol

content and viscosity of polymer polyol would be as

shown in figure 4, if polymer particle size did not vary. The amount of soluble oligomer increases in
proportion to the polymer content.
We tried to remove the soluble oligomer from
polymer polyols and measure the viscosity of polymer
polyols without soluble oligomer. It was possible to
obtain the polymer polyols without soluble oligomer by
using the dispersion of the polymer particles separated
from polymer polyol in the same base polyether polyol.
T he re la tio n sh ip of p olym er c o n te n t a n d viscosity
of polym er polyol w h en soluble oligom er to
polyether polyol does n o t e x ist in polym er polyol is
show n in figure 6.

Figure 4 . The relationship o f polym er content and

viscosity o f polym er polyol when polym er particle size
does not vary (estimated graph)

Tomosoda, Ohta, Kono, Furuta, Horiguchi / 445

 The relationship of monomer A content and viscosity
of polymer polyol when polymer content is 50 weight %
is shown in figure 8.

Polym er Content. (%)

Figure 6 . The relationship of polymer content and

viscosity o f polym er polyol when soluble oligomer M onom er A content. (% )

to polyether polyol does not exist in polym er polyol Figure 8. The relationship o f monomer A content and
viscosity of polym er polyol when polym er content is 50
So we concluded t h a t we could o b tain th e weight %
polymer polyols with high polymer content and low
viscosity by preventing the coalescence of polymer Viscosity of polymer polyol drops as the amount of
particles and generation of the soluble oligomer to monomer A increases up to 2 weight %. But viscosity of
polyether polyol. polymer polyol using over 3 weight % of monomer A is
slightly higher than 2 weight % monomer A polymer
The means fo r decreasing the viscosity o f polym er polyol. Therefore, best additional amount of monomer A
polyol is 2 weight %.
We developed the new polymerization process of We used monomer B to prevent the generation of
polymer polyol by using special monomers to obtain the soluble oligomer to polyether polyol. Monomer B has so
high polymer content and low viscosity. special formula that the polymer generated from the
We used monomer A to prevent the coalescence of co-polymerization of acrylonitrile, styrene, monomer B
polymer particles. The relationship of amount of and other monomers is hard to dissolve in polyether
monomer A and polymer particle size of polymer polyol polyol. Polymers are easy to become insoluble to
when polymer content is 50 weight % is shown in figure polyether polyol and formulate the stable particles. The
7. relationship of monomer B content and soluble oligomer
content is shown in figure 9.

Figure 7. The relationship of amount o f monomer A and Figure 9. The relationship of monomer B content and
polym er particle size o f polym er polyol when polymer soluble oligomer content viscosity of
content is 50 weight % polym er polyol when polym er content is 50 weight %

446 / Tomosada, Ohta, Kono, Furuta, Horiguchi

 The relationship o f monomer B content and viscosity the same amount o f “ H A R D U P F S - 103 ” as conventional
o f polymer polyol when polymer content is 50 weight % polymer polyol, foam hardness increased about 20%.
is shown in figure 10 Other properties are slightly superior or equal to those
from conventional type.

TABLE 1. Foam property o f the polyurethane foam produced

from “H A RDUP F S -10 3 ”
HARDUP FS -103 80 64 —

SANNIX FL -555(conventional POP) — — 80

SANNIIXGP -3000(conventional PPG) 20 36 20
Water 3 3 3
DABCO 33LV 0.25 0.25 0.25
SRX -280A 0.8 0.8 0.8
Stannous octoate 0.19 0.19 0.16
TDI -80 38.2 38.2 38.2

Figure 10. The relationship o f monomer B content and NCO INDEX 110 110 110

viscosity o f polym er polyol when polym er content is 50 Viscosity of PPG premix(mPa -s) 2260 1850 2120
weight % Rise Time (sec) 90 92 97
Density (kg/m3) 3 2 .4 3 2 .2 3 3 .6
Viscosity o f polymer polyol drops as the amount o f
25% ILD ( kgf/314cm 2) 3 7 .6 3 3 .6 3 0 .4
monomer B increases. But polymer chain flexibility Resilience (% ) 27 28 30
containing over 6 weight % o f monomer B is so soft that Tensile strength (k P a ) 1 69 1 58 145
Elongation (% ) 75 75 70
load bearing capacity o f polyurethane foam produced
Tear strength (N/cm) 7 .2 7 .4 7.1
from this polymer polyol is lower than that o f w e expect. Compression S E T (% ) 2 .4 2 .4 2 .6
The best additional amount o f monomer B is 5 weight %. Air flow(ft3/min) 1.4 1.5 1.6
We obtained the polymer polyol “ H A R D U P F S - 103 ”
with 50 weight % o f polymer content and 4,800m P a, s We also evaluated the foam properties by varying the
viscosity by using best amount o f monomer A and amount of “H A R D U P F S - 103 ” . It showed good
monomer B. The relationship o f polymer content and properties at any level o f polymer content. Table2 and
v is c o s ity o f p o ly m e r p o ly o l when monomer A and figure 11 show the foam properties at some level o f
monomer B are added in best amount is shown in figure 2 polymer content.
shown before.
TABLE 2 . Foam properties o f the polyurethane foam when
The evaluation results of the polyurethane foam the amount o f “H ARDUP F S -103 ” varied
produced from the new polymer polyol HARDUP FS -103 20 50 80
SANNIIXGP -3000(conventional PPG) 80 50 20
We evaluated our new polymer polyol “ H A R D U P
Water 3 3 3
F S - 103 ” by measuring the properties o f the polyurethane DABCO 33LV 0.25 0.25 0.25
foam produced from this polymer polyol. This new SRX -280A 0.8 0.8 0.8
Stannous octoate 0.19 0.19 0.19
polymer polyol indicated the characteristics w e had TDI -80 40.7 39.4 38.2
expected. The properties o f the polyurethane foam NCO INDEX 110 110 110
produced from “ H A R D U P F S - 103 ” compared with that Viscosity of PPG premix(mPa -s) 640 1100 2260
from the conventional type polymer polyol “ SA N N IX Rise Time (sec) 97 87 90
FL - 555 ” is shown in table 1. Density(kg/m3) 3 3 .4 3 3 .2 3 2 .4
25% ILD (kgf/314cm 2) 2 0 .4 2 5 .5 3 7 .6
T ablel shows the application data for slabstock foam.
Resilience(%) 42 38 27
Largest merit o f “ H A R D U P F S - 103 ” is the increase o f Tensile strength (kP a) 107 1 47 169
the 25%ILD (load bearing capacity). Even we adjusted Elongation (% ) 118 1 04 75
the polymer content same as conventional polymer polyol, Tear strength (N/cm) 6 .0 7 .4 7 .2
Compression S E T (% ) 1.8 2 2 .4
the foam hardness increased about 10% and the viscosity
Air flow(ft3/min) 2.1 2 .4 1.4
o f polyol premix decreased about 10%. When w e used

Tomosoda, OhtayKorio, Furuta, Horiguchi / 447

 Figure 11. Foam properties of the polyurethane foam
when the amount o f “HARDUP FS -103 ” varied

CONCLUSION
*The factors o f high viscosity o f high polymer
content polymer polyol are not only the coalescence o f
polymer particles but also soluble oligomer to base
polyether polyols.
•It is possible to prevent generation o f the soluble
oligom er to polyether polyol by using special monomer at
the same time with preventing the coalescence o f polymer
particles.
• Polyurethane foam from the new high content
polymer polyol produced by this new technology
indicates better properties than conventional type.

REFERENCES
1. T. Shimada, “The effect o f morphology o f polymer
particles on the viscosities o f polymer dispersed polyols ”
Polyurethane World Congress 1987
2. H.R. Van Der Wal, “Epoxy dispersion polyols: A novel
polymer - modified polyol for improved load bearing and
resiliency ” Polyurethane World Congress 1987
3. Stanley L. Hager, “Advances in high solids polymer
polyol technology for new slabstock foam processes and

applications ” S P I 1999
4. International publication No.: W 001/0 9 2 4 2

448 / Tomosada, Ohta, Kono, Furuta, Horiguchi

BIOGRAPHIES TAKESHI FURUTA
TSUYOSHI.TOMOSADA Takeshi Furuta received a
.Tsuyoshi Tomosada B.S. and M .S. degrees in
received a B .S. degree in chemistry from Kyoto
organic chemistry from University in 1996 and
Kyoto University in 1982. 1998. He joined Sanyo
He joined Sanyo Chemical Chemical Industries, ltd.
Industries, ltd. in 1982 and in 1998 and has engaged
has engaged in development in development of
o f flexible, semi- rigid and polyurethane foam and
rigid polyurethane foam elastomer.

HIDEMASA HORIGUCHI
HIDEFUMI. OHTA Hidemasa Horiguchi
.Hidefumi Ohta received a received a B.S. and M.S.
B.S. and M.S. degrees in degrees in chemistry from
chemistry from Kumamoto Toyama University in
University in 1980 and 1982. 1996 and 1998. He joined
He joined Sanyo Chemical Sanyo Chemical Industries,
Industries, ltd. in 1982 and ltd. in 1998 and has
has engaged in development engaged in development
of flexible polyurethane o f flexible polyurethane
foam foam.

SHOICHIRO KONO
Shoichiro Kono received a
B.S. and M.S. degrees in
industrial chemistry from
Nagasaki University in
1990 and 1992. He joined
Sanyo Chemical Industries,
ltd. in 1992 and has engaged
in development o f flexible
polyurethane foam.

Tomosada, Ohta, Kono, Furuta, Horiguchi / 449

Flexible Foam Formulating Technology Development via a
New Modeling Approach

ROBERT J . LOCKWOOD
Huntsman Polyurethanes
2190 Executive Hills Blvd.
Auburn Hills, Michigan 48326
ABSTRACT
benefits o f these two distinctive isocyanates and formulate
Flexible foam system s based on MDI are com plex and a rational technology. The objective w as to develop a
muti- dimensional in their response and typically require formulating technology based predominantly on MDI with
many iterative experiments to achieve specific targeted maybe 5 to 20% TDI. This w ould allow for many o f the
formulations. W hile practical/workable formulations are key benefits o f MDI to be maintained: high hardness, fast
eventually achieved by use o f an iterative approach, the cure, good cell opening. A t the same time it was hoped
broader technological basis for the formulation is never that these levels o f TDI would solve some o f M DI ’s
fully understood nor mastered. deficiencies: low blow ing efficiency, low tear strength and
A program was undertaken to explore the technology elongation, and poor hysteresis.
potential for M DI/TDI blends in flexible foams. Due to Traditionally, polyurethane formulating is done via many
the com plexity o f this undertaking, it w as decided that an iterative experiments and the ‘art’ or ‘gut - feel ’ o f the
experimental design approach might elucidate the working formulator. It is particularly difficult to do experimental
behavior o f M DI/TDI blends in flexible foams much better design with flexible polyurethane foams due to the
than the old iterative approach. The predictive models formulation holes (i.e. collapsed or shrinking foam) over a
achieved with experimental design clarified the range o f selected variables. It can be done, however, with
formulation ranges achievable. These predictive models good planning, trial screenings, and the fortitude to achieve
were then further analyzed via a ‘fundamental polymer a viable design. W hile this still requires significant work
physics calculation m odel ’ called FormCalc© to further and experimentation, the p ayoff is in the great
define the 4Root Polymer ’ basis for the key formulation understanding, insight, and predictive models which fully
responses. U sing this approach it was possible to model define the new technology.
formulation requirements in terms o f the structure property
relationships for various applications and processes. EXPERIMENTAL DESIGN APPROACH

INTRODUCTION Due to the known com plexity o f flexible polyurethane

foams, it was decided that simple designs with only a few
Huntsman introduced MDI into the slabstock market in variables would be the best approach. This approach has
the 9 0 ’s with the introduction o f the Waterlily® two benefits: It means fewer experiments to complete a
technology. This new technology has developed into a design and easier pre -rationalization o f the variable ranges
small, but significant commercial segment in Western to be selected. The use o f such an approach meant that a
Europe. North American slabstock producers were not number o f designs would have to be run, to properly cover
interested in the niche or specialized system approach o f a wider range o f response in the key variables.
W aterlily®, so other alternative w ays to introduce MDI Y ou should always use as much scientific knowledge and
understanding at your disposal when designing
were looked at. One o f the main problems with MDI is the
experiments. Huntsman uses a proprietary formulation
many different types required for various applications.
MDI is very good for the higher density and higher calculation program called FormCalc© to assist in fully
hardness products, but they represent only a very minor rationalizing the experiments [1]. In addition, the novel
portion o f the slabstock market. TDI excels in the approach o f reanalyzing the experimental designs with the
preparation o f low density foams which makeup the bulk input o f the FormCalc© parameters proved to be extremely
o f die slabstock market. W hy not consider a blend o f MDI beneficial to get to a 4Root Polymer ’ and m odeling basis
and TDI to merge the for some o f the responses.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 451

 A very simple and user friendly software package called T a b le 1 . C o n tro l V a ria b les , L e v e ls & D esig n s f o r
ECHIP specifically designed for scientists and engineers, E X P E S IG N -2 & E X P D E S IG N -3
was used. It has embedded statistics that automatically
determine lack-of-fit and interactive process optimization
pictures that help with the visualization and understanding Experimental Designs
the relationships between the control and response Sim ple 3 V ariable
variables.
The design intent or objective o f this study was to create • EXPDESIG N - 2 • EXPDESIGN 3
formulations yielding free rise foams covering 25% IFD Hard Foams >21 IFD Soft Foams <21 IFD
hardness in the range o f 10 to 35 lbs. and densities in the - 10% - 15% TDI - 5% -10% TDI

range o f 1.6 to 3.0 pcf. Enough early iterative and pre­ - 2.75 3.75 pbw W ater - 2.5 - 3.25 pbw Water

experimentation work was done to know that two - 0-40% H S 100 graft - 0 - 20 % H S 100 graft
experimental designs would be necessary to cover this • Box Behnken Design • Q uadratic Design
range. In keeping with a simple design only three control - 13 discrete experim ents - 15 discrete experim ents
variables were selected: % TDI, Water level and % High - 5 replicates - 5 replicates
Solids Graft Polyol. EXPDESIGN - 2 covered the hard
foams or those > 21 IFD and EXPDESIGN - 3 respectively
the soft foams or those < 2 1 IFD. These two designs are T able 2. Formulations Run for EXPDESIGN - 2 & 3
described in Table 1. N ote that it was very necessary to
F o rm u la tio n s Used
run different ranges o f TDI in the soft and high hardness
ExpDesign-2 ExpDesign-3
foams and the level o f Graft polyol was appropriately
Hard Foams Soft Foams
adjusted. The use o f only 3 control variables led to no Materials PBW PBW
more than 15 discrete experiments and 5 replicates for a 1 ^ Variable 5500 MW HR Triol 6 0 -1 0 0 8 0 -1 0 0
total o f 20 trial experiments to complete the design. The ^ Variable HS100- Graft 0 -4 0 0 -2 0

two experimental designs, a Box - Behnken for Constant Surfactant 0.7 0.7
Constant T-12 0 .1 2 0 .1 2
EXPDESIGN-2 and a Quadratic for EXPDESIGN-3 were Constant Jeffcat ZF-22 0 .1 2 0.24
run via laboratory hand mixing. Since the main objective 2 Variable Water 2.75 - 3.75 2.5 - 3.25
was to characterize the hardness and density response for
Constant NCO Index 102 102
these design series, the only measurements necessary were
free rise density and 50% CFD. It w as already know that a Wt% Blend Wt% Blend
very good relationship exists between 50% CFD and 25% 3 ^ Variable Rubinate 7400 85% - 90% 90% - 95%
IFD (i.e 25% IFD,lb. = 1 2 .5 x 50% CFD,lb), so this ^ Variable Rubinate TDI 8020 10% -15% 5% -10%
allowed for the easier production o f smaller scale handmix
foams. The objective was to characterize the hardness
range and response o f this new technology rather than to I f the resultant free rise densities are compared to the
argue about the exact definitional 25% IFD. theoretical density at 100% assumed water reaction, then a
The formulations were kept as simple as possible using % B low ing E fficiency can be calculated as shown below:
standard catalysts and additives. A low functionality MDI
variant Rubinate 7400 was blended with standard 80/20 ( 1)
TDI. The basic formulation for each set o f experiments is
shown in Table 2.
The nice thing about % B low ing Efficiency is that it w ill
RESULTS & DISCUSSION illustrate just how open / unstable or tight / stable a
formulation is on a com m on basis for all different water
Designed Experiments levels and formulation variations. From this information
and som e know ledge o f handmix versus machine blowing
One may wonder what could be learned from only efficiency differences, the best utilizable range or process
analyzing the free rise density and the hardness, however, window o f the control variables can be determined. Such
with the proper treatment o f the data and the predictive information also gives a clear understanding as to the key
m odeling via the designed experiment a wealth o f formulation parameters controlling cell openness and
knowledge can be obtained. Besides just characterizing stability. Figures 1 & 2 show the blow efficiency
the response free rise density and hardness o f these series response for the two designs run. ExpDesign - 2 only had a
o f formulations, a complete understanding o f these small region o f low blow ing efficiencies due to the higher
responses in relation to the control variables or levels o f % TDI used. Handmix and machine correlations
formulations was the objective. Thus, a deeper explanation studies demonstrated that blow ing efficiencies o f 80% or
o f the resultant densities and hardness from a fundamental above on a handmix basis would most likely cause
scientific basis was an important secondary goal. shrinkage on a machine due the more efficient mix. Thus,

452 / Lockwood
the machine process w indow for ExpDesign - 2 is very
small. This is very much the reason w hy ExpDesign - 3 was
designed with lower levels o f % TDI along with the lower
water levels to assure open processable foam on a machine.

F ig u r e 3. E X P D E S IG N -2 3 -D im en sio n a l R e sp o n se o f
D e n sity

F ig u r e 1. E X P D E S IG N - 2 S m a ll P r o c e ssa b le W in dow a t
le ss than 8 0 % B lo w E fficien cy

F ig u r e 4 . E X P D E S IG N -2 3 -D im en sio n a l R esp o n se o f
B lo w efficien cy

F ig u r e 2 . E X P D E S IG N - 3 P r o c e ssa b le W in dow a t le ss
than 8 0 % B lo w E fficien cy (c o v e rs e n tire ra n g e)

One o f the main things that the experimental design

pictures emphasize about this and any series o f flexible
foam formulations, is the extreme com plexity o f the
pictures. The curvilinear 3 - dim ensional graphs in Figures
3,4,5 illustrate just how com plex and confounding the
responses can be. The B low Efficiency curvature
ultimately influences the actual density yield and then the
density ultimately affects the 25% IFD hardness. F ig u r e 5. E X P D E S IG N -2 3 -D im en sio n a l R esp o n se o f
2 5 % IF D

Lockwood / 453
 It is no wonder that polyurethane formulators call it’s
practice an art or black science. These responses are
com plex and all interrelated! Indeed, experimental design
is the only w ay to capture some understanding o f the
com plex relationships between these variables. The 3 -
dimensional models do predict the true com plex reality o f
the control variables and how they effect the property
responses o f the formulations.

Normalization Techniques

Can this normal com plexity o f flexible formulations be

simplified further? The answer is yes, if you are w illing to
sim plify the model. The density is in fact the result o f the
blow ing efficiency and ultimately hardness is a function o f
both the hard block com position o f a formulation and the
resultant density. One can use the very fundamental
relationship between density and hardness or the good
approximation, that final hardness is a function o f the
density squared:
F ig u r e 6. 2 5 % IF D a t 1.8 p c f N o rm a lize d D e n sity
S h o w in g R e d u c e d C o m p lex ity
( 2)

If Hi and Di are known, then it is possible to approximate

the new hardness H2 at any density D 2 with the follow ing
equation:

(3 )

U se o f this equation allows one to approximate the

hardness at any density based on the actual raw data. For
example, the normalization o f all the densities to a 1.8 pcf
density and recalculation o f the 25% IFD at this density
leads to a less com plex and more rational picture than in
Figure 5. See Figure 6. The term [ (DO2] in the
equation above can actually be treated as the ‘R o o t
P o ly m e r H a rd n e s s ' term. B y dividing, the hardness by the
density squared, the hardness is in fact normalized and then
the hardness becom es independent o f the density . When
this is done, the previously com plex curvilinear 3-
F ig u r e 7. ‘R o o t P o ly m e r H a rdn ess* In d ep en d en t o f
dimensional graphs becom e an almost smooth plane as
D e n sity
shown in Figure 7. N ow w e are view ing the real hardness
changes due to the control variables as a result o f the
polymer hard block content o f the formulations.
and hard segments and even polyurea hard block levels.
These calculated results were then re -inputted into the
FORMULATION / POLYMER MODELING
design. Before getting into a discussion o f the results with
combined ECHIP™ & FormCale© analysis, it is necessary
The EC H IP™ designs brought much understanding to
to explain som e o f the output parameters o f FormCale©.
these MT based formulations in terms o f their behavior, FormCale© output calculates the Polyurea and Total Hard
trends and base properties. W ould it be possible to go B lock contents on a different basis, when particulate solids
further to the (ro o t p o ly m er* basis o f these changes? In are introduced into a formulation due to the addition o f
order to attempt this approach, it w as necessary to input the Graft or PIPA polyols. In one case, described as the
formulation experiments into FormCale©, in order to F o a m B a sis *, the added graft or Pipa polymer particles are
easily calculate the polymer building blocks such as soft treated as part o f the total foam com position and they
contribute to the total hard block content. In the other

454 / Lockwood
case, the polymer particles are subtracted from the total
foam and the resultant becom es the final ‘Polymer Basis
The addition o f polymer particles actually dilutes the soft
segment polyol level thereby leading to higher calculated
hard block levels whether as polyurea hard block or total
polymer hard block. A further explanation o f the above
terms and the software FormCalc© can be found in another
Huntsman paper [1],
Logically, w e should be able to correlate the polymer
hardness or normalized values directly to the FormCalc©
output values. Figure 8 show s graphical analysis o f the
EXPDESIGN - 2 data, Hardness/Density2 versus the %
Polyurea Hard Block contents calculated by the ‘Foam and
Polymer Basis ’ methods. The results clearly show that
there is very poor correlation via the ‘Foam B asis ’ with an
R2value o f only 0.469 and a much better correlation via the
‘Polymer B asis ’ with an R2 value o f 0.874. Next, the same
analysis was done using the % Total Hard Block via the
two methods. Figure 9 shows pretty good correlation
in both cases, but the ‘Polymer B asis ’ does better with an
R2 value o f 0.923. Clearly, the ‘Polymer B asis ’ method o f
calculation gives a much better correlation with
formulations incorporating different levels o f graft polyol.
This is because the Polymer Basis reflects the true
influence o f graft on the formulation or the fact that graft
dilutes the polyol content and thus forms more Figure 8. Statistical Fit o f ‘Root Polymer Hardness'
concentrated or higher hard block polymer experimental versus the % Polyurea Hard Block
design and treated as responses [2,3]. A s expected, there
was a very high degree o f correlation versus the design
equation terms.
N ow that it was known that the ‘Polymer Basis '
treatment was best, the data o f both experimental designs
was combined and evaluated for the best statistical fit for
hardness prediction. Plotting o f both experiments led to an
overall very good statistical fit with an R2 value o f 0.93.
This fit is quite incredible considering the wide range o f
the control variables covering ExpDesigns 2 & 3. This
ultimately means that by the use o f hardness/density2
normalization and FormCalc© calculation o f the true
polymer hard block basis one can account for most o f the
prediction for hardness. Further analysis was done, by
treating the data as a series o f %Graft levels and refitting
the data. A s shown in FigurelO, the R2 values were
between 0.96 and 0.98 and the regression lines show
clearly that the slope o f the lines increases with increasing
% Graft as would be expected.
Broad application modeling is also possible using the
‘Root Polymer Hardness' data obtain from these designed
experiments. The ‘Root Polymer Hardness' data for free
rise foams at atmospheric pressure can be extrapolated by
means o f Equation 3 to any other densities achievable by
processing; This would include density reduction by
auxiliary blowing agent or lower than atmospheric pressure
and density increase via overpacking or molding the Figure 9. Statistical Fit o f the ‘Root Polymer Hardness *
formulation. Table 3 shows that the wide range o f final versus the % Total Hard Block
products that, in theory, could be achieved by processing.
The wide range o f product hardness achievable, is quite
phenomenal with changes in both the hard block o f the

Lockwood / 455
 T able 3. M odeling H ardness at A ny D ensity for V arious
A pplications U sing Equation 3

APPLICATION MODELING OF FLEXIBLE FOAM

O N T H E B A S IS O F
_____________ 'ROOT POLYMER _
H
A_
1
1 Vacu. F ree M o ld M o ld
1 C02 Rise 1.2X- 1.6X-
Density-pcf [ @ 1 0 1.65 2.25 2.7 3.6
ExpDsgn Polymer 25% 25% 25% 25%
‘ T ria l i — 1EQ — IEO ___ IEO ___ LED
"9 I 3.6 9.9 18. 26. 47.
6 ' 4.7 12. 24. 34. 61.
2 ! 5.2 14. 26. 38. 68.
1 ; 6.2 17. 31. 45. 45.
5 ' 6.3 17. 32. 46. 82.
14 ! 7.3 20. 37. 53. 95.
10 ; 7.5 20. 38. 55. 97.
13 . 7.7 21. 39. 56. 99.
11 l 8.0 22. 40. 58. 104.
4 ' 9.3 25. 47. 68. 121.
8 ! 9.9 27. 50. 72. T28.
3 | 10.2 28. 52. 74. 133.
7 i 11.1 30. 56. 81. 144.
12 | 13.0 35. 65. 94. 168.
‘

Figure 10L Best Prediction Model for Hardness for the

Combined EXPDESIGNS

polymer and density o f the end product. Hard block

changes alone account for a 3.6X multiple o f the polymer
hardness. Then, depending on the range o f product density
selected the final foam hardness can cover multiples o f
10X, 20X or even more than 50X! Such wide
extrapolation w ill not be expected to have a high degree o f
fit. Actual IFD or CFD o f foams is a function o f many
variables including sample size and thickness, cell size,
cell openness and anisotropocity. A number o f predicted
points were experimentally reproduced and correlations o f
from 90% to 75% were achieved. See Figure 11. This is
still quite useful for the purpose o f targeting new grades o f
foam via a process and the author believes having such a
model is still better than no model. A significant amount
o f work and time can be saved by fully understanding the
hardness yield relationships rather than starting with a
formulation and density that may not even have the
potential to achieve the targeted goal.

SU M M A R Y

N ew flexible foam formulating technologies normally

take considerable time to develop due to the com plexities
Figure 11. Application Modeling Fit for 3.6 p c f Molding
involved with adjustments for optimum foam stability and
o f Selected Trial Experiments
the development o f basic data such as density and
hardness.

456 / Lockwood
 It has been shown that the use o f an experimental designed BIOGRAPHY
approach to develop an M DI/TDI based flexible foam
technology, is extremely useful in gaining considerable Robert J. Lockwood
knowledge and understanding with a minimum o f Robert J. Lockwood received an
experimentation and time invested. A A S in Chemical Technology
Experimental design in itself clarifies the response trends from the Waterbury State
o f formulation changes, but the m odels in them selves are Technical C ollege in 1969 and a
still quite com plex. It was found necessary to use various B.S. Degree in Chemistry form
normalization treatments and techniques in order to simply the University o f Hartford in 1971
these com plex models. The simple correlation o f hardness - both in Connecticut. After
versus the square o f the density brings great clarity to the serving 13 years at the Donald
data. It has also been shown that the normalization to the S.Gilmore research laboratories
root polymer hardness allow s application modeling o f the Upjohn Company, he worked at ICI Polyurethanes
covering a broad spectrum o f possible applications o f the from 1984 - 1999 and now is with Huntsman Polyurethanes.
formulations. Further normalization o f die formulations Mr. Lockwood ’s work over 30 years has covered all o f the
them selves back to the basic polymer building blocks via main disciplines o f polyurethanes including flexible and
the program FormCalc© brings even more understanding. rigid foams, RIM and microcellular elastomers. His current
Hardness as a function o f the formulation hard block position in Huntsman Polyurethanes is a 'Senior
content allow s much easier extrapolation o f interpolation. Technologist ' in the Fumiture/Slabstock group. K ey areas
Such fundamental basic information is translatable to other o f expertise are a fundamental understanding o f polymer
formulations and base technologies, saving a considerable morphology applied to formulating and predictive
amount o f time in determining appropriate formulation m odeling o f polyurethane technologies. He holds (9) U.S.
starting points. patents.
The potentiality o f M DI/TDI blends to achieve various
flexible foam products was brought to fruition with
experimental design with a minimum o f time and effort.
Huntsman is continuing to develop and explore avenues o f
this new technology in a number o f application areas
including automotive and furniture molded and slabstock
foams. Further experimental designs are in progress to
fully define all the physical properties and processing
attributes o f this new flexible foam technology.

ACKNOWLEDGEMENTS

The author would like to th an k , co - op student, Christopher

A. Arendoski for his diligence in carrying out the design o f
experiments.

REFERENCES

1. Robert J. Lockwood, “Form calc© Formulation

Calculator - A Predictive & A nalyzing Formulator’s
Tool, ” Proceedings o f the October Polyurethanes
Foam Association - 1998.
2. Lee Lawler, “M odels for Foam Formulating,”
Proceedings o f the Polyurethane Foam A ssociation -
1993.
3. D.R.Grier, W .A. Lidy, R.D. Priester Jr.,M.R. Adams,
R.E. O ’N eill, E.G. Rightor, C.G. Barnes, B.L. D avis,
“Fillers, Hard Phases and Copolymer Polyols - Their
Different Mechanism to Reinforce Flexible
Polyurethane Foam,” SPI -Expo 1998 Dallas, Texas,
September 17 - 20, 1998.

Lockwood / 457
 SESSION L
APPUANCE

T E C H N IC A L SE S S IO N O R G A N IZ E R S

David Williams, Chair

HONEYWELL

Jeff King
THE DOW CHEMICAL COMPANY

T E C H N IC A L SE SSIO N M O D ER A TO R

Larry Wethje
ASSOCIATION OF HOME APPLIANCE MANUFACTURERS

T E C H N IC A L S E SSIO N V IC E - M O D ER A TO R

Ken Davis
GENERAL ELECTRIC APPLIANCES
An Assessment of the Performance of Rigid Polyurethane Foam
Insulation for Use in Residential Water Heaters Produced with
SNAP-Approved Non-ODP Blowing Agents

KERRY A. INGOLD, RAYMOND A. YOURD a n d V. MICHAEL MAUTINO

Bayer Corporation
100 Bayer Road
Pittsburgh, Pennsylvania 15205

ABSTRACT addition, the Department o f Energy (DOE) has recently

issued the revised NAECA standards for residential water
In this paper, we have produced laboratory and end - use heaters [1]. These standards will increase the minimum
data for the major SNAP approved blowing agents available required energy efficiency by a range o f 4.43% to 5.18% for
for use in rigid polyurethane insulation foams. These electric and 8.59% to 11.63% for gas residential heaters.
include HFC-245fa, HFC-134a, and cyclopentane. This These are significant increases and, in most cases, will
data includes a comparison o f physical testing results from require each manufacturer re-evaluating and re-designing
laboratory panels and data from EF testing and stand -by loss their products to comply with the new requirements. The
testing (Bayer SBLT) o f heaters produced with these candi­ effective date for the new standard is January 20, 2004. In
dates. Specific results include data on k -factor, hot and cold this paper, we w ill examine the leading candidates to replace
dimensional stability, EF and heat-loss data from the Bayer HCFC-141b for use as insulation material in water heaters.
SBLT. We will do this by review o f data from molded foam panels,
stand-by loss (SBL) and EF test data obtained from 50-gal-
INTRODUCTION lon electric units foamed with each candidate.

The water heater industry has been faced with the chal­ BLOWING AGENT PERFORMANCE IN FOAMS
lenge o f finding an insulation material that will help maxi­ USED FOR WATER HEATER APPLICATIONS
mize the energy efficiency o f its products. The choice for
this insulation for many years was fiberglass. The energy A review o f blowing agent levels, vapor thermal conduc ­
conservation policies o f the 1980 ’s launched a search for tivity and k -factor is presented in Table 1.
new insulation materials. This led to a conversion from The boiling point and vapor thermoconductivity (VTC) o f
fiberglass to R - ll blown rigid polyurethane foams. During each blowing agent varies based on molecular weight and
the past several years, the industry in the United States and composition; with each property having its own distinct
Canada has made the change to alternative blowing agents effect on the processing and performance o f the foam.
for R - ll as required by the Montreal Protocol. HCFC-141b HCFC-141b, HFC-245fa and cyclopentane are liquids at
has largely replaced R - ll as the predominant blowing agent. temperatures above 59 ° F, making the masterbatch blends
The use o f water -blown foam systems in the industry has readily transportable via common containers. HFC - 134a
been limited due to the high k-factors o f open cell water- boils far below room temperature, which makes it a require­
blown and the higher density required for dimensional sta­ ment to transport and store the masterbatch under higher
bility o f typical closed-cell water-blown systems. pressure or add it to the blend via third stream mixing or
The next challenge faced by the industry will be to replace batch blending in the customer facility. HFC - 134a also pro­
the HCFC-141b with a non-ozone depleting material. Many duces a froth foam. This effect causes air entrapment and
new blowing agents are currently being tested for this pur­ uncontrolled pour placement in the heater, which can result
pose, including hydrofluorocarbons, such as HFC - 134a and in voids occurring within the cavity.
HFC - 245fa, and hydrocarbons such as cyclopentane. In Indexed differences in the VTC ’s at 110° F, based on

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 461

 T a b le 1. Blowing Agent Levels and Vapor Thermoconductivity

P r o p e rty H C F C -1 4 1 b L ow H F C -2 4 5 fa H ig h H F C - 2 4 5 f a C -P e n ta n e H F C -1 3 4 a

Boiling Point ( ° F ) 90° 59° 59° 120 ° -1 5 °

B A M o le % in B - side 59% 38% 57% 62% 65%

% B A at M ixhead 8 .7 % 5 .6 % 8 .4 % 4 .7 % 9 .1 %

C alculated N^por T herm al

Conductivity @ 110 ° F
(Btu -in / hr4t2 -° F) 0 .0 7 6 0 .1 0 2 0 .1 0 2 0 .0 8 4 0 .1 1 5

Index for V T C at 110 ° F 1 .0 0 0 1.3 52 1 .3 52 1.117 1 .5 1 8

(H C F C - 1 4 1 b = 1)
110 ° F k -factor for 2 inch Panel
(B tu - in /h r - ft2 -° F ) 0 .1 5 8 0 .1 6 0 0 .1 5 4 0.161 0 .1 6 7

Figure 1. 2 -inch Bosch Panel Foam

k -factor vs. Temperature

HCFC-141b being equal to one, range from a low o f 1.12 for the percent at the mixhead ranging from 9.1 % for H FC-134a
cyclopentane to a high o f 1.52 for HFC - 134a. These differ ­ to 4.7% for cyclopentane.
ences combined with the molecular weight, impact the The k -factor o f the foams was examined at mean temper ­
amount o f blowing agent needed to obtain the targeted k - atures ranging from 50 ° F to 110 ° F. Data for these results
factor. Masterbatch blow ing agent m ole percent levels o f are presented in Figure 1. The k -factors are more clustered
38% for HFC - 245fa to 65% for HFC - 134a were used, with at the 50 ° F mean than at the 110° F mean. This is indica

462 / Ingold, Yourd, Mautino

 Figure 2. 2 -inch Bosch Foam
k -factor vs. Days Aged

Table 2. Machine Processing Parameters

P r o p e rty H C F C -1 4 1 b L ow H F C -2 4 5 fa H ig h H F C - 2 4 5 f a C -P e n ta n e H F C -1 3 4 a

M a c h in e H K -2 4 5 H K -2 4 5 H K -2 4 5 H K -5 5 0 H K -5 5 0

M a te r ia l te m p . ( ° F )
is o c y a n a te 77 70 70 80 70
p o ly o l m a s te rb a tc h 77 70 70 70 70

M o ld T e m p e r a tu r e ( ° F ) 105 105 105 105 105

T o ta l T h ro u g h p u t (g /s e c ) 540 329 525 447 453

(A + B to ta l)

P o u r P r e s s u re s
is o c y a n a te 1700 1500 1500 1500 1500
p o ly o l m a s te rb a tc h 1700 1500 1500 1500 1500

F r e e R is e D e n s ity (p c f) 1 .4 4 1 .4 6 1 .5 5 1 .5 8 1 .5 8

tive o f the boiling points and VTC behavior o f each blowing coefficient o f the foam [2]. HFC-245fa, HFC-134a and
agent influencing the cell gas conductivity component o f the cyclopentane produce fine celled foams. This is due in part
k-factor. The radiation component o f the k-factor is also a to their lower level o f solubility in the foam matrix.
larger influence at the 110° F temperature. The strut frac ­ Changes in k - factor resulting from foam aging also impact
tion, and thus the cell size, directly affects this contribution, the performance o f the foam. The results are presented in
as they have a direct relationship in regards to the extinction Figure 2. Data taken from 2-inch thick molded panel sam-

Ingold, Yourd, Mautino / 463

 Table 3. Physical Test Data for 2 in. Bosch Panels

P r o p e rty H C F C -1 4 1 b L o w H F C -2 4 5 fa H ig h H F C - 2 4 5 f a C -P e n ta n e H F C -1 3 4 a

M in im u m Fill D e n s ity (p c f) 2 .2 3 2 .1 9 2 .2 9 2 .2 6 1 .9 6

P a c k (% ) 5 .4 5 .5 5 .7 4 .9 5.1

P a n e l D e n s ity (p c f) 2 .3 5 2 .3 1 2 .4 2 2 .3 7 2 .0 6
A STM D 1622

k - fa c to r @ 1 1 0 ° F 0 .1 6 0 0 .1 5 8 0 .1 5 4 0 .1 6 3 0 .1 6 6
(B tu in ./ft2 h r ° F )
A STM C 518

k - fa c to r @ 7 5 ° F 0 .1 4 2 0 .1 4 5 0 .1 3 9 0 .1 4 2 0 .1 4 9
(B tu in ./ft2 h r ° F )
A STM C 518

( % ) C lo s e d C e ll 8 7 .8 9 2 .3 9 2 .1 8 7 .1 8 9 .2
ASTM D 2856

P e r c e n t V o lu m e C h a n g e 0.1 - 0 .4 0 .6 -0 .1 0 .2
1 4 d a y @ -2 2 °F
ASTM D 2126

P e r c e n t v o lu m e c h a n g e 0 .6 - 1 .6 - 0 .4 0 .8 1 .2
1 4 d a y @ 1 5 8 °F
ASTM D 2126

pies tested at periods o f 1, 14 and 28 days produced slopes BASIS FOR FOAM SELECTION
that varied from 0.0025 for HFC-245fa to 0.0065 for HCFC-
141b. Foam aging occurs due to the exchange o f cell gas Water blown foams were not included due to the higher k-
between the atmosphere and the foam. Water is used as a co­ factor inherent to the systems. Redesigns o f current produc­
blowing agent in most rigid PU foams. The C 0 2 formed tion units to meet the new NAECA standard for EF mini ­
from the reaction with the isocyanate has a very high diffu ­ mum requirements would require a significant increase in
sion rate through the foam relative to most o f the blowing cavity wall thickness to provide enough insulation to mini­
agents used today. It leaves the foam while the air moves m ize stand - by lo sses. The DOE Technical Support
into the foam, albeit at a much slower rate. The TVC o f the D ocum ent (T SD ) prepared by Environm ental Energy
cell gas increases as this exchange takes place, causing the Technologies D ivision o f Lawrence Berkeley National
k - factor to increase (worsen). Laboratory [3] evaluated a baseline heater in regards to pro­
Foam aging in a water heater w ill not occur at the same ducing equivalent insulation values between HFC - 245fa and
rate as in the cut foam samples sitting in an open room. In open cell water blown foams. The data was generated by the
most cases, a large percentage o f the foam surface will be WATSIM model for electric storage - type water heaters [4].
faced to the steel o f the tank and exterior shell o f the unit. It was determined that a cavity thickness o f 2.12 inches
Where this occurs, the steel facing prevents the exchange o f would be needed using an open cell water blown foam to
gases into and out o f the foam. H owever all heaters have equal the insulation performance o f a cavity thickness o f 1.5
openings at the juncture o f the top and bottom pans and the inches using an HFC - 245fa foam [4]. In addition, the TSD
openings for piping, electrical components and flues. In also reported estimated costs for manufacture associated
addition, the bottom and sometimes the top o f the heater is with the HFC - 245fa and open cell water blown foams,
normally insulated with fiberglass or expanded polystyrene including costs for steel used in the jacket and pans. The
to serve as a foam dam to prevent leakage from the bottom conclusion was that despite a lower material cost for the
pan area. This is another area where the foam is not faced to open cell water blown foam, the overall cost to produce a
the steel, and aging can occur. heater with equivalent insulation performance was signifi -

464 / Ingold, Yourd, Mautino

 Table 4. Physical Test Data for 3 in. Bosch Panels

P ro p e rty H C F C -1 4 1 b L o w H F C -2 4 5 fa H ig h H F C - 2 4 5 f a C -P e n ta n e H F C -1 3 4 a

M in im u m Fill D e n s ity (p c f) 2 .0 8 1 .9 5 2 .0 4 2 .1 8 1 .8 2

P a c k (% ) 4 .8 5 .6 5 .4 5 .0 4 .9

P a n e l D e n s ity (p c f) 2 .1 8 2 .0 6 2 .1 5 2 .2 9 1.91
A S TM D 1622

k - fa c to r @ 1 1 0 ° F 0 .1 6 3 0 .1 6 2 0 .1 5 9 0 .1 6 4 0 .1 6 9
(B tu in ./ft2 h r ° F )
ASTM C 518

k - fa c to r @ 7 5 ° F 0 .1 4 6 0 .1 4 7 0 .1 4 0 0 .1 4 4 0 .1 5 0
(B tu in ./ft2 h r ° F )
ASTM C 518

( % ) C lo s e d C e ll 8 7 .7 8 9 .9 9 2 .7 8 9 .5 8 9 .9
ASTM D 2856

P e r c e n t V o lu m e C h a n g e - 1 .9 - 4 .8 0 .7 0 .3 - 0 .3
14 d ay @ -3 0 °C
ASTM D 2126

P e r c e n t v o lu m e c h a n g e - 0 .3 - 2 .8 0 .6 0 .8 1 .7
1 4 d a y @ 7 0 °C
ASTM D 2126

cantly higher vs. the HFC - 245fa [3]. The cost to manufac ­ regression analysis performed to estimate a minimum fill
ture using closed cell water blown foam would most likely density. Test panels were then made for each system at 5
fall somewhere in between these two options. percent pack over minimum fill.

MACHINE STUDY Test Evaluation Methods

Foam Preparation Panel and sample density was determined according to

ASTM D 1622. Thermal conductivity was measured using a
The processing parameters for the machine work are list ­ Lasercomp FOX 200 heat flow meter at 75 and 110 °F
ed in Table 2. The machine processing evaluations were according to ASTM C 518. Percent (%) closed cell content
conducted using a Hennecke HK - 245 high pressure foam was measured using a Micrometric Accupyc 1330 using a
machine with an MQ 12-2 mixhead for the HCFC-141b and steel ball standard as described in ASTM D 2856.
HFC-245fa system evaluations. A Hennecke HK-550 high- Dimensional stability was evaluated by storage for 14 days
pressure foam machine with an MQ 18 - 2 mixhead was used at 158 ooF / ambient relative humidity and at - 22 accord ­
for the cyclopentane and HFC - 134a due to the H K - 550 ’s ing to ASTM D 2126.
capability o f processing hydrocarbon and gaseous blowing
agents via a Hennecke Pentamat 10 unit. Machine foamed RESULTS AND DISCUSSION
test panels were molded in two different volume Bosch
panel molds with internal measurements o f 200 cm x 20 cm 2-inch Bosch Panels
x 5 cm (2-inch) and 200 cm x 20 cm x 7.6 cm (3-inch).
W hile in the vertical position, the liquid foam is injected into Table 3 shows the physical testing data for the 2 - inch
the mold through the pour hole located 4-inches from the Bosch panels. The 2-inch Bosch mold minimum fill density
bottom. The mold temperature for all test panels was held o f the control HCFC - 141b system was 2.23 pcf. The low
constant at approximately 105°F. A minimum fill density and high - level HFC 245fa and cyclopentane systems had
was determined for each foam system by making three pan ­ similar minimum fill densities as the control HCFC 141b
els having different weights and heights. The volume and system. The HFC - 134a system had the low est minimum fill
density were calculated for each o f the three panels and a density (12% lower than the control) due to the gaseous

Ingold, Yourd, Mautino / 465

blowing agent producing acceptable dimensionally stable HFC-245fa system produced the best k-factors (0.159 at
foam at lower densities. The control HCFC - 141b system had 110° F and 0.140 at 75 ° F). The low - level HFC - 245fa sys ­
a k - factor o f 0.160 at a 110 °F and 0.142 at 75 °F. The high - tem (0.162 at 110 ° F and 0.147 at 75 ° F) and the cyclopen ­
level HFC - 245fa system produced the best k - factors (0.154 tane system (0.164 at 110 ° F and 0.144 at 75 ° F) had com ­
at 110 ° F and 0.139 at 75 ° F). The low - level HFC - 245fa sys ­ parable k - factors to the control. The HFC 134a system
tem (0.158 at 110° F and 0.145 at 75 ° F) and the cyclopen ­ (0.169 at 110 ° F and 0.150 at 75 ° F) was slightly higher than
tane system (0.163 at 110° F and 0.142 at 75 ° F) had com ­ the control HCFC 141b system. N o significant differences
parable k-factors to the control. The HFC-134a system were noted in closed cell content for each system evaluated.
(0.166 at 110° F and 0.149 at 75 ° F) k factors were slightly Except for the higher % volume change values noted in the
elevated as compared to the control HCFC-141b system. No low -level HFC-245fa system, all o f the 3-inch Bosch test
significant differences were noted in closed cell content for panels exhibited 2% or less volum e change in dimensional
each system evaluated. All o f the 2 ” Bosch test samples stability testing in both high heat and cold temperature
exhibited 2% or less volume change in dimensional stability extremes.
testing in both high heat and cold temperature extremes.
DISCUSSION ON HEATER TESTING
3-inch Bosch Panels
Foam k-factor, EF and Stand-by Loss (SBL) Testing
Table 4 shows the physical test data for the 3 - inch Bosch
panels. The 3 - inch Bosch mold minimum fill density o f the The rigid polyurethane foam is the primary insulation com ­
control HCFC - 141b system was 2.08 pcf, a 7% decrease in ponent o f a water heater. The foam covers roughly 90+% o f
density compared to the HCFC - 141b minimum fill density the tank surface in the units used for this study. The thick ­
in the 2 ” Bosch mold. A decrease in density between the 2- ness o f the foam varies from approximately 5/8 inch to over
inch and 3-inch Bosch mold is typical for m ost rigid foam 3-inches in thickness due to the geometry differences
systems due to the foam ’s ability to flow better in a larger between the inner steel tank and the outer steel jacket and
less restrictive cavity. Compared to the control HCFC - 141b end pans. These thickness changes have a strong influence
system, the high-level HFC-245fa had a similar density at on the foam flow during rise, influencing the density and cell
2.04 pcf, while the low -level HFC-245fa had a slightly structure which in turn impacts the physical performance (k-
lower density at 1.95 pcf. The cyclopentane system (2.18 factor, dimensional stability, etc.) o f the foam in the heaters.
pcf) was 5% higher in minimum fill density compared to the The foam is optimized through formulation and process con ­
control HCFC-141b system. The HFC-134a system (1.82 ditions to minimize the effects o f varying thickness on the
pcf) had the lowest minimum fill density in the 3 - inch pan ­ final performance o f the foam.
els, 14% lower than the control HCFC-141b system. In the The EF (energy efficiency - no units) value for a given
3-inch panels the control HCFC-141b system had a k-factor heater is determined by testing the unit using the 24-hour
o f 0.163 at a 110° F and 0.146 at 75 0 F. The high level Simulated U se Test as mandated by the DOE [5]. For the

Table 5. Average EF and SBL Data

Avg. E F D ays E F In d e x B a s e d % D iffe r e n c e Avg. S B L S B L In d e x B a s e d % D iffe r e n c e

S in c e on H C F C - to In d e x D a y s S in c e on H C F C - to In d e x
B lo w in g A g e n t Foam ed 141b C o n tro l Foam ed 141b C o n tro l

H F C -1 3 4 a 4 1 .0 0 4 0 .3 8 % 81 1 .0 7 1 7 .1 2 %
Hi H F C -2 4 5 fa 17 1 .0 1 3 1 .3 1 % 23 0 .9 0 1 - 9 .9 3 %
Hi H F C -2 4 5 fa 21 1 .0 1 2 1 .2 3 % 29 0 .9 2 1 - 7 .9 1 %
Hi H F C -2 4 5 fa 25 1 .0 0 8 0 .8 3 % 35 0 .8 9 8 - 1 0 .2 4 %
Hi H F C -2 4 5 fa 29 1 .0 1 1 1 .0 6 % 40 0 .9 5 2 - 4 .8 5 %
Hi H F C -2 4 5 fa 52 1 .0 0 5 0 .4 7 % 62 0 .9 7 2 - 2 .7 5 %
Hi H F C -2 4 5 fa 55 0 .9 9 7 - 0 .3 1 % 66 0 .9 8 6 - 1 .4 0 %
C - p e n ta n e 13 1 .0 0 1 0 .1 1 % 18 1 .0 2 7 2 .6 6 %
C - p e n ta n e 19 0 .9 9 5 - 0 .4 6 % 24 1 .0 1 2 1 .2 0 %
H C F C -1 4 1 B 40 0 .9 9 0 - 0 .9 8 % 45 1 .0 2 0 1 .9 9 %
H C F C -1 4 1 B * 44 1 .0 0 0 0 .0 0 % 50 1 .0 0 0 0 .0 0 %
L ow H F C -2 4 5 fa 62 1 .0 0 7 0 .6 9 % 71 0 .9 9 7 0 .3 4 %
Low H F C -2 4 5 fa 66 1 .0 0 5 0 .4 9 % 76 0 .9 8 3 1 .6 9 %

* U s e d a s c o n tro l fo r In d e x C o m p a r is o n s

466 / Ingold, Yourd, Mautino

 Table 6. Averages for E F Testing

Avg. E F D ays In d e x o f A v e r a g e
B lo w in g A g e n t E F C o n tro l G ro u p S in c e F o a m e d E F V a lu e % D iff

H i-le v e l H F C - 2 4 5 f a A 19 1 .0 0 0 0 .0 0 %
C - p e n ta n e A 16 0 .9 8 6 - 1 .4 5 %
H F C -1 3 4 a A 4 0 .9 9 2 - 0 .8 9 %

H i-le v e l H F C - 2 4 5 f a B 54 1 .0 0 0 0 .0 0 %
H C F C -1 4 1 B B 42 0 .9 8 5 - 1 .7 7 %

H i-le v e l H F C - 2 4 5 f a C 54 1 .0 0 0 0 .0 0 %
L o w -le v e l H F C - 2 4 5 f a C 64 1 .0 0 6 0 .5 9 %

purpose o f this paper, we examined electric heaters only.

The EF is more heavily influenced by the insulation on these
models than on the gas units. The heater is installed in a
temperature -controlled room (65 ° to 70 ° F) using copper pip ­
ing o f a required length and internal diameter. Insulation or
heat traps (check valves) are used on the piping if it is sup ­
plied with the unit from the manufacturer. Power is supplied
to the heater from a regulated source, as is the incoming
water (tempered to 58 ° F ± 2 ° ) and outgoing water volum e (3
gallons ± 0.3 / minute) during the draws. Six draws are
made; one at the start o f each o f the first six hours o f the test.
Approximately 18 hours o f the test time is spent in what is
called the stand -by mode, meaning that no draws o f water
are being made from the heater.
It is during this period that the foam insulation has the
largest influence on the energy usage o f the heater. Btu / hr F ig u re 3 . O new ay A nalysis o f E F By A verage E F days
losses from the unit are directly related to the % insulation
coverage o f the inner tank and the thickness and type o f
insulation (som e fiberglass and expanded polystyrene is
used as w ell) surrounding the tank within the heater. The
installation o f the heat traps or pipe insulation are also a
major influence on Btu / hr losses.
It is difficult to isolate the contribution o f the foam insula ­
tion to the total Btu / hr losses o f a heater because o f the rea­
sons just discussed. We therefore developed the Bayer
SBLT as a method to m inim ize the non - foam influences on
heat loss [6]. The test is performed by filling the water heater
with hot water (140 ± 2 ° F), removing the inlet and outlet
supply hoses and then capping - off and insulating the inlet
and outlet pipe nipples. A vent is provided to allow the tank
to breathe as the water cools without introducing air into the
main body o f the tank. The internal water is collected for
approximately 24 hours using a Fluke Hydra 2625 data log ­
ger. Six thermocouples are used to monitor the water tem ­
Figure 4 . Oneway Analysis of Btu / hr By Ave SBLT days
perature inside the tank. The thermocouples are placed to

Ingold, Yourd, Mautino / 467

 Figure 5. Oneway Analysis of EF By Blowing Agent Figure 6. Oneway Analysis of EF By Blowing Agent

Figure 7. Oneway Analysis of EF By Blowing Agent Figure 8. Oneway Analysis of Btu/hr By Blowing Agent

create six equal volume zones, the same placement as spec ­ EF and Stand-by Loss Results
ified for the DOE 24 - hour Simulated U se Test [5]. The
ambient temperature is controlled to maintain a range o f 65 ° A total o f 39 heaters were tested during this study. Test and
to 70 °F. The SBLT heat loss rates, Qsblt ’ ^^ are cal foaming equipment availability im posed limitations on the
culated from equation (1): timeline o f the testing. The result was that we were unable
to complete the production o f equal numbers o f units and
specifically control the aging effect on our units in the time
( 1)
span available to us. The number o f units evaluated as w ell
as the span o f time from when each unit was foamed to when
where M is the mass o f water in lb., Cp is the specific heat it was tested varies. In our review and analysis o f the data
o f water at the average water temperature (Btu/lb.° F), and that follow s, w e have used formulation, blowing agent prop ­
msbjt is the slope o f a least squares fit line to the tempera­ erties and foam physical test data discussed earlier as a basis
ture loss curves (°F/hr). Our original test method also to support our conclusions.
included a specially designed tank that greatly reduced non ­ The units used for this study were 50 - gallon electric resi ­
foam heat loss paths. This tank was not used for this proj­ dential heaters produced with the HCFC-141b, low -level
ect, as EF data was needed as w ell for our analyses. HFC-245fa, high-level HFC-245fa, cyclopentane and HFC-

468 / Ingold, Yourd, Mautino

 Figure 9. Oneway Analysis of Btu/hr By Blowing Agent Figure 10. Oneway Analysis of Btu/hr By Blowing Agent

Table 7. Averages for Stand-by Loss Testing

S B L C o n tro l A vg. S B L D ays In d e x o f A v e r a g e

B lo w in g A g e n t G ro u p S in c e F o a m e d B tu /h r % D iff

H i-le v e l H F C - 2 4 5 f a A 26 1 .0 0 0 0 .0 0 %
C - p e n ta n e A 21 1 .1 1 9 1 0 .6 5 %

H i-le v e l H F C - 2 4 5 f a B 38 1 .0 0 0 0 .0 0 %
H C F C -1 4 1 B B 48 1 .0 9 2 8 .4 6 %

H i-le v e l H F C - 2 4 5 f a C 64 1 .0 0 0 0 .0 0 %
L o w -le v e l H F C - 2 4 5 f a C 74 1 .0 11 1 .0 7 %
H F C -1 3 4 a C 81 1 .0 9 4 8 .5 9 %

134a foams. A s stated earlier, the cavity thickness varied another, resulting in an average age (since foamed) o f 17 to
from approximately 5/8-inch to over 3-inches in thickness 55 days. A statistical analysis o f the EF data is shown in
due to the geometry differences between the inner steel tank Figure 3. The mean range covers an EF value o f 0.140. This
and the outer steel jacket and end pans. Each unit was is a - 1.6% change in the mean EF over a span o f 40 days.
foamed under the optimum process conditions specific to The confidence limits shown (95%) reflect the variation
each formulation, as shown in Table 3. In - place foam den ­ present in the EF test and heater construction. The first five
sity averaged approximately 2.0 p cf with the exception o f units do not have statistically significant EF values over the
the L ow - level HFC - 245fa at 1.8 pcf. The testing was con ­ first 53 days. The sixth unit shows a statistical difference,
ducted such that the 24 - hour Simulated U se Test testing (EF) but it is minimal at best. We can use this data to gauge the
was completed and the SBL testing was begun within 5 to 11 performance o f the other foam systems. If you recall, the
days afterward, with the exception o f the HFC - 134a units, rates o f aging for the different foam systems were fairly
which were separated by 76 to 78 days between the two equal with the exception o f the HCFC - 141b foam, which
tests. Table 5 presents the averaged data from the EF and was significantly higher. This allows for us to compare
SBL testing. It has been shown here as indexed values, with heaters at a relatively equal age (since foamed) and assess
the highest HCFC - 141b EF value equal to 1.00. This was whether there is a major difference in EF values between the
done due to the proprietary nature o f the data. foams. A summary o f the EF testing is listed in Table 6.
To begin our discussion, w e w ill examine the high - Level The stand -by losses (Btu/hr) o f these units were analyzed
HFC - 245fa units. These six heaters were foamed and then as w ell. This is presented in Figure 4. Each unit was tested
tested three times per unit, in consecutive order one after three times using the m odified Bayer SBLT, in consecutive

Ingold, Yourd, Mautino / 469

order one after another, resulting in an average age (since C O N C L U SIO N S
foamed) o f 22 to 67 days for SBL testing. The mean SBL
increases approximately 19.2 Btu/hr. This is a 9.5% increase We have examined the three leading blowing agent candi­
in the stand -by losses over a span o f 45 days. The confi ­ dates for replacing HCFC - 141b in rigid polyurethane water
dence limits shown (95%) reflect the small amount o f varia­ heater insulation. EF and stand -by loss values (Btu/hr) for
tion present in the SBL tests, with the a range o f 25.3 Btu/hr. each foam system were obtained using 50 - gallon electric
The first three units do not have statistically significant SBL units tested using the 24 -hour Simulated U se Test and the
values over the first 36 days. However, the next three units, modified Bayer SBLT [5], [6]. Our analysis has shown that
having average aging values o f 40 to 66 days, show a high ­ a significant change in the heat loss rate is needed to impact
er rate o f stand-by loss (Btu-hr). It is interesting to note that the EF values on the units tested. An examination o f the data
nearly a 20 Btu/hr increase in stand -by loss was seen before shows the overall Btu/hr performance follow s the trend o f
an EF change o f any significance was seen. This will be an the 110 ° F k factor data in that the hierarchy o f the best to
important point later in our discussion. Our largest sampling worst insulation is the higher level HFC - 245fa, lower level
o f units was foamed with the high-level HFC-245fa. We will H FC -245fa, HCFC-141b, cyclopentane and H FC -134a
therefore use this as a control for comparison against the respectively. The EF data also showed that the variation
other four foam systems. inherent to the test and to the design o f the units was signif ­
Figures 5 through 7 illustrate the statistical analysis o f the icant enough to subdue the influence o f the foam insulation
EF values by blowing agent. Each o f these data sets com ­ on the results. Comparison o f the stand -by loss data on
pares the higher level HFC - 245fa units to heaters relatively heaters o f similar age supported the panel data in regards to
close in age (days since foamed). Figure 5 reveals that the the k -factor performance the foams. This supports the need
high and low -level HFC-245fa systems eleven days apart in to use the stand -by loss testing as the method o f choice to
age, which did not produce significantly different EF values. evaluate foam performance in residential water heaters.
Figure 6 shows the higher level HFC - 245fa units with a
range o f 24 to 56 days aged compared to the HCFC - 141b ACKNOW LEDGEM ENTS
units that were 39 to 45 days aged. There is a notable offset
in the two means, although they are statistically different by The authors wish to thank Bayer Corporation for the sup ­
the smallest o f margins. Figure 7 compares the HFC - 134a, port o f this work, Kevin M olyneaux, Victoria Taylor and Bill
higher level HFC-245fa and cyclopentane units. The HFC- Tingler for their tireless work on the foam equipment, and
134a unit presents no significant difference relative to the Dr. Michael Blaszkiewicz and Dr. Irene M cGee for their sta­
other two foams. However the higheir level HFC - 245fa and tistical expertise.
cyclopentane foams are quite different, with the former pro­
ducing the higher EF mean o f the two. R E FE R E N C E S
Figures 8 through 10 illustrate the statistical analysis o f the
SBL values (Btu/hr) by blowing agent. Each o f these data 1. Department o f Energy, O ffice o f E fficiency and
sets compares the higher level HFC - 245fa units to heaters R enew able Energy; 10 CFR Part 430, Energy
relatively close in age (days since foamed). Figure 8 shows Conservation Standards for Water Heaters; Final Rule;
the higher level HFC-245fa units and the HCFC-141b units. Federal Register; January 17, 2004.
A significant difference in the means is reported. This cor ­ 2. Beidermann, A., Kudoke, C., Merten, A., M inogue, E.,
relates w ell to the EF data discussed earlier (see Figure 6). Rotermund, U., Ebert, H. R, Heinemann, U., Fricke, J.,
Figure 9 illustrates the higher level HFC - 245fa against the Seifert, H.; “Analysis o f Heat Transfer Mechanisms in
cyclopentane units. Again, a significant difference is shown, Polyurethane Rigid Foam ” ; Polyurethanes ’99
with a spread between the means o f roughly 25 Btu/hr. International Technical Conference and Exposition;
Figure 10 compares the higher level HFC - 245fa units to the September 12, 1999.
HFC-134a and low level HFC-245fa heaters. The HFC- 3. Lutz, James D ., “Energy E fficiency Standards
134a mean is much higher than the other two, with the 95% for Consumer Products: Water Heaters” ; Lawrence
confidence limits being about 14 Btu/hr apart at the furthest B erkeley National Laboratory, Technical Support
point. The HFC - 245fa units (high and low) at not statisti­ Document; April 14, 2000.
cally different, producing means that are just over 2 Btu/hr 4. Hiller, C. C., A. I. Lowenstein, and R. L. Merriam;
apart. An examination o f the data listed in Table 7 shows the “WATSIM ’ S User Manual, Version 1.0: EPRI Detailed
overall Btu/hr performance follow s the trend o f the 110° F k- Water Heater Sim ulation M odel U ser ’s M anual” ,
factor data in that the hierarchy o f the best to worst insula ­ Electric Power Research Institute, Palo Alto, CA Report
tion is the higher level HFC-245fa, lower level HFC-245fa, No. T R -101702; October 1992.
HCFC-141b, cyclopentane and HFC-134a respectively.

470 / Ingold, Yourd, Mautino

5. “Uniform Test Method for Measuring the Energy R aym ond Yourd
Consumption of Water H e a t e r s 10 CFR Part 430
Subpart B; Appendix E, 1990; pp. 54-67. Dr. Raym ond Yourd is a
Technical M anager in the
6. Sm ith, Todd S., Ingold, Kerry A ., B laszk iew icz,
Insulation Department o f Bayer
M ichael, Tully, John M. and W illson, Keith R.; “Bayer
Corporation Stand -by Heat Loss Test Method (Bayer Corporation ’s Polyurethanes
SBLT) for Water Heater Applications ” ; Polyurethanes Division. Prior to joining Bayer
in 1988 as a scientist in the
’99 International Technical Conference and Exposition;
Corporate Polym er Research
September 12, 1999.
group, he received a PhD in
Macromolecular Science from
B IO G R A P H IE S C ase W estern R eserve
University. He joined the Polyurethanes D ivision in 1993 as
K erry A. Ingold a Developm ent Scientist with the responsibility for devel ­
oping rigid foam insulation systems for appliances. In 1997
Kerry A. Ingold is a Technical he transferred to the Polyurethane Marketing group o f Bayer
Project Leader in Application and AG in Leverkusen, Germany where he was responsible for
D evelopm ent,T echnical developing rigid foam insulation system s for a number o f
Insulation Group of the
applications in the European market. He transferred back to
Polyurethanes D ivision o f Bayer
Bayer Corporation into his current position in the summer o f
Corporation. He holds a B. S. in
Chem istry from M iddle 2000.
Tennessee State U niversity,
Murfreesboro, TN.. Kerry came
to Bayer in 1992 as a Project
E ngineer for H ennecke
M achinery and m oved into the A pplications and
Developm ent Group in 1994. He is currently responsible for
all technical projects, technical service and system develop ­
ment relating to rigid PU foam insulation used in the manu ­
facture o f water heaters. Kerry presented a technical paper at
the SPI Polyurethanes EXPO ’96 which reported on the use
o f carbon black fillers and their effect on k -factor in rigid
foam water heater insulation. His paper the EXPO ’99
reported on the development o f the Bayer Stand -by Loss
Test for foam evaluation in water heaters.

V. M ichael M autlno

V. M ichael Mautino has worked

for Bayer Corporation for the past
eleven years and is currently a
Research and D evelopm ent
Specialist in the Polyurethanes
D ivision. During this time he has
worked in the applications and
developm ent area, formulating
and evaluating polyurethane rigid
foam insulation for a variety o f
market applications including refrigerators, entry doors,
vending machines, picnic coolers, and most recently for the
waterheater industry. M ichael has authored and presented
oral and poster presentations at the American Plastics
Industry meetings since 1993. He has been a member o f the
American Chemical Society (ACS), the Pittsburgh Section
o f the ACS, the A C S ’s D ivision o f Chemical Technicians,
and the Western Pennsylvania Technician Affiliate Group
for the past five years.

Ingold, Yourd, Mautino / 471

HFC-134a: The Other HFC

JE FF KING, IKE LATHAM a n d CHUCK MARTIN

The D o w C h em ica l C o m p a n y
2 3 0 1 N. B r a zo sp o rt B lvd., B - 1 6 0 8
F reep o rt , Texas 77541

ABSTRACT

With the elimination o f HCFC-141b in the United while maintaining the k-factor o f the system. This
States over the next few years, appliance manufacturers combination reduces the overall cost o f the system by
have two major categories o f blowing agents to choose greater than 10%. Both o f these improvements make
from as potential replacements. Hydrocarbons offer low foams blown with HFC - 134a more attractive than in the
raw material costs, high capital costs and a moderate to past.
high cabinet energy consumption. A large amount o f
work has been performed outside the United States on INTRODUCTION
hydrocarbon usage, with successful manufacturing
implementation. However, it appears that the Government regulation requires that manufacturing o f
combination o f cabinet energy penalties, capital HCFC - 141b be discontinued in the United States
conversion costs and flammability concerns are currently beginning January 1st 2003. Several alternatives have
detouring appliance manufacturers from choosing this been identified as candidates for replacement o f HCFC -
route for the next blowing agent in the U .S. 141b including: HCFC-22, HCFC-142b, HFC-365mfc,
The second major option is gaseous blowing agents, HFC-245fa, HFC-134a and cyclopentane. Cyclopentane
either Hydrofluorocarbons (HFC ’s) or is widely regarded as the first logical option among the
Hydrochlorofluorocarbons (HCFC’s). The most common hydrocarbons blowing agents.
o f which include HFC-245fa, HFC-365m fc, HCFC-142b, Both the HCFC’s and cyclopentane options for
HCFC-22 and HFC-134a. O f the potential HCFC-141b appliance Original Equipment Manufacturers (OEM ’s)
replacements, many appliance manufacturers have chosen have encountered obstacles to implementation in the form
to use HFC - 245fa since this yields the low est cabinet o f government restriction. United States environmental
energy consumption. The biggest drawback to HFC - regulation targeted at all HCFC ’s will cause HCFC 22
245fa has been the cost o f the material and the perceived and HCFC - 142b to be phased out o f usage by 2010. In
difficulty in passing the increased costs along to the addition, safety regulation has caused capital costs to
consumer in the appliance market. A blend o f HCFC - deter appliance manufacturers from using cyclopentane in
142b/HCFC-22 offers another option with similar k-factor the United States.
and factory costs, but a significantly lower raw material O f the HFC’s, HFC - 365mfc is not available for
cost at the head. The major problem with this blend is its purchase in the United States. This leaves only HFC -
timeline for acceptance. Depending on the E PA ’s latest 245fa and HFC - 134a as realistic alternatives for use.
ruling, the phase out date for these materials could HFC - 245fa has been identified as a solid option for
potentially m ove from 2010 to 2005 in the United States. OEM ’s. Low thermal conductivity, zero ozone depletion
Another option is HFC - 134a, as it produces a moderate potential and good solubility performance have led the
cost part but at a penalty in cabinet energy consumption. push for using this blowing agent. However, HFC -
Recent developments using HFC - 134a provide data that 245fa ’s high cost has caused som e O EM ’s to continue
address these two shortcomings prompting another look at looking for other options.
this potential blowing agent. A new polyol was HFC - 134a has long been identified as a candidate for
developed which significantly improved the thermal the replacement o f HCFC - 141b, but several issues have
conductivity o f foams produced using HFC-134a. The deterred the use o f HFC-134a as such a replacement. As
new polyol is responsible for lowering thermal can be seen in Table #1, the thermal conductivity o f HFC -
conductivity o f the foam from 0.152 to 0.147 btu. This k- 134a is higher than all o f the competing HCFC - 141b
factor improvement could provide near equivalent cabinet replacement options except C 0 2 [1]. This thermal
energy performance to a hydrocarbon based appliance conductivity contributes to a k-factor in a typical HFC-
system without the flammability issue. Another option 134a system that is well above that o f HCFC - 141b or
recently identified is a reduced cost HFC - 134a system. other competing blowing agent system s (Table #2).
This system has both a lower head cost, due to a reduced
level o f the HFC-134a, and a reduced in-place density,

POLYURETHANES EXPO 2001 , SEPTEMBER 30 -OCTOBER 3, 2001 473

 Table #1: Blowing Agent Properties Comparison
H C F C -1 4 1 b H C F C -2 2 H C F C -1 4 2 b H F C -2 4 5 fa C y c lo p e n ta n e C 02 H F C -1 3 4 a
F o r m u la C C I2 F C H 3 C H C IF 2 C H 3 C C IF 2 C F 3C H 2C H F 2 C 5 H 10 C 02 C F 3C H 2 F
B o ilin g P o in t, ° F 8 9 .6 - 4 1 .4 1 4 .4 5 9 .5 120 -1 0 8 - 1 5 .7
O zone
0 .1 0 .0 5 0 .0 6 0 0 0 0
D e p le tio n P o te n tia l
G lo b a l
630 1600 1600 790 11 1 1300
W a r m in g P o te n tia l
V a p o r T h e rm a l 0 .0 7 2 @ 0 .0 7 8 @ 0 .0 8 8 @ 7 7 0 .0 8 3 @ 0 .0 8 6 @ 0 .1 1 4 @ 0.101 @
C o n d u c tiv ity , B tu 77 °F 7 7 °F °F 7 7 °F 7 7 °F 7 7 °F 77 °F
u n its
U. S. Phase 2003 2010 2010 None None None None
o u t/B a n
R a w M a te r ia l C o s t $ $ $ $$$ $ $ $$

The cost for HFC-134a is higher than HCFC-141b, but performance for HFC-134a foamed systems. W hile
it is much more favorable than the anticipated costs for overpack levels less than 6% can be achieved consistently
other HFC candidates. The shift to any o f the alternative in the laboratory, variability in the manufacturing
blowing agents will also carry a need for increased environment requires the safety margin warranted by the
capital. For the HCFC and HFC candidates, this need is additional overpack level.
due to their gaseous nature. Work has been done at The Thermal conductivity analysis was measured using a
D ow Chemical Company to address the deterrents to the LaserComp Fox 200. Test samples (8" x 8" x 1") were
use o f HFC-134a. The resulting advances in HFC-134a analyzed using a upper cold plate and a lower warm plate.
technology will be the focus for this paper. The plates were set to 50/100 °F for a 75 ° F mean
temperature and 15/45 ° F for a 30 °F mean temperature
analysis. K -factors were determined using heat flow
EXPERIMENTAL
meters for the upper and the lower plate. These two
measurements were then averaged for the final mean k-
System development and laboratory evaluations for this
factor measurement. All k -factor measurements in this
paper were done at The D ow Chemical Companies
Polyurethanes R&D facility in Freeport, TX. Foam . . . btu* in
paper are reported in the units — ------------ . These units
processing was done using a Cannon HP - 60 and a Hi ­ F ft *hr*° F
Tech CS - 50 high pressure machine and was foamed at a are abbreviated btu in this paper.
throughput o f 25 to 30 lb/min. Foam samples were Scanning Electron M icroscopy (SEM ) was performed
generated using molds preheated to 45 °C. The majority using a Hitachi S - 4000 SEM. Imaging was accomplished
o f data was gathered from Brett mold samples (200 cm x at 2 keV after samples were coated with a minimal gold
20 cm x 5 cm) which were foamed while in the vertical splutter coat. A ll samples were evaluated at 30x
position. K -factor data were generated from foamed magnification. Cell size was calculated using the
panels (24 " x 24" x 2") using the same foaming equation:
equipment utilized in Brett foaming operations. Freeze
evaluations were performed on all foam samples using a d = ? /(0 .7 8 5 )2 = 110.616 Equation #1
6% overpack sample that was allowed to cure for 24 where:
hours and then frozen for 24 hours at - 30 °C. All d = cell size, microns
samples shown in this article passed a visual inspection t = average cell chord length, microns
follow ing the freeze test, which is consistent with typical

T able #2: B ase System K - F actor R esults fo r A ltern ative B low ing Agents
H C F C -141b H F C -245fa C yclo p en ta n e H F C -134a
P o ly o l B len d “A ” “B " “C ” “D ”
Isocyan ate P A P I™ 6 1 4 6 PAPI 6146 PAPI 6146 PAPI 6146
B lo w in g agent lev e l, pphpp 34.5 4 1 .5 16 .0 2 4 .0
k - factor @ 75 ° F,
0 .1 2 3 0 .1 3 4 0 .1 4 0 0 .1 5 2
B tu .in /ft2.hr. ° F
k - factor @ 3 0 ° F,
0 .1 2 2 0 .1 1 8 0 .1 2 9 0 .1 3 2
B tu .in /ft2.hr. ° F

™ PAPI is a trademark of The Dow Chemical Company

474 / King, Latham, Martin

 Table #3: HFC - 134a Initial Main Effects Screen vs Baseline Blend
Polyol Blend A B
Isocyanate PAPI 6146 PAPI 6146
HFC-134a PPhpp 24 23
Total Blowing cc/g 30.94 30.93
Minimum Fill
pcf 2.00 2.07
Density
Top Panel
btu 0.153 0.151
k-factor @ 75 ° F
Bottom Panel
btu 0.152 0.148
k-factor @ 75 ° F
Demold expansion
in 0.058 0.130
(10% OP at 3 min)

Equation #1 is consistent with ASTM Method D 3576 - this test and shown in Table #3 is the HFC-134a discussed
77. Four samples were tested for every foam evaluated in the Proceedings o f the Polyurethanes EXPO ’99 [1].
using this method. Four chords were measured for each As shown in Table #3, the main effects screening
resulting image giving a total of 16 data points per foam provided a notable improvement in k-factor, but a
sample. The resulting mean chord length was then significant ‘increase in demold expansion. Next we
calculated and reported. moved to a second screening of 5 polyols, again added to
the base blend, but this time individually. In this
RESULTS AND DISCUSSION screening, we were looking for demold improvement
while maintaining or further improving k-factor. The
Low K-factor HFC-134a System results of this testing are shown in Table #4. Both of the
reported systems, polyol C and D, show a marked
One argument for not switching to HFC -134a in the decrease in demold expansion. However, the second
past has been the high k-factors that are delivered in a system (D) shows a high k-factor in the top of the panel.
normal HFC -134a system. From Table #2, we see that For this reason, we proceeded with polyol C.
HFC-134a typically produces a 8.5% higher 75 °F k- Once the components necessary for an improved blend
factor and a 2.3% higher 30 °F k-factor than a typical were identified, we ran one last main effects screen in
cyclopentane system. Similarly, HFC -134a produces a which varying levels of polyol B and C were added to the
13.4% higher 75 °F k-factor and a 11.8% higher 30 °F k- base polyol blend. These were used to identify the
factor than a HFC -245fa system. However, optimal levels of each component in the base blend and to
developmental work has produced a new polyol that test for synergies among the components. Table #5 shows
delivers performance which exceeds past performance. the best system resulting from this foaming operation.
Starting with the base system, a main effects screen was Polyol blend E did a good job of maintaining
done using 3 different polyols at different levels in the properties, but did not provide the desired decrease in k-
base polyol blend. The three polyols were added to the factor. The final system, using polyol blend F, showed
system individually and blended together in a series of 7 the synergy hoped for when the pieces from the first two
runs. The best resulting run is displayed in Table #3 main effects analysis were put together. This provided
(Polyol B). The baseline polyol blend (Polyol A) used in the desired improvement in k-factor. In addition, the

Table #4: Main Effects Screen #2 vs Baseline Blend

Polyol Blend A C D
Isocyanate PAPI 6146 PAPI 6146 PAPI 6146
HFC-134a PPhpp 24 25 24
Total Blowing cc/g 30.94 30.95 30.92
Minimum Fill
pcf 2.00 2.03 2.05
Density
Top Panel
btu 0.153 0.152 0.156
k-factor @ 75 ° F
Bottom Panel btu
0.152 0.150 0.149
k-factor @ 75 ° F
Demold expansion
in 0.058 0.033 0.032
(10% OP at 3 min)

King, Latham, Martin / 475

 Table #5: Component Levels Main Effects Screen Results

Polyol Blend A E F
Isocyanate PAPI 6146 PAPI 6146 PAPI 6146
HFC -134a pphpp 24 25 24
Total Blowing cc/g 30.94 30.93 30.96
Minimum Fill
pcf 2.00 2.08 2.07
Density
Top Panel
btu 0.153 0.151 0.148
k-factor @ 75 ° F
Bottom Panel btu
0.152 0.150 0.147
k-factor @ 75 ° F
Demold expansion
in 0.058 0.137 0.119
(10% OP at 3 min)

change in demold impact was minimized. Polyol blend F Xs= heat transferred by convection
still is characterized by all the benefits of using HFC-
134a: good dimensional stability, acceptable demold and For rigid foams, the convection term in Equation #2 has
lower cost than HFC-245fa. In addition, the k-factor been determined to be negligible due to the small cell
improvement of this system will allow it to be used in size.
applications previously closed to HFC - 134a systems. If we focus on the conductivity of the gas per Chart #1,
we can see that some degradation in k-factor is expected.
Low Cost HFC -134a System In its simplest form, we would expect the contribution of
the gas to be linear according to the vapor thermal
A second factor considered in the decision to use HFC- conductivities of HFC-134a and C 0 2 and the
134a as the blowing agent to replace HCFC-141b is the compositions of each gas in the cell. This relationship is
cost. Because HFC -134a is a higher cost blowing agent shown graphically in Chart #1.
than HCFC-141b, the cost of the HFC-134a system Development work for the new low cost system was
increases at the head. In addition, similar to HFC-245fa, performed using a standard HFC -134a system with a
special considerations and equipment must be used while constant amount of total gas in the foam (cc of gas/gram).
handling this system that have an up front cost of capital The relative amounts of C 0 2 and HFC -134a were varied
associated with implementation. Dow has performed according to the following percentages of C 0 2: 38%,
development work to reduce the cost of the HFC -134a 44%, 50% and 61%. PAPI 6146 and a consistent catalyst
system in an effort to combat the elevated costs associated and surfactant package were used in all four formulations.
with this blowing agent. Results of this foaming operation can be seen in Table #6.
The approach used for this work has focused on As expected, the head cost of the system dropped as
reducing the amount of high cost blowing agent in the more HFC -134a was replaced by C 0 2. This will provide
cell, and substituting this with an inexpensive gas, C 0 2 cost savings to the user estimated at 5.8%. In addition to
formed in situ. The subsequent HFC -134a that is the head cost reduction, we also see two benefits that
displaced in the formulation provides an expected cost were unexpected. First, we see noticeable reduction in
advantage. the density of the system. Second, we see in the
The main concern associated with dropping the level of experimental data that the k-factor in the system did not
HFC-134a in the cell is that the k-factor of the system will degrade as initially expected.
suffer. The k-factor of a system can be described using
the following heat transfer equation [2]: Density Reduction

X = Xs+Xg+Xr+Xc Equation #2 Density reduction in this system will provide a cost

Where: reduction to the user. Using the above data, this
A.s = conductivity of the solid calculates to be a 6.25% density reduction. This
Xg= conductivity of the gas reduction in density is the result of increasing blowing
Xr= heat transferred by radiation

476 / King, Latham, Martin

 Chart 01: Vapor Thermal Conductivity vs Mole Fraction for
H F C - 1 3 4 a /C 0 2 System

efficiency as we raise the level of CO2 in the blend. One Lack of K - factor Degradation
could speculate that improvement in efficiency is due to
the increased heating due to the high reaction rate and Investigation into the reason for the lack of k-factor
exotherm of the additional water in the blend. This degradation shows that our overly simplified theory about
heating would directly affect the gas in the mixture and vapor thermal conductivity does not translate directly into
could explain the better blowing efficiency. a foam system. Other factors influence the conductivity
Another explanation for the improved blowing of foam. First, in our simplified model we have ignored
efficiency is related to the solubility limits of the HFC- any molecular interactions that may be taking place
134a in the formulated polyol. As documented by between the C 0 2 and the HFC -134a in the gas phase of
Latham [3], the solubility limit of HFC -134a in a system the cell.
is dependent on the molecular interactions between the Molecular interactions in the gas phase modify the ideal
polyol and the HFC-134a. In this publication, Latham vapor thermal conductivity when the molecules in the gas
reports solubility limits between 10 and 20% in the B- phase of a system are of different sizes or have strong
side. At the loading level of 17.6% in our base system, interactions that effect the mixture’s ability to transfer
we are close to or even exceeding the solubility limit of heat. When molecule sizes are vastly different, the
our polyol blend. However, as we drop the HFC -134a heavier molecule can impede the efficient transfer of
level, blowing efficiency improves. This improvement energy by the smaller gas molecules. This effect is
can be observed visually for the four runs in Table #6. adequately modeled mathematically using the Wassiljewa
Inspection revealed a noticeable improvement could be equation [4]. The Wassiljewa equation uses a weighted
distinguished in the form of fewer blowholes in the foam sum of the thermal conductivities to calculate the thermal
with rising C 0 2 levels. conductivity of the mixture, k m.

Table 06: HFC- 134a System at Differing Levels of CO2

Polyol Blend A A A A
C02 % 38.58 44.34 50.00 61.30
Total Blowing cc/g 30.94 30.91 30.95 30.94
% HFC-134a of Total
% 17.64 16.48 15.33 12.70
Polyol Side
Gel Time sec 40 42 35 35
Minimum Fill Density pcf 2.08 2.10 2.04 1.95
Demold expansion
in 0.068 0.066 0.074 0.046
(10% OP @ 3 min)
Top Panel
btu 0.153 0.153 0.151 0.152
k-factor @ 75 °F
Bottom Panel
btu 0.151 0.152 0.152 0.152
k-factor @ 75 °F
Normalized Head
1.000 0.985 0.970 0.942
Cost (wet)

King, Latham, Martin / 477

 In addition to the oversimplification in the gas phase differences in the thermal conductivity of the polymer, Xp
model as described in the previous paragraphs, we also [6]. Using Equation #3, Booth substituted the maximum
discounted the solid and radiative heat transfer terms and minimum values from the range of Sinofsky’s data,
when creating our model. Calculating the solid heat resulting in variation of only 0.003 units of k-factor.
transfer term for foam has been well documented. This Although 0.003 btu is significant, in a system with a small
contribution was modeled by L. D. Booth [5] using the change in polymer structure, such as in this case, the error
following equation: in the present calculation should be much less than 0.003
btu.
Xs= (2/3 - Sf/3)( \ - ? ) X p Equation #3 The final heat transfer term required is for the radiative
Where: heat transfer. Calculation of XT is possible using the
Xp= thermal conductivity of the solid polymer following equation:
Sf = fraction of the polymer in the struts
P = porosity of the foam
Equation #6
Equation #3 does an adequate job of modeling the solid’s
thermal conductivity contribution. However, because it is
based on a cubic model of the cell, we have found that it
falls short in some rigid foam systems. Nevertheless Where:
Equation #3’s three building blocks are still accurate and a = Stefan Boltzmann constant
can be broken down into the following relationship: T = mean temperature
Cd= average cell diameter
Xs= F(cs, pf, Xp) Equation #4 pd= thermal conductivity of the polymer
Where:
cs = cell structure of the foam Equation #6 simplifies to Equation #7 when the constants
pf = density of the foam are plugged in and 0.8 is used for Sf as described by L. D.
Booth:
Experimental work allowed modification of Equation #4
into the following form for calculating the solid’s thermal XT= 16 Cd (W/ m2 ° K) Equation #7
conductivity contribution.
K - factor calculation
Xs= .0140 (pf) Equation #5
By substituting values for the foam displayed in Table
Equation #5 holds for similar polymers. In this case, we #6, we can calculate the expected k-factor of the systems.
are assuming that the change in water concentration in our For this paper, we will calculate for the 38% and 61%
foam is not significantly affecting the polymer. This is a cases since these are the most rigorous, in an effort to
reasonable assumption based on Booth's and Sinofsky's explain the lack of k-factor increase in this system.
work. Sinofsky screened 25 different polymers for

Chart #2; Non-ideal Thermal Conductivity vs Mole

Fraction of C 02

478 / King, Latham, Martin

 C hart # 3 ; Cell Size vs C 0 2 Level in a H FC - 134a System

Referring to Equation #2, we have now shown that we calculations for the radiative thermal conductivity
can calculate each of the components that make up the contribution, XTt based on the cell size data in Chart #3 are
thermal conductivity of the foam. The first component is found in Table #7.
the thermal conductivity of the gas, X g. This value is
calculated using the Wassiljewa equation. Based on this C O N C L U SIO N S
equation, when molecular interactions are included, the
straight line that we originally used for our assumption Due to HFC -134a’s k-factor, systems using
becomes the curve shown in Chart #2. Using the new HFC -134a will never reach a thermal conductivity
correlation data, the thermal conductivity contribution for equivalent to that of HFC-245fa. However, new
the gas phase is calculated and can be seen in Table #7. developments in HFC -134a technology have improved
The second component we need to calculate is the A.s. HFC -134a systems in two previously deficient areas.
The thermal conductivity of the solid is calculated by This paper has shown notable improvements in both
substituting the core densities into Equation #5. The HFC-134a system cost and HFC-134a system energy.
resulting values are shown in Table #7. Table #8 shows the final comparison of the new low k-
Finally, we need to calculate the contribution of the Xr. factor system and the baseline systems used in this paper.
To use Equation #7, we need cell size analysis to Polymer development has produced a polyol system
determine the contribution of X T. Cell size analysis was which provides 0.005 btu improvement in the k -factor
performed on the materials in Table #6 using a scanning measured at mean temperatures of both 75 °F and 30 °F.
electron microscope. The results of this analysis are in This quantity is large enough to give a measurable
Chart #3. This analysis clearly shows a drop in cell size improvement in a cabinet’s energy results. Often in the
with increasing C 0 2 levels. This effect is well known to appliance industry, the most important temperature range
occur as the rate of reaction in a system rises. In our set is between these two reported temperatures, as these are
of systems, the rate of reaction increase is reflected by the the operating range of a typical refrigerator. Interpolation
quicker gel time due to the additional water in the system. of the data in Table #8 shows that at the midpoint of 75 °F
Equation #7 shows how this drop in cell size contributes and 30 °F, 52.5 °F, the new polyol system should deliver
to the reduction of k-factor for our system. The final a k-factor within 2% of the k-factor for a cyclopentane

Table #7: Calculated Thermal Conductivity Contributions

38% C 0 2 61% C 0 2 Change in X

Xq 0.1032 0.1057 +0.0025

x$ 0.0291 0.0273 -0.0018

h 0.0218 0.0174 -0.0044

X \o \a \ 0.1541 0.1504 -0.0037

King, Latham, Martin / 479

system and within 8% of the k-factor of a HFC-245fa is still the best replacement for HCFC-141b. However, if
system. These improvements may allow OEM’s to utilize cost is an issue, or if the application can withstand some
HFC-134a in applications previously thought to be too k- sacrifice concerning k-factor performance, HFC-134a is
factor demanding for HFC-134a. In addition, this k-factor now well positioned to fulfill the needs of the OEM.
improvement may ultimately allow reduction of costs for
the OEM via removal of costly equipment in the A cknow ledgem ents
refrigerator that help with cabinet energy efficiency.
Development work has also been done to increase the The authors of this paper would like to extend their
C 0 2content in HFC -134a systems which has resulted in a thanks to the Dow Chemical personnel who assisted in
new system which reduces head cost and density cost. this paper including the test lab personnel for their
The normalized reduction in head costs shown in Table analytical efforts. Special thanks is given to Perry
#6 can be extended to include HFC-245fa for comparison Brooks, Steve Kahla and Robert Kuklies for their time
purposes. In this comparison, the reported 5.8% head cost and energy in producing foam to support this paper.
reduction adds to established cost advantages of HFC-
134a. In addition, a density improvement is realized with R E FER EN C ES
the new polyol system due to the lower loading of HFC-
134a in the polyol and the consequent improved blowing 1. Irwin, P., Latham, D. D., Morgan, R. E., Haworth, J.
efficiency. The resulting 6.3% cost reduction via density and Sinaga, J., “New Developments with Appliance
improvement further positions the OEM in today’s highly Rigid Foam Systems Utilizing Liquid and Gaseous
competitive environment. Blowing Agents to Achieve Optimum Density,
Surprisingly, both of these cost reductions occur Flowability, Cabinet Energy and Product
without any penalty in k-factor. This lack of penalty, Appearance,” Proceedings of the Polyurethanes
though not intuitive, can be explained and calculated. EXPO '99, 1999.
Calculation shows that foam density reduction contributes 2. Ball, G. W., Hurd, R., Walker, M. G., “The Thermal
to the reduction of the thermal conductivity of the solid. Conductivity on Rigid Urethane Foams,” Journal of
In addition, reduced cell size decreases the radiative Cellular Plastics, March/April, 1970.
thermal conductivity. These two terms in Equation #2 are 3. Latham, I., “Factors Affecting the Solubility of
often overlooked, but in this case explain the lack of Gaseous HCFC and HFC Blowing Agents,”
degradation in the overall thermal conductivity of the Proceedings o f the Polyurethanes Expo '96, 1996.
foam. 4. Smits, G. F., Thoen, J. A., “Mathematical Modeling
The head cost reduction and the density reduction add up Techniques to Predict Important Aspect of k-Factor
to a total cost savings of 11.7% compared to the baseline Aging of CFC-11 Free Rigid Insulation,”
HFC-134a foam system. In today’s continued Polyurthanes World Congress, 1991.
commoditization of the appliance marketplace, this 11.7% 5. Booth, L. D., “Radiation Contribution as an Element
reduction in overall raw material costs, without a k-factor of Thermal Conductivity,” Polyurethanes World
penalty, can be a huge asset to an OEM. The new polyol Congress, 1987.
blends developed in this paper provide OEM’s a new 6. Sinofsky, M. and Glicksman, L. R. May, 1983.
choice when selecting blowing agents that fulfill their “Property Measurements and Thermal Performance
needs. If the primary requirement is k-factor, HFC-245fa Predictions of Foam Insulation,” M. S. Thesis, MIT.

Table # 8 : Final Thermal Conductivity Comparison

System % k-factor increase % k-factor increase System % k-factor increase % k-factor
k-factor vs Cyclopentane vs HFC-245fa @ k-factor vs Cyclopentane increase vs HFC-
@ 75 °F System @ 75 °F 75 °F @ 30 °F System @ 30 °F 245fa @ 30 °F
Baseline
Cyclopentane 0.140 - - 0.129 - -
System
Baseline
HFC-245fa 0.134 - - 0.118 - -
System
Baseline
HFC-134a 0.152 8.5 13.4 0.132 2.3 1 1 .8
System
Low k-factor
HFC-134a 0.147 5.0 9.7 0.127 - 1 .6 7.6
System

480 / King, Latham, Martin

B IO G R A P H IE S C huck M artin

J e ff K ing Chuck Martin received his BS

degree in Chemical Engineering
Jeff King received a Masters of and Petroleum Refining at the
Science degree in organic Colorado School of Mines in
chemistry from Abilene Christian Golden, Colorado. He joined
University. He joined The Dow The Dow Chemical Company
Chemical Company in the Rigid upon graduation in December
Foams Research and 1993 and worked for six years in
Development Laboratories in Dow’s Epoxy Products and
1989. He worked on polyol Intermediates business. In
synthesis and new product January 2000 he began working
development until 1993 when he in his current assignment, offering technology
moved to the Appliance Section development improvements to the North American
of the Polyurethanes Technical Appliance Market.
Service and Development Labs. Currently he is leading
Dow's efforts developing technical solutions for HCFC-
141b replacements.

Ike Latham

Ike Latham received his BS

degree in Chemistry from
Washington College in
Chestertown, Maryland and his
Ph. D. in physical chemistry
from Texas A&M University.
After obtaining his degree, he
spent two years as a postdoctoral
fellow at the University of
Southern Mississippi. He joined
The Dow Chemical Company in
1989 as a member of the material
science group in the Polyurethane Product Research Lab.
He is currently working on technology development for
rigid polyurethane foam applications.

King, Latham, Martin / 481

HFC-245fa: Foam Equipment Considerations

J . M. BOWMAN a n d D. J . WILLIAMS
Honeywell
20 Peabody Street
Buffalo, New York 14210

HFC-245fa (1,1,1,3,3-pentafluoropropane) has been IN T R O D U C T IO N

under development by Honeywell as a replacement for
HCFC-141b as a blowing agent for polyurethane and The polyurethane foam insulation industry has utilized
related foam. Production of HCFC-141b for use as a fluorocarbon blowing agents over the years due to a
blowing agent in the U.S. will be phased out January 1, variety of desirable properties that these blowing agents
2003. At that time, many applications will transition to impart to the foam and the application.
HFC-245fa. Most, if not all, of the foam processing Trichlorofluoromethane (CFC-11) was the liquid
equipment modifications must be completed prior to fluorocarbon blowing agent of choice until the phase out
conversion to HFC-245fa. of this material due to its ozone depleting potential
Early third generation blowing agent development (ODP). 1,1-Dichloro-1-fluoroethane (HCFC- 141b)
criteria were to match as closely as possible the chemical became the replacement for CFC-11 as it had similar,
and physical properties of CFC-11 and HCFC-141b. desirable properties as a blowing agent and much lower
HFC-245fa emerged as the preferred HFC because the ODP. The industry was cognizant that HCFC-141b
chemical and physical properties of this molecule were lifetime as a transition fluorocarbon blowing agent was
very similar to CFC-11 and HFC-141b in terms of boiling limited due its ODP, and the production of this material
point, molecular weight, non-flammability, and the ability for use in the United States would be phased out on
to produce foams with superior properties. Despite the January 1, 2003.
similarity of these molecules, prior to conversion to a new Honeywell - formerly AlliedSignal Inc. - began a
blowing agent, any differences in the properties must be research program to identify the third generation
carefully considered. fluorocarbon blowing agent even before HCFC -14lb was
HFC-245fa has chemical properties more analogous to fully commercialized. At that time, the polyurethane
CFC-11 than to HCFC-141b and physical properties industry iterated the desirable attributes of a zero-ODP
bordering a low boiling point liquid. Compatibility of blowing agent that closely paralleled the chemical and
HFC-245fa with conventionally used plastics and physical properties of CFC-11 and HCFC-141b. The
elastomers also must be addressed prior to commercial molecule chosen for commercialization was 1,1,1,3,3-
utilization. Because of these differences in properties, to pentafluoropropane (HFC-245fa) because of these
aid efficacious adoption of HFC-245fa as the replacement properties, along with favorable toxicity considerations,
blowing agent, Honeywell is preparing a conversion and the ability to commercially manufacture this molecule
handbook for use in the polyurethane foam industry. The in a time frame to meet the phase out of HCFC-141b.
conversion handbook will provide recommendations to As a part of Honeywell’s commitment to a smooth
equipment and processing considerations without specific transition to HFC-245fa from HCFC-141b, a compilation
detail to the foam formulation. This paper outlines the of data has been assembled to answer questions and to
information and material that will be detailed in the make recommendations on storage and handling of, and
handbook. processing with HFC -245fa in polyurethane foam
applications. This paper summarizes this data and
provides a road map for this conversion.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 483

P H Y SIC A L A N D C H E M IC A L P R O P E R T IE S aluminum with any halogenated material, which includes
HFC-245fa, due to the reactive nature of aluminum -
HFC -245fa is a zero ozone depletion, non flammable, particularly if aluminum fines are present. This is
liquid fluorocarbon . In addition to HFC-245fa, other especially true if the oxide layer on the surface of the
chemicals are currently being used or are under aluminum is removed from the surface of the aluminum.
consideration for use as zero ozone depletion blowing In general, HFC -245fa is less aggressive toward plastics
agents to replace HCFC’s, especially HCFC-141b. For and elastomers than HCFC-141b. Gaskets and seals that
comparative purposes, the key chemical and physical were changed to accommodate HCFC-141b should be
properties of some of the more common zero ozone compatible with HFC-245fa. Honeywell has evaluated
depletion blowing agents are shown in Table 1. CFC-11 plastics and elastomers for use in the construction of the
and HCFC-141b are included, as these blowing agents commercial HFC-245fa plant in the event that any
were the progression of fluorocarbon blowing agents anomalies in compatibility were observed. Table 2
leading to HFC-245fa. summarizes the findings of this study. Elastomers that
HFC -245fa has many physical properties similar to may find application in both static conditions (for
CFC-11 and HCFC-141b which make HFC-245fa a example gasketing between flanges) versus dynamic
nearly ideal third generation blowing agent. The boiling conditions (for example seals on rotating shafts) may have
point of HFC -245fa allows this material to be handled as varying degrees of suitability in use.
a liquid at room temperature conditions. Because the Prudence and consideration of these three statements
molecular weight of HFC-245fa is nearly identical to will dictate decisions on gasket and seal change out for
CFC-11, similar use levels can be realized in foam use with HFC-245fa:
formulations. HFC -245fa is a non flammable material • What is the risk exposure if the gasket/seal fails
exhibiting no flash point or vapor flame limits, which in terms of safety, environmental, and
obviates the need for flammability considerations in the economics?
handling and processing from a blowing agent • Compatibility with neat HFC­245fa does not
perspective. predict compatibility with a polyurethane premix
The importance of a liquid third generation blowing containing HFC-245fa.
agent stems from the polyurethane industry’s experience • PTFE is always the best choice !
with liquid blowing agents as well as specific formulation
and processing considerations required for gaseous F L A M A B IL IT Y C O N SID E R A T IO N S
blowing agents. HFC-245fa, as well as CFC-11 and
HCFC-141b, is a liquid, which can be seen from Figure 1 HFC -245fa is a non-flammable fluorocarbon liquid,
illustrating the vapor pressure curves of these materials that is, HFC -245fa exhibits no flash point, has no vapor
compared with true gaseous blowing agents. flame limits in air, and has a very high minimum ignition
energy. These properties compared with other common
M A T E R IA L S C O M P A T IB IL IT Y blowing agents are show in Table 3.
Evaluating blowing agent storage vessels and
HFC -245fa is non-reactive and non-corrosive toward all polyurethane processing equipment for flammability
commonly used metals in polyurethane processing hazard and risk exposure is a function of both the
equipment. This includes carbon steel, stainless steel, flammable characteristics of the material and possible
copper and brass. There is a concern with use of ignition sources. Open flames and sparks are clearly

Table 1. Blowing Agent Properties

CFC-11 HCFC-141b HFC-134a HFC-245fa HFC-365mfc Cyclopentane
Formula CCI3F CH3 CCI2 F CH2 FCF3 CF3CH2 CF2 H CF3 CH 2 CF2 CH3 C5 H 10
Molecular Weight 137.4 116.9 1 0 2 .0 134.0 148.0 70.0
Boiling Point (°C) 23.7 32.1 -26.2 15.3 40.0 49.3
Boiling Point (°F) 74.8 89.7 -15.1 59.5 104.0 120.7
Pvap (Bar, abs. @ 2 0 ° C ) 0.883 0.689 5.26 1.23 0.469 0.338
Pvap(PSIA @ 6 8 °F) 1 2 .8 1 0 .0 76.3 17.8 6 .8 4.9
Vapor Thermal Conductivity
(m W /m K ) @ 10°C [1] 7.4 8 .8 12.4 13.5 [2] 10.6 [3] 1 1 .0
Vapor Thermal Conductivity
(BTU in/ft2 hr°F) @ 50°F 0.051 0.061 0.086 0.093 0.073 0.076
Flash Point (°C) [4] None None None None -25 [5] -37
Vapor Flame Limits (Vol. %) None 7 .6 -1 7 .7 None None 3.5 - 9.0 [3] 1 .4 -9 .4
VOC Status (US) Exempt Exempt Exempt Exempt Unpublished Yes
Ozone Depletion Potential 1 .0 0 .1 1 0 0 0 0
Global Warming Potential [6 ]
(C 02 = 1,100 yr) 4600 700 1600 990 [7] 910 1 1 181

4 8 4 / Bowman, Williams
 Figure 1. Blowing Agent Vapor Pressure Curve

ignition sources, however, there are common static HCFC-141b also presents a very low risk of ignition from
electricity sources that are often overlooked. The energy from common static electricity sources. HFC-
magnitude of energy from some common static electricity 365mfc illustrates a material that has a relatively low
sources relative to the minimum ignition energy, at the minimum ignition energy, which generally increases its
lower explosive limit (LEL) for various alternate blowing risk from common static electricity sources.
agents is show in Figure 2. It is evident from this chart
there is virtually no flammability risk with HFC -245fa BULK STORAGE VESSELS
from any of these potential ignition sources, since HFC-
245fa, as well as CFC-11, do not exhibit vapor flame Storage of HFC -245fa will necessarily require a
limits in air and have very high minimum ignition pressure vessel due to the boiling point of HFC-245fa -
energies. HCFC-141b illustrates the example of a 59.5°F (15.3°C) - and the potential for vapor pressure
blowing agent that has vapor flame limits in terms of generation in the tank at elevated temperatures to 120°F
LEL, however has high minimum ignition energy. (49°C). Honeywell engineering recommendations for bulk

Table 2. Materials Compatibility

Plastics
Application Dynamic Static
Polyethylene Satisfactory Satisfactory
Polypropylene Satisfactory Satisfactory
PTFE (Teflon) Satisfactory Satisfactory
PVDF (Kynar) Satisfactory Satisfactory
Nylon Satisfactory Satisfactory
PFA Satisfactory Satisfactory

Elastomers
Application Dynamic Static
Perfluoroelastomer (Kalrez) Unsatisfactory Marginal
Fluoroelastomer (Viton) Unsatisfactory Unsatisfactory
Nitrile Rubber (Buna N) Unsatisfactory Marginal
Neoprene Unsatisfactory Marginal
EPDM Unsatisfactory Marginal
Butyl Rubber Satisfactory Satisfactory
TFE encapsulated Viton Satisfactory Satisfactory
Urethane Satisfactory Satisfactory

Bowman f Williams / 485

 Table 3. Flammability Considerations
CFC-11 HCFC-141b HFC-245fa HFC-365mfc Cyclopentane
Formula CCI3F CH3 CCI2 F CF3 CH2 CF2H CF3 CH2 CF2 CH3 C5 H10
Molecular Weight 137.4 116.9 134.0 148.0 70.0
Boling Point (°C) 23.7 32.1 15.3 40.0 49.3
Boiling Point (°F) 74.8 89.7 59.5 104.0 120.7
Pvap (Bar, abs. @ 2 0 ° C ) 0.883 0.689 1.23 0.469 0.338
Flash Point (°C) None None None -25 [9] -37
Vapor Flame Limits in
Air
Lower, Vol. % None 7.6 None 3.5 1.5 [9]
Upper, Vol. % None 17.7 None 9.0 8.7 [9]
Min. Ignition Energy, mJ >100,000 20,000 [9] >100,000 10.4 <0.1 [51

storage of HFC -245fa are: tank installation. Parameters to be considered should be

• Tank pressure design rating for vapor pressure of consumption rates, available space, economics and
HFC -245fa at 120°F (49°C) plus a 10% safety proximity to supply points.
factor for safety device tolerances. Complete pre-engineered bulk storage systems range in
• Storage vessels should be designed for full size up to 12,000 gallons standard sizes are 6565, 9200,
vacuum. 12000 gallons capacity. These packaged units consist of a
• Carbon steel is the recommended material of storage tank, dual pumps (one of which is a dedicated
construction. spare), level gauge, filters, starter motors, piping,
• Pressurization or inerting of the tank must only completely assembled and shipped via flat bed truck for
be done with an inert gas such as dry nitrogen. installation on a concrete pad. Larger storage tanks will
Air should no be used. require foundations and pillars for support and are
Table 4 illustrates the design criteria for HFC-245fa considerable more involved with respect to labor and
storage vessels and the recommended tank pressure material costs.
rating. All storage tanks should be built in accordance to the
Generally, storage tanks range in size from 6500 latest edition of ASME Code, Section VII, must have a
gallons for tank trailer truck deliveries to 25,000 - 30,000 “U” stamp and must be National Board registered.
gallons for rail car deliveries. Individual manufacturing
locations should be evaluated for the most practical size

Figure 2. Minimum Ignition Energy

486 / Bowman, Williams

 Table 4. Tank Pressure Rating Recommendation
Blowing Agent Vapor Pressure @ 120°F (49°C) + 10% Tank Design Pressure Recommended
(minimum)
HCFC-141b 11.8 psig (81.5 kPa) 50 psig (345 kPa)
HFC-245fa 36.3 psig (25000 kPa) 50 psig (345 kPa)
HFC-134a 188.2 psig (1300 kPa) 250 psig (1725 kPa)

LEVEL MEASUREMENTS construction. The pressure relief device setting should

equal the vessel pressure rating.
A variety of level measuring devices are commonly in
use for determining the level in the storage vessel. Some POLYOL PREM IX STORAGE
of the most common and least expensive are level sight
glasses and float gauges with a % full dial ( Standard 0- From the Honeywell bulk delivery trailer, through the
100% Scale Magnatel ). Use of these is limited to HFC -245fa bulk storage tank, through to the polyol blend
determining the point at which another delivery is tank ( master batch tank ), the material stream is “neat”
required, and not normally used for inventory purposes. HFC-245fa. All handling and processing operations with
An advanced remote monitoring system ( Sentry respect the pressure ratings of vessels, gasket and seal
System™) is available. material compatibility, and all safety precautions relate
directly to the blowing agent - HFC-245fa.
More sophisticated instruments can provide exact Once the HFC -245fa is introduced into the polyol blend
temperature and level measurements simultaneously for tank, any or all of the components of the polyol system
more accurate determination of the number of pounds of side will be present - polyol, surfactant, catalyst, flame
product in the vessel ( MTS Float / Temperature device). retardant, water, etc. Other than the water present, the
Redundant level instruments are not uncommon in the specific compatibility of the other components of the
event one of the unit’s malfunction or fail. polyol blend should be considered separately. Prediction
of material compatibility with gaskets and seals, and
VALVES pressure rating requirement of vessels becomes distinct
from “neat” HFC-245fa. Compatibility of plastics and
Full port ball valves with PTFE ( Teflon ) seats and elastomers in service from the polyol blend system
seals, carbon steel body and stainless steel ball and stem through any auxiliary holding tanks, day tanks, including
are recommended. Gate valves with appropriate packing pumps and valves through to the foam head should be
and adjustable packing gland are suitable for use. evaluated and assessed on an individual basis. It is
unlikely that any polyol system will have system
PUMPS pressures exceeding Honeywell’s design recommendation
of 50 psig ( 345 kPa ) at 120 °F ( 49°C ). Individual
Turbine, side channel multi-stage, or centrifugal pumps operations may chose design criteria less stringent for
with casing pressures to match ANSI 150# class ratings equipment commencing with the polyol blend tank
are used to deliver HFC -245fa to the polyol blend tank. through to the foam head on the basis of the polyol blend
These pumps should utilize PTFE mechanical seals. physical and chemical characteristics, or experience. Risk
exposure of failure is the prudent guidance for decisions
GASKETS relative to this section of the process.

Flexitallic stainless steel spiral wound with PTFE filler BULK STORAGE TANK CONVERSION
are recommended. Other types of gaskets such as W”
Teflon sheet are adequate. Gaskets of other elastomeric 1. Tank Suitability for HFC-245fa
materials should reference the compatibility chart earlier • Set of tank drawings that identify tank and
in this paper. Consideration of the risk exposure should design criteria
the gasket fail is the prudent guidance on these choices. • ASME Data Plate Pressure / Vacuum Rating
• Material of Construction Carbon Steel
PRESSURE R ELIEF DEVICES
2. Tank Inspection External
• Evidence of external corrosion may affect
Storage tanks must be protected against excessive
pressure rating
pressure over the rating of the vessel. A positive sealing
valve or a rupture disc and a spring-loaded relief unit is • Tank shell, tank nozzles , support saddles
recommended. Satisfactory materials for the rupture disc • Ultrasonic thickness gauge to determine shell
include Monel, nickel, stainless steel and silver. Relief corrosion
valves used after a rupture disc can be carbon steel 3. Liquid removal (HCFC-141b) to drums or auxiliary
tank

Bowman , Williams / 487

4. Vapor removal with vacuum pump to 29” vacuum. 4. Service and test safety relief system on the blend
5. Introduce dry air to atmospheric pressure. tank.
6. If internal inspection is planned: CAUTION !! 5. Assess and change pump seals as indicated by
• Tank must be isolated from the process with compatibility requirements.
blinds or disconnected piping. 6. Assess and change agitator seal as indicated by
• Air must be purged through tank to breathing compatibility requirements.
concentration. 7. Inspect polyol blend circulation cooling / heat
• Oxygen detector should be used to determine exchanger.
safe breathing levels of oxygen. 8. Repair / replace all grounding straps.
• Man watch is essential and life line on personnel
entering vessel is required. Polyol Blend systems are in processes that are more
• Other confined space entry requirements may likely to receive inspection and preventative maintenance
exist. Check with the appropriate person in your in a routine fashion. During the conversion from HCFC-
organization prior to any personnel entering a 141b to HFC -245fa is the time for a thorough inspection
tank for inspection. and maintenance procedure to avoid risk of failure in the
7. Change all gaskets to compatible gasketing future.
material. Inspection and preventative maintenance in a like
8. Change all valves as indicated by compatibility fashion should be performed from the polyol blend tank
assessment. through any auxiliary equipment to the foam head during
9. Safety relief valve preventative maintenance and test. the conversion from HCFC-141b to HFC-245fa.
10. Pressure tank to 50 psig and check for leaks.
11. Evacuate tank to 29” vacuum. PACKAGED MATERIALS
12. Bulk storage tank is ready for HFC -245fa delivery.
13. Change pump seal if indicated by compatibility. Packages, such as drums, tote tanks, and small pressure
14. Change all gaskets from pump section to polyol vessels, are currently utilized in applications with HCFC-
blend tank as indicated. 141b where bulk tank trucks or bulk railcars are
15. Inspect and change valves from pump suction to impractical due to size. With the conversion to HFC-
polyol blend tank as indicated. 245fa, segments of the industry will still find commercial
16. Repair / replaced all grounding straps from bulk utility for packages over bulk. These packages may
storage tank to polyol blend tank. contain “neat” HFC -245fa or a formulated polyol blend
with HFC -245fa included in the mix.
Judicious inspection and maintenance of the bulk The over-riding criteria, as with bulk tanks, will be the
storage tank system during the conversion from HCFC- package temperature, through shipping and storage, as
141b to HFC -245fa will minimize the risk of failure in the this will determine the pressure to which the package is
future. Most bulk storage systems have not been subjected. The best practice design consideration is a
inspected and preventative maintenance performed since package that will withstand the vapor pressure of “neat”
the conversion from CFC-11 to HCFC-141b, which in HFC -245fa at 130°F ( 49°C ). Any package less robust
most cases is six to eight years. It is unlikely that an will require specific assessment of the risk exposure to
opportunity to do a though inspection and maintenance of over pressure.
the bulk storage system will occur in most facilities in the Honeywell has identified a drum that meets this design
near term future. criteria for “neat” HFC-245fa - Greif Bros.Corporation
HP Drum. The HP Drum was filled with 525 pounds of
POLYOL BLEND TANK CONVERSION HFC -245fa and subjected to 49°C ( 130 °F ) for four
days. There was no deformation of the drum with an
1. Determine suitability of tank for HFC-245fa. internal pressure of 295 kPa ( 42 psi ). HP Drum
information is contained in Table 5.
• Tank drawings and pressure rating.
An extractor valve system Micro Matic USA, Inc. has
2. Tank inspection internally and externally for
been identified to be used in conjunction with the HP
corrosion.
Drum to meet the safety criteria of opening a drum that is
• Inspect the internal tank liner if one is present for
potentially under pressure. The extractor valve system
repair. CAUTION ! Personnel tank entry safety
incorporates a dry disconnect and locking coupler to
requirements as outlined in Bulk Storage
minimize any hazard in opening the drum. The drum is
Conversion.
factory fitted with a dip tube such that the drum can be
• Determine that HFC 245fa will be introduced
filled or emptied in the upright (vertical) position. Figure
below liquid level in the blend tank.
3 illustrates this assembly in a drum.
3. Assess and change gaskets and valves as indicated by
Bulk pressure vessels design criteria includes “full
compatibility requirements
vacuum” in addition to pressure rating. Small package

488 / Bowman, Williams

 Table 5. Grief Bros. Corp. H P Drum Specification
Criteria Value/Description
Product Identification 55 gallon. Tight head high pressure drum
Construction: Drum Head/Body 1.4 mm steel
Height 35 1/16 inches
Rolling Hoop 7/16 inch
Diameter over rolling hoop 23 7/16 inches
Seams 1.4 mm double seam s
Chime Rings 12 gauge
Shipping Data
Capacity 55 gallons
Weight 70 pounds *
UN Markings UN 1A1/X1.4/500
Pressure rating 600 kPa (90 psi) w/o failure

♦Current drum weight is 42-44 pounds. compatibility data, and a walk through the process
approach for systematic conversion to HFC-245fa.
design may not have this design consideration, which is Honeywell is committed to commercialization of HFC-
important if packages are filled at room temperature, and 245fa prior to phase out of HCFC-141b. Honeywell’s
then subjected to low temperature (frigid winter continued support of the polyurethane industry is a
continuation of this quest for excellence.

DISCLAIM ER

A ll statements, information and data given

herein are believed to be accurate and reliable
but are presented without guaranty, warranty or
responsibility o f any kind, expressed or im plied.
Statements or suggestions concerning possible
use o f our products are made without
representation or warranty that any such use is
free o f patent infringement and are not
Figure 3. Diagram of Micro Matic USA inc. Extractor
Vaive Assmebly in Drum recommendations to infringe any patent. The
user should not assume that all safety measures
are indicated, or that other measures may not be
conditions) shipping and storage. There have been
required.
isolated instances of packaged (drum) collapse due to less
than atmospheric pressure in the package due to low
REFERENCES
temperatures during shipping and storage.
Small cubicle “totes” are under investigation for
designs suitable for this application as the polyurethane 1. 1998 UNEP Rigid and Flexible Foams Technical
industry has adopted these for use with HCFC-141b. It is Options report, Table 1, unless otherwise noted.
likely that both package styles will have some value
attached or be “returnable” for reconditioning and reuse. 2. Geller, Bivens, Yokozeki, ‘Transport Properties and
Surface Tension of Hydrofluorocarbons HFC 236fa
CONCLUSION and HFC -245fa” 20th International Congress of
Refrigeration, IIR/IIF, Sydney, 1999.
With the scheduled phase out of HCFC-141b by
3. Zipfel, et al, “HFC-365: A Versatile Blowing Agent
December 31, 2002, many polyurethane applications will
for Rigid Polyurethane Foams”, Polyurethane Expo
find commercial utility in the use of HFC -245fa as a third
’98, 1998. Note data at 25°C.
generation, zero ODP blowing agent. Major raw material
conversions, such as this, require a conservative, prudent
4. By ASTM D-3828-87 / ASTM D 1310-86 unless
evaluation of the process equipment involved. Honeywell
otherwise specified.
has provided data relative to physical and chemical
properties of HFC-245fa, along with materials
5. Albouy et al, “A Status Report on the Development
of HFC Blowing Agents for Rigid Polyurethane

Bowman, Williams / 489

 Foams”, Polyurethane World Congress ’97, 1997 8. From 1998 UNEP Rigid and Flexible Foams
(method not specified). Technical Options report, Table 1.

6. “Scientific Assessment of Ozone Depletion: 1998” 9. Phillips Petroleum Company, “Nine Light Petroleum
WMO Global Ozone Research and Monitoring Hydrocarbons” 1942B-96 A 01 1996.
Project - Report No. 44, unless otherwise noted.

7. Magid et al, “Atmospheric Lifetime and Global BIOGRAPHIES

Warming Potential of HFC -245fa”, J. Geophysical
Research, Vol. 104, No. D7, pp. 8173-8181, 1999.

J. M. Bowman D. M. Williams

Jim holds a B.S. degree in Chemical Engineering from Dave holds a B.S. degree in Chemistry from the
Iowa State University, Ames, Iowa, an M.Eng. degree University of New Haven, New Haven, Connecticut,
from McNeese State University, Lake Charles, LA, an USA. From 1975 to 1994, he worked as a Research and
MBA from Northern Illinois University, DeKalb, Illinois, Development Chemist and Technical Service
and is a registered professional engineer in the State of Representative for the Upjohn Company and later the
Louisiana. Jim joined Allied Corporation (AlliedSignal, Dow Chemical Company in a wide variety of rigid
now Honeywell) in 1984 and has worked in sales, polyurethane and polystyrene foam application areas.
marketing, and technical service capacities. He is Dave joined AlliedSignal (now Honeywell) in 1994 and is
currently a principal engineer in Honeywell’s Blowing currently the manager of the Blowing Agent Technical
Agent Technical Service and Development Group with Service and Development Group. In this capacity, he is
primary responsibility for appliance industry applications responsible for technical service and product development
of fluorocarbon products. of Honeywell’s line of CFC, HCFC, and HFC foam
blowing agents. Dave is a member of the UNEP Flexible
and Rigid Foam Technical Options Committee.

490 / Bowman, Williams

Development of Amine Catalysts for the Next Generation
Blowing Agents

HIROYUKI KISO, KATSUMI TOKUMOTO, KENNETH M. GAY

YUTAKA TAMANO, HIROYUKI YOSHIMURA TOSOH U.S.A., Inc.
TOSOH Corporation 1100 Circle 75 Parkway, Suite 600
Chemical Research Laboratory Atlanta, Georgia 30339-3097
4560 Kaisei-cho ShinNanyo-city
Yamaguchi 746-8501
Japan

ROGER VAN MARIS

TOSOH EUROPE B.V.
Crown Building-South, Hullenbergweg 359
1101 CP Amsterdam, Z.O.
The Netherlands
for the next generation blowing agents. In contrast to the
ABSTRACT
conventional amine catalysts, the new TOSOH catalyst
systems result in foam properties identical to HCFC-141b
This paper describes the fundamental investigation of
blown foams. It is possible to produce a low density foam
amine catalysts for appliance blown with the next generation
with good properties such as, higher flowability, lower k-
blowing agents.
factor better dimensional stability at low temperature and.
HCFC-141b has been widely used as a blowing agent in
In this paper, HFCs blown systems are discussed from a
rigid polyurethane foams, still it will be phased out by 2002
standpoint of amine catalysts and TOSOH’s new catalyst
in the United States and Europe and by 2003 in Japan due to
systems are introduced for the next generation blowing
its higher ozone depletion potential (ODP).
agents.
Although HFC-245fa and HFC-365mfc (HFCs) have been
proposed as candidates of the next generation blowing agents, INTRODUCTION
their lower solubility in polyols and differences in physical
properties compared with HCFC-141b cause serious Rigid polyurethane foams are widely used in appliance
problems in production of rigid foams, like: and construction applications due to their high performance
(1) Higher thermal conductivity of the foam. Due to the as insulating and structural materials.
high price of HFCs, a higher dosage level of water as co­ In the appliance industry much emphasis is put on the foam
blowing agent is used to reduce the cost. properties such as the thermal conductivity (k-factor),
(2) Lower adhesion on substrates at low temperature. dimensional stability, low-density as well as adhesion
(3) Inferior flowability resulting in a higher density foam. strength.
(4) Inferior dimensional stability at low temperature. 365- In these applications, a number of alternative blowing
mfc is easy to be liquefied at low temperature due to its high agents such as HCFCs, HFCs, HCs, HFEs and all water
boiling point (40°C). blown systems have been proposed and evaluated, since the
(5) HFCs are difficult to handle because of the low boiling regulation of CFCs reduction started. Today CFC-11 has
point (15°C) of245fa and flammability of 365mfc. been eliminated completely and HCFC-141b has been widely
In order to solve these problems, various investigations used as replacements of CFC-11 in the United States and
have been reported from the view of blowing agents and main Japan for the rigid polyurethane foam, however it will itself
raw materials such as polyol. However there are only a few be phased out by 2002 in the United States and Europe due to
reports related to the development of catalysts. its higher ozone depletion potential (ODP). The new
TOSOH Corporation has carried out detailed examinations generations of blowing agents being considered as HCFCs
associated with various catalysts in HFC blown rigid foams replacements are HFC -245fa and HFC-365mfc [1,2,3],
and has successfully developed novel amine catalyst systems however, HFCs blown systems have many problems because

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 491

of differences in physical properties compared with HCFC- based polyol (polyol A), which showed moderate solubility
141b. For example, the thermal conductivity of HFCs (40-60%) of HFCs, was used for HFCs blown foam.
becomes much higher. Due to the high price of HFCs, the
Table 1. Amine catalysts
dosage level o f water as co-blowing agent is increased to
reduce the overall cost of the formulation. Moreover, in the Catalyst Chemical Formula
case of HFCs blown systems, processability and handling
become worse due to the lower solubility in polyol and
different physical properties of HFCs. In HFC -245fa blown MR
systems, it is believed that the adhesive strength decreases
due to the low boiling point (15°C) of HFC-245fa. In 365mfc
blown systems, the dimensional stability at low temperature
TE
worsens compared to HCFC-141b blown systems. 365mfc
can easy be liquefied at a low temperature due to its high
boiling point (40°C). Moreover, HFCs are difficult to handle
because o f the low boiling point (15°C) o f 245fa and the PMA
flammability of 365mfc.
In order to improve the above-mentioned problems,
TOSOH Corporation has evaluated a variety of amine DMCH
catalysts and developed novel amine catalyst systems for
HFC blown foams. These provide a good flowability during
processing and so a low density foam even with good NP
properties such as lower k-factor and better dimensional
stability at low temperature. This paper describes the effects
of amine catalysts on the foam properties from the standpoint
of catalytic activity, i.e. ratio of blowing/gelling activity and DT
the cell structure. In addition, the effectiveness of newly
catalyst systems is discussed when applied to HFCs blown
L33 Triethylenediamine 33wt% in DPG
foams.
DM70 imidazole based amine catalyst
EXPERIM ENTAL

Catalysts and Form ulations RX5

The lists of catalysts evaluated in this paper and then-

abbreviations are shown in Table 1. TOYOCAT -DT and RX7 Specialized reactive amine catalyst
DMCH are standard catalysts in the production of appliance
TMF Specialized amine catalyst
foams. DT is a strong blowing catalyst and DMCH is a
balanced activity between gelling and blowing. TOYOCAT- N,N\ N"-Tris(3-dimethylaminopropyl)-
MR, TE, PMA, and NP also exhibit a balanced activity TRC
hexahydro-s-triazine
between gelling and blowing, especially NP is used as a basic
catalyst to improve k-factor. TEDA-L33 and TOYOCAT- TR20 Specialized trimerization catalyst
DM70 are both tertiary amine catalysts, which have strong
gelling activity. In addition to these catalysts, general amine
Table 2. Formulation for HFCs blown foams
catalysts were evaluated. TMF is a nucleation catalyst having
strong blowing catalytic activity, which improves flowability Recipe-A Recipe-B
and k -factor (4). TOYOCAT-RX5 and RX7 are reactive Polyol1 100 100
amine catalysts. Silicone 2.0 2.0
Table 2 shows the standardized model formulations used to Water 1.8 2.2
examine the HFCs blown appliance rigid foams. Since the HFC-245fa 45 —
HFCs have a lower solubility in polyols compared with HFC-365mfc — 45
HCFC-141b, the polyol was selected carefully. Therefore, Catalyst Varied Varied
the solubilities o f HFCs in different type of polyols were
Isocyanate2 Index = 110 Index = 110
evaluated, see Figure land 2 (next page). As a result, it was
proven that aromatic amine based polyether polyol as well as 1Sucrose/aromatic amine based polyol
aromatic polyester polyol exhibited poor solubilities, while OH value = 440mgKOH/g
aliphatic amine based polyether polyol provided better 2Crude MDI, NCO content =30.8%
solubility of HFCs. In this paper, a sucrose/aromatic amine (Nippon Polyurethane Industry Co.,Ltd.)

492 / Kiso, Tokumoto, Tamano, Yoshimura, Gay, Van Maris

 Table 3. Type of polyols

Sucrose/aromatic amine based polyether polyol

Polyol-A
(OH value = 440 mgKOH/g)
Sucrose based polyether polyol
Polyol-B
(OH value = 451 mgKOH/g)
Sucrose based polyether polyol
Polyol-C
(OH value = 375 mgKOH/g)
Aromatic amine based polyether polyol
Polyol-D
(OH value = 450 mgKOH/g)
Sorbitol based polyether polyol
Polyol-E
(OH value = 451 mgKOH/g)
Aliphatic amine based polyether polyol
Polyol-F
(OH value = 500 mgKOH/g)
Aromatic based polyester polyol
Polyol-G
(OH value = 315 mgKOH/g)

Figure 1. Solubility o f245fa to polyols at 14°C

The foam flowability and the density distribution of the
molded foam were measured using an aluminum mold with
the internal dimensions o f 50 x 50 x 4.5 cm conditioned at
40°C. To evaluate the density distribution in the foam, pieces
were cut from each section (of a total of 9 sections) every 5
cm in height from the bottom. The dimensional stability was
evaluated with ten test specimen from the bottom, middle and
top sections of the panel foam at -30°C and 70°C for 48 hours.
Thermal conductivity (k -factor) was measured on each test
sample at a mean temperature of 25°C using an Anacon
Model 88 apparatus after 24 hours.

Viscosity and Pressure M easurem ent System

The evaluation of viscosity and pressure build up during

foaming has been considered an important parameter for the
development of polyurethane systems as well as quality
control for the production. The viscosity rise profiles were
measured using the viscometer (VM -1A-M, YAMAICHI
DENKI KOGYO Corp.). Foaming pressures were also
Figure 2. Solubility o f 365mfc to polyols at 20 ° C
evaluated simultaneously using a load cell sensor (MR-5K,
JC -500 indicator, KEYENCE Corp.)
Reaction Profiles and Foam Properties
RESULTS AND DISCUSSION
Foam tests were performed using conventional hand
HFC -245fa blown system
mixing methods and in the case of HCF-245fa foam, tests
were performed using a low-pressure foaming machine. In
In general, 245fa blown system offers higher foam thermal
the HFC-245fa blown systems the premix temperature was
conductivity and higher foam density due to a high dosage
adjusted at 15°C and in the HFC-365mfc it was controlled at
level of water as co-blowing agent. 245fa blown system also
20°C. The reaction profiles, such as cream time, gel time, is difficult to handle because of the low boiling point of
tack-free time and rise time, were measured on free rise test in 245fa.
a two liter polyethylene cup at 40°C. The concentration of A variety of amine catalysts were screened in a 245fa
each catalyst was adjusted to give an identical gel time blown foam. Table 4 shows the reaction profiles and foam
(245fa: 60 seconds, 365mfc: 45 seconds respectively) in properties of conventional catalysts. The minimum fill
order to evaluate each catalyst system on the same criteria. density indicates the blowing efficiency as well as flowability

Kiso, Tokumoto, Tamano, Yoshimura, Gay; Van Maris / 493

 Table 4. Reaction and physical properties on 245fa blown rigid foam system

Catalyst MR TE PMA NP DMCH L33 MR/DT DMCH/DT MR/TMF

DM70 RX7
(3/1) ! (3/1) (2/1)
___ (£bw)________________________ 4.80 5.50 5.75 8.40 6.50 5.85 7.60 7.90 3.00/1.00 | 3.54/1.77
Blowing/Gelling Ratio1 I 2.85 ! 2.72 ! 3.05 ! 4.44 I 3.76 I 1.34 I 0.77 I 18.6 I 11.5 I 12.1 14.1
Free rise foam2 I
Reaction profile (sec)
Cream time 11 10 11 11 11 15 16 5 7 7
Gel time 59 57 59 59 60 60 67 60 59 58
Tack-free time 64 69 65 65 72 65 76 67 66 - 66
Rise time 73 76 80 82 77 83 77 83
Core density (ka/m3) 2*1 24.2 24.3 24.7 24.0 25.2 24.0 23.8 24.8 24.1
------------- 1
Panel mold foam3 I L
(pbw) 4.25 j 4.20 4.00 5.90 4.40 4.30 “ 5.40 5.60__ 2.78/0.93 |[ 3.15/1.05 |I 2.77/1.38 I
Reaction profile (sec) i
Cream time 5 | 5 5 5 5 6 5 4 5 4 4
Gel time 57 ! 61 60 55 55 58 57 56 52 58 57
Tack-free time 64 82 67 62 63 65 64 64 60 69 62
Rise time 85 | 84 ______ 80 77 75 76 73 67 79 72

494 / Kiso, Tokumoto, Tamano, Yoshimura, Gay, Van Maris

___ 79 ;
100% packing foam
Minimum fill density (kg/m3) 26.3 25.7 25.7 26.0 26.1 26.8 28.6 26.3 25.6 25.8 25.1
Core density (kg/m3) 23.7 | 22.7 23.7 23.5 23.2 23.7 26.3 23.7 22.3 22.9 21.3
Density distribution (kg/m3) Ii
Max.-Min. 4.8 I 4.1 4.6 5.9 4.3 6.4 4.1 5.1 6.3 5.4 4.9
115% packing foam i
Over all density (kg/m3) 30.3 29.7 29.7 30.5 29.9 31.4 32.3 30.3 29.4 29.3 28.9
Packing ratio (%) 115.1 115.6 115.5 117.3 114.3 117.0 112.9 115.2 114.8 113.7 115.2
Dimensional stability (%)
-30°C x 48hr (Max.) -4.75 -1.11 -1.87 -0.98 -2.77 -10.78 -2.62 -1.11 -9.88 -0.77 ! -7.59
70°C x 48hr (Max.) 1.88 2.46 1.43 2.54 2.99 1.97 0.54 0.9 1.33 1.33 1.97
Compressive strength (kgf/cm2) l
Parallel 2.23 2.28 2.42 2.28 2.14 2.11 2.69 2.34 2.5 1.244 2.17
Perpendicular 1.42 1.45 i 1.53 1.5 1.44 1.77 1.96 1.61 1.44 1.37 1.29
Ratio 1.57 1.57 1.58 1.52 1.49 1.19 1.37 1.45 1.74 0.91 ! 1.68
K-Factor (mW/m-K) 19.2 i 19.9 19.9 18.7 19.2 19.8 19.7 18.7 19.4 19.4 19.9

*0
m

in
o
o

a
a
a
00

1
CO

CM
CM
CM
o

CM

■v
CM

c
CM

c
c

£
CO
T-
T—

c\i
r—
r—

h-

060 |
O
U

!£
E
E
1"-

13
1.28 I
o>

Reaction rate constant ratio : H20-TDI(/ /eq.mol.h)/DEG-TDI(/ /eq.mol.h).

!Free rise foam, mold = 2 liter polyethylene cup.
‘Panel mold foam, mold = 50 x 50 x 4.5 cm.
‘Panel mold foam, mold = 25 x 25 x 8 cm.
o f the foam at identical gel time. As shown by the results, HFC -365mfc blown system
although the use of a strong gelling catalyst, such as TEDA -
L33 and TOYOCAT-MR, provided relatively higher Evaluation o f Conventional Amine Catalysts
minimum fill density foam as well as the distribution in foam
density. There is no correlation between the minimum fill In general, a 365mfc blown system offers a lower blowing
density and the blowing/gelling ratio. L33, MR, DM70 and efficiency and flowability due to the higher boiling point of
DMCH all exhibited a poor dimensional stability at low 365mfc, which results in a higher foam density. Although the
temperature. NP provided a lower thermal conductivity (k- reduction of foam density is an important factor for cost
factor) compared with conventional amine catalysts. It is saving and weight reduction in the appliance industry, low-
shown that this result is contributed primarily to the initial density foam normally provides inferior dimensional stability
increase in viscosity of reaction mixture. This is due to the because of the poor polymer strength. In addition, it is also
relatively high blowing activity of NP and the improvement reported that 365mfc is easy to be liquefied at low
of cure speed as well as polymer strength because of temperature, therefore, the reduction o f foam density in
thermosensitive activity of NP. TOYOCAT-RX7, which is 365mfc blown system has become rather difficult.
special reactive amine catalyst, also exhibits a lower thermal Table 5 shows the reaction profiles and foam properties of
conductivity due to its strong blowing activity. In the case of conventional catalysts in 365mfc blown system. The
blended catalyst systems with the strong blowing amine minimum full density indicates the blowing efficiency as well
catalyst DT present in the blend, showed an improved as flowability o f the foam at identical gel time. As shown by
flowability and a lower minimum fill density foam. The the results, the use of a strong gelling catalyst like TEDA-L33
thermal conductivity became worse. This phenomenon or TOYOCAT -DM70 provides a higher minimum fill density
indicates that the common blowing amine catalyst improves foam as well as a wider distribution in foam density. However,
the isotropy o f the cell structure but increased the cell size, MR also indicated a higher minimum fill density foam, RX5
which resulted in the inferior thermal conductivity. It is and RX7 provided lower minimum fill density foam due to its
previously reported that TMF is a nucleation catalyst and strong blowing activity. As a result, the minimum fill density
improves the flowability as well as thermal conductivity [4]. is strongly correlated to the blowing/gelling ratio (Figure 4).
However, there is no effect on nucleation of bubble in 245fa
blown foam, because a 245fa blown system is processed as a
froth.
As shown in Table 4, the swelling at 120% packing was
also evaluated. The demold time was 5 minutes and the after
expansion level was measured 1 minute after demolding. As a
result, the strong blowing amine catalyst such as DT, TMF
and RX7 indicated low expansion level due to its low
foaming pressure (Figure 3). Especially, RX7 produced the
lowest expansion level among amine catalysts evaluated.

Figure 4. The relationship between blowing/gelling ratio of

amine catalyst and minimum fill density (H F 0 3 6 5 m fc blown
system)

NP and the blend catalyst systems of MR/TMF provided

lower thermal conductivity compared with conventional
amine catalysts. The effect of NP is considered to be due to
Figure 3. Foaming pressure o f amine catalyst systems
the improvement of the cure speed and depress the expansion
(HFC - 245fa blown system)
of cell. On the other hand, for TMF this is considered to be
due to the fine cell structure. This fine cell effect of TMF
contributes to the acceleration of nucleation formation of the

Kiso, Tokumoto, Tamano, Yoshimura, Gay; Van Maris / 495

 Table 5. Reaction and physical properties on 365mfc blown rigid foam system
------------- 1
Catalyst MR NP DMCH L33 MR/DT MR/DT MR/TMF
te DM70 RX5 RX7
(3/1) (2/1) (2/1)
(pbw) 3.50 3.60 6.00 4.50 4.60 5.80 5.30 5.30 2.25/0.75 1.80/0.90 2.60/1.30
Blowing/GellinqRatio1 2.85 __ 2.72__ 4.44 | 3.76 | 1.34 I 0.77 | 15.0 L 18.6 I 11.5 14.3 14.1
Free rise foam2
Reaction profile (sec)
Cream time 13 13 12 12 15 15 6 6 8 8 11
Gel time 46 46 45 45 45 ! 47 45 45 45 44 45
Tack-free time 61 79 80 77 72 ! 77 67 64 86 i 69 64
Rise time 78 80 80 81 82 82 79 78 78 73 79
Core density (kg/m3) 25.6 28.4 24.9 25.3 27.8 27.9 24.2 25.0 — 26.3 26.0 25.0
Panel mold foam3 i
j
100% packing foam
Minimum fill density (kg/m3) 27.3 26.6 26.8 26.6 29.2 29.4 25.9 25.9 26.4 26.5 25.9
Core density (kg/m3) 24.5 24.1 23.8 23.6 26.8 27.2 24.5 24.1 24.6 24.5 24.2
i
Density distribution (kg/m3)

496 / Kiso, Tokumoto, Tamano, Yoshimura, Gay, Van Maris

Max.-Min. 5.3 5.4 4.2 4.4 6.0 6.2 3.6 4.3 4.0 4.0 3.7
Cell size (pm)
Parallel 357 403 338 447 443 464 351 347 423 360 299
Perpendicular 268 294 278 325 278 299 264 247 288 259 240
Ratio 1.33 1.37 1.22 1.38 1.60 1.56 1.33 1.40 1.47 1.39 1.25
115% packing foam
Over all density (kg/m3) 31.5 30.8 31.2 30.1 33.6 34.4 29.9 29.4 30.2 30.7 30.0
Packing ratio (%) 115.5 116.0 116.3 113.2 114.8 116.7 115.2 113.6 114.5 116.0 115.9
i
Dimensional stability (%)
-30°C x 48hr (Max.) -0.89 -1.06 -0.57 -1.13 -0.44 -0.22 -0.99 -1,12 -0.68 -0.79 -1.01
70°C x 48hr (Max.) 0.44 0.22 0.29 0.33 0.88 0.77 0.56 0.56 0.38 0.22 0.56
Compressive strength (kgf/cm2)
Parallel 2.06 2.01 2.39 2.00 2.27 2.34 1.58 1.53 1.17 1.74 1.87
Perpendicular 1.34 0.74 1.37 0.60 0.83 1.13 0.99 0.77 0.84 0.94 0.94
Ratio 1.54 2.72 1.74 3.33 2.73 2.07 1.60 1.99 1.39 1.85 1.99
K-Factor (mW/nrK) 19.6 21.4 19.2 20.7 20.6 20.3 19.8 19.9 20.2 19.9 18.9

r-
Reaction rate constant ratio : H20-TDI(/2/eq.mol.h)/DEG-TDI(/2/eq.mol.h).

CM
'Free rise foam, mold = 2 liter polyethylene cup.

tO
‘Panel mold foam, mold = 50 x 50 x 4.5 cm.
 Table 6. Reaction and physical properties on 365mfc blown rigid foam system
Catalyst DMCH/DT/TRC MR/DT/TRC NP/RX7/TR20 NP/TMF/TR20 DMCH
(2/1/1) (2/1/1) (2/2/1) (2/1/1)
(pbw) 3.20 3.00 6.00 6.00 4.50
I------------------------
Free rise foam1
Reaction profile (sec)
Cream time 7 7 8 7 10
Gel time 45 44 45 46 45
Tack-free time 93 74 89 84 98
Rise time 77 70 71 76 80
Panel mold foam2
100% packing foam
Minimum fill density (kg/m3) 27.2 27.3 27.3 27.2 26.6
Core density (kg/m3) 25.3 25.9 25.3 25.0 23.6
Density distribution (kg/m3)
Max.-Min. 2.8 4.0 3.2 3.7 4.4
114% packing foam
Over all density (kg/m3) 31.3 29.8 30.7 31.1 30.1
Packing ratio (%) 115.1 109.2 112.5 114.3 113.2
Dimensional stability (%)
-30°C x 48hr (Max.) -0.38 -4.79 -0.77 -0.21 -1.13
70°C x 48hr (Max.) 0.63 1.18 0.75 0.68 0.33
Compressive strength (kgf/cm2)
Parallel 2.68 2.46 2.62 2.50 2.00
Perpendicular 1.46 1.34 1.51 1.55 0.60
Ratio 1.84 1.84 1.74 1.61 3.33
K-Factor (mW/nvK) 20.0 19.4 19.6 19.1 20.7
3min-Swelling at 120% packing3
After 1min (mm) 1.24 0.67 2.10 2.02 2.76
After 5min (mm) 1.75 0.48 1.90 1.83 2.90
After 10min (mm) 1.47 0.43 1.71 1.68 2.63
After 24hours (mm) 0.88 | 0.05 0.95 1.01 1.94
Free rise foam, mold = 2 liter polyethylene cup.
Panel mold foam, mold = 50 x 50 x 4.5 cm.
‘Panel mold foam, mold = 25 x 25 x 8 cm.

Kiso, Tokumoto, Tamano, Yoshimura, Gay, V&rc Maris / 497

cells at the initial stage of the foaming. DMCH and TE, which CONCLUSIONS
exhibit balanced activity between gelling and blowing,
showed extremely higher thermal conductivity compared A variety of amine catalysts were screened in HFC -245fa
with the other amine catalysts. It is found that these catalysts and HFC-365mfc blown systems. As a result, TOSOH
increased the cell size, which resulted in the inferior thermal Corporation has successfully developed the novel amine
conductivity. Figure 5 shows the viscosity build up during catalyst systems for the next generation blowing agents. In
foaming. DMCH and TE produced much smoother viscosity contrast to the conventional amine catalysts, with the new
rise profiles than the other catalyst systems. This poor cure TOSOH catalyst systems foam properties identical to
profile is responsible for the large cell structure. HCFC-141b blown system are obtained.
As shown in Table 5, no significant differences were found In 245fa blown systems, TOYOCAT NP and RX7 provides
on conventional amine catalysts in the terms of dimensional a lower foam thermal conductivity (k-factor) to be equal to
stability at temperature. All evaluated amine catalyst systems 141b blown systems. Especially, RX7 also indicated the
exhibited good dimensional stability. lowest expansion level in this evaluation of amine catalysts
due to the low pressure during foaming.
In 365mfc blown systems, NP and the blend catalyst system
MR/TMF provides a lower thermal conductivity compared
with the conventional amine catalysts. While, DMCH and TE,
which exhibit a balanced activity between gelling and
blowing, provides a higher thermal conductivity due to the
increase in foam cell size. In case of optimized blended
catalyst systems, the blend catalyst system o f NP/RX7
provides both superior foam density distribution and lower
thermal conductivity.

REFERENCES

[1] B.Logsdon and D.J.Williams. Proceedings of the Poly­

urethanes EXPO’99, p.421,1999
[2] H.P.Doerge and S.L.Schilling. Proceedings of the Poly­
Figure 5. Viscosity rise profiies o f the balanced activity
between gelling and blowing amine catalysts (HFC -365mfc
urethanes 2000 Conference, p.445,2000
blown system) [3] L.Zipfel and P.Doumel. Proceedings of the Polyure­
thanes 2000 Conference, p.171,2000
Optimization of Catalyst Systems [4] Yoshimura.H., Tokumoto.K., Okuzono.S., and Lowe
D.W. Proceedings of the Polyurethanes 1994 Confer­
The combination systems of the catalysts were evaluated on ence, p.300, 1994
the basis o f the above-mentioned results. Table 6 shows the
reaction profiles and foam properties of optimized catalyst BIOGRAPHIES
systems in a 365mfc blown foam. Trimerization catalysts
such as TRC and TR20 were added in order to improve the H IR O Y U K IK ISO
cure speed and the expansion of foam at demolding. The Hiroyuki Kiso joined TOSOH Corporation in 1993 after
blend catalyst systems NP/RX7 and NP/TMF provided both receiving a B.S. and M.S. degree in Organic Chemistry from
superior distribution in foam density and lower thermal Hiroshima University in 1991 and 1993 respectively. After
conductivity compared with the typical blend amine catalyst development work in organic chemicals, he has been engaged
systems (DMCH/DT/TRC, MR/DT/TRC). in development of amine catalysts and other related materials
The catalyst blend system DMCH/DT/TRC showed a for polyurethanes.
higher thermal conductivity due to the presence of DMCH,
which produced large cell structure. Furthermore, the KATSUMI TOKUM OTO
measurements o f 3min-after expansion showed a two step Katsumi Tokumoto joined TOSOH Corporation in 1986.
expansion with this catalyst blend. MR/DT/TRC indicated a After development work in organic chemical products, he
wider density distribution in the foam due to the MR. studied organic chemistry in Kyusyu University. Currently he
NP/TMF/TR20 provided a lower thermal conductivity than has been engaged in development of amine catalysts and
the blend catalyst systems of NP/RX7 because of the TMF other related materials for polyurethanes
resulting in fine cells.
All catalyst systems produced a low level of expansion, and YUTAKA TAMANO
no significant differences in expansion at 120% packing were Yutaka Tamano joined TOSOH Corporation in 1970. After
obtained in amine catalyst systems. development work in organic chemical products, he has been

498 / Kiso, Tokumoto, Tamano, Yoshimura, Gay, Van Maris

engaged in development of amine catalysts and other related Schering-Plough Health-Care Products as a new product
materials for polyurethane. He is currently the group leader development chemist. Most recently, he worked with Isotec
of development and technical service of amine catalysts. International, a polyurethane systems developer, having
served as R&D director and sales manager. He is responsible
HIROYUKIY OSH IM URA for the support o f the customers not only in Northern USA
Hiroyuki Yoshimura joined TOSOH Corporation in 1986 but also in the Southern American countries.
after receiving a B.S. and M.S. degree in Applied Chemistiy
from Kagoshima University in 1984 and 1986 respectively. Roger van M aris
After development work in organic chemicals, he has been Roger van Maris obtained a MSc degree in Polymer
engaged in development of amine catalysts and other related Technology from the Eindhoven University o f Technology in
materials for polyurethanes. Currently, he is the project 1987. In 1990 he completed a two year research course in
development manager. Polymer Technology at the same university. After working in
R&D and technical support in flame retardancy o f mainly
Kenneth M .Gay flexible polyurethane foams, he joined TOSOH Europe in
Kenneth Gay joined TOSOH USA, Inc. in January 2000 as a April 1998. He is the technical manager for polyurethane
technical sales manager. He graduated from the University of catalysts in Europe, in the Near and Middle East & African
Southern Mississippi with a degree in Polymer Chemistry countries.
and Engineering. After graduation, he worked with

Kiso, Tokumoto, Tamano, Yoshimura, Gay; Van Maris / 499

 Do the Existing Silicone Surfactants Match the Requirements of
HFC-245fa-Blown Appliance Systems?

GEORG BURKHART a n d MANFRED KLINCKE KENNETH WILLOUGHBY

Goldschmidt AG Goldschmidt Chemical Corp.
Goldschmidtstrasse 100 914 East Randolph Road
45127 Essen Hopewell, Virginia 23860
Germany

ABSTRACT to identify the most important requirements on silicone

surfactants in HFC -245fa blown appliance formulations.
The HCFC-141b - the dominating blowing agent for
Information will be provided on questions like: Do we
rigid polyurethane foam in the United States for a number
need emulsification o f HFC -245fa in the raw materials or
o f years now - will be phased out on January 1st, 2003.
in the polymer being formed? How are the typical froth
New formulations and systems with alternative blowing
defects minimized? Can we focus on superior nucleation to
agents for the variety o f insulation foams have to be
achieve the finest cell structure and lowest k-factors?
developed and approved by manufacturers and the
The results o f this work will enable us to quickly
necessary adjustments in the production have to be made
develop surfactants for the new HFC -245fa blown
under a pretty tight schedule. As it is imperative to use the
appliance formulations from the different system suppliers
most efficient insulation material to match the
and appliance manufacturers.
requirements o f the Department of Energy, HFC -245fa has
become an accepted option for the appliance industry in
INTRODUCTION
the US, meeting many o f the additional requirements, also.
Initial screenings with HFC -245fa, including silicone Not only the continuing efforts to identify more
surfactants, were done to have foam products for basic environmentally acceptable blowing agents, as required by
tests needed to prove the principle suitability of HFC -245fa the multiple phase out regulations in the different regions,
as a blowing agent. With the change ahead it is urgent now but also the more stringent energy regulations and cost
to get back for final optimization o f the chemistry and the considerations have forced the appliance manufacturers to
processing. constantly modify their products. This includes the basic
Each blowing agent has its very specific properties, design, the individual components, and the manufacturing
some effecting the final foam properties such as vapor processes. The polyurethane foam insulation is one o f the
thermal conductivity, others being most relevant for individual components and die appliance manufacturers
processing, such as solubility in the raw materials and the demand foam systems which comply with the coming
boiling point. All these characteristics have to be addressed regulations. The directions the industry in the different
in formulating a new system around a blowing agent. regions has taken have been influenced by many factors:
Therefore further system modifications are anticipated
which will influence the choice o f the silicone surfactant. • the phase out schedule for the CFC - and HCFC-
This paper reports on work being done at Goldschmidt blowing agents,

POLYURETHANES EXPO 2001 , SEPTEMBER 30 -OCTOBER 3, 2001 501

• the time when possible alternatives such as HFCs blowing agent to use based on the experience they have
become available, gained using the potential candidates, and their analysis of
• energy requirements, factors considered to be most relevant for the future
• political issues, development.
• cost considerations and so on. HFC -245fa most likely will be the blowing agent for
household refrigerated appliances in the US substituting
When the European manufacturers had to phase out HCFC-141b. Its properties meet many of the requirements
CFC -11 in the early nineties there were strong political o f the industry such as zero ODP, low vapor phase thermal
objections against the environmentally harmful, halogen conductivity, low toxicity, manageable boiling point, and
containing blowing agents. The favorable ecological miscibility with the polyols. The performance of HFC -
properties of cyclopentane - zero ODP and low GWP - as 245fa in a variety o f rigid polyurethane foam systems has
well as the immediate availability and relatively low cost, been investigated [4-16] including an initial screening of
were higher valued than the lower gas thermal conductivity
some available silicone surfactants [17]. Appliance foams
and flammability and the better solubility of HCFC-141b
could be produced with existing machinery in an already
in the polyols. The logical consequence was the
acceptable quality and good energy performance [11, 16]
modification of the refrigerator factories to safely handle
with a processing window that “has not shrunk radically in
the flammable cyclopentane and the introduction of
comparison with HCFC -141b systems” [13].
cyclopentane blown systems as they could meet the
This paper reports on work Goldschmidt has done
environmental and energy legislation of the European
addressing the critical issues o f HFC -245fa blown
market. In the meantime the initial drawbacks of
appliance systems with silicone surfactants. As the
cyclopentane blown foams such as higher k-factors
compared to the 50% reduced CFC -11 systems, higher development o f those systems is still in the initial stage and
foam densities, and reduced cycle times could be appropriate systems for a surfactant optimization were not
minimized by coblowing techniques and continuous available at the time when this study was carried out, we
optimization o f foam flow properties and foam structure had to use formulations which were originally developed
[1,2,3]. for other blowing agents. These systems were first foamed
In the US the home appliance industry is of a different with a variety of surfactants including the specified type
structure and also the political priorities were quite and with the specified blowing agent, also. Then we
different when the use of CFC - 11 was prohibited. There replaced the blowing agent by an equivalent amount o f
was a much higher concern about an increase in the energy HFC -245fa and again varied the silicone surfactant. The
consumption of the units, the flammability of cyclopentane tests give a clear indication on how sensitive the
and the necessary plant adjustments than in Europe and formulations are concerning the change of the blowing
therefore HCFC-141b became the preferred blowing agent. Additionally it becomes obvious what type o f
option. With that, another change was accepted as the surfactants fulfill the requirements of HFC -245fa blown
phase out scenario for HCFC-141b which still had some foams best.
significant ODP, was already determined with the end of
2002. Consequently next to HCFC-141b the first CRITICAL PROCESSING ISSUES • • •
generation o f available HFCs, all showing zero ODP, has
found acceptance in the US, namely HFC-134a, despite of Before outlining a development concept for a silicone
its low boiling point. The most promising candidate to stabilizer for a HFC -245fa blown appliance formulation
replace HCFC-141b when it will be phased out is HFC- the critical parameters o f the whole process, especially in
245fa. comparison with the existing technology, have to be
The elimination o f blowing agents, CFCs and HCFCs, identified. Then the effect o f the surfactant on each critical
showing some ozone depletion, in other countries around parameter has to be valued and finally it has to be agreed
the world is underway and there is still quite some on an acceptable compromise. O f course there are some
uncertainty regarding the choice of the next blowing agent. requirements always high on the list of priorities but the
What makes the choice so difficult is the large number of ranking indeed depends to a certain degree on the
selection criteria that have to be considered. Finally it will individual OEM or on the system supplier.
depend on the phase out regulations the governments have The critical properties o f a blowing agent in the
agreed upon and those are mandated by environmental processing o f an appliance formulation which have to be
considerations. Within this given frame the OEMs considered when choosing the components for a
(Original Equipment Manufacturers) will decide which formulation are:

502 / Burkhart, Klincke, Willoughby

1. the boiling point as it determines the pressure build up and emulsification o f incompatible components in the
in the tanks and in the pipes and more important premix, in the A and B blend, and in the expanding
whether the system frothes after pouring, polymer, a precondition for a smooth transition o f the
2. the heat o f vaporization as it effects the temperature various phases into a homogeneous polymeric material.
profile o f a foaming process,
The degree of nucleation is a function of the surfactant as
3. the molecular weight and the specific gravity as they
well as the stability of the nuclei, both most important for
is are decisive for the amount of blowing agent
required in the system to achieve the desired density, the final cell structure. The third function o f the surfactant
4. the solubility in the B-side and in the total blend as in a rigid foam system is the stabilization o f the gas
insoluble blowing agents require equipment and bubbles during all the stages o f the foaming process [18].
formulation adjustments and
5. the solubility in the polyurethane matrix as better EXPERIM ENTAL
soluble blowing agents cause a softening of the matrix,
what has to be compensated for with a higher density. The number of experiments we carried out in our study
had to be limited to a sometimes unjustifiable low value as
O f course ODP and GWP, vapor phase thermal the available quantity o f HFC -245fa was so low.
conductivity, toxicity, flammability, and chemical as well In our study we used preformulated appliance systems
as thermal stability of the blowing agent are of outstanding which might be considered as being close to actual systems
importance for the choice o f a blowing agent, but these in the US and the European market. The systems are
properties don’t effect the foaming process as such. originally blown with HCFC-141b, cyclopentane or HFC-
134a. Further details are shown in Table 1.
• • • AND HOW SILICONE SURFACTANTS CAN
SUPPORT?
Com patibility Tests
Now, what are the main functions of silicone surfactants
First of all we evaluated the maximum solubility o f
in rigid polyurethane foam manufacturing and how can
HFC -245fa in the polyols, as they were provided. In a
silicone surfactants be used to optimize the processing of
second series we added the surfactant and the amount of
HFC -245fa blown appliance systems?
HFC -245fa, corresponding to the amount of blowing agent
Silicone surfactants are key ingredients in rigid foam
in the original formulation. Our interest was not to identify
formulations as they control the surface phenomena during
the maximum solubility o f HFC -245fa in this fully
mixing and foam expansion until the material finally sets.
The ability to do so is principally based on their surface formulated B -side we only wanted to know, if the
activity and their physical interactions with the chemical calculated equivalent was soluble.
structures of the raw materials and the polymer being The polyol, respectively, the polyol/surfactant blend
formed. For that reason the molecular structure o f a (surfactant concentration as used in the original system)
silicone surfactant has to be adjusted to the characteristics was mixed at lOOOrpm for 30sec. 50g of this blend were
of the individual foam formulation, especially to the nature filled into a pressurizable glass tube equipped with a
of the polyol and to the type o f blowing agent, also. The pressure gauge. The calculated amount o f HFC -245fa was
control of the surface phenomena within a foaming process added, homogenized by just shaking and allowed to
implies the control of the cell structure and cell stability equilibrate at ambient temperature. After 30min the
and all related properties such as k-factor, flowability, pressure was recorded (pressure 1) and the material
density distribution, void formation, aging, and so on. allowed to clear up further. A second measurement was
Silicone surfactants indeed lower the surface tension of done when the blend was totally clear or separated
the different raw materials and therefore support mixing (pressure 2).

Table 1: Appliance Foam S ystem s Used in the Study

System 1 System 2 System 3 System 4
Original BA Cyclopentane HCFC-141b HCFC-141b HFC-134a
Original BA Level 16 pphp 31,5 pphp 39 pphp 19 pphp
Conversion Factor 1,914 1,146 1,146 1,314
HFC-245fa 31 pphp 36 pphp 45 pphp 25 pphp
Surfactant 3 pphp 2 pphp 2,5 pphp 3 pphp
Crude MDI 169 pphp 140 pphp 180 pphp 145 pphp

Burkhart, Klincke, Willoughby / 503

Foam Tests based on the minimum fill.

In a first foaming series we used the original blowing PROCESSING AND FOAM PRO PERTIES
agent and surfactant including some surfactants that cover
a wider spectra o f chemical structures. Then we replaced Flowability
the blowing agent by an equivalent amount o f HFC -245fa
The amount o f material poured into the lance mold was
and repeated the trials. The intention was to learn whether
adjusted to allow an approximately 80%-90% filling. The
the best performing surfactant in the original system after
lengths of the panels are measured and the variation from
replacing the blowing agent with HFC -245fa is still the
a standard is given in percent. The changes were always
superior product and also in which direction we have to calculated back to normal conditions (760 mmHg) and the
move to find the best performing stabilizer. same amount of material.

Handmix Procedure Density Distribution

For formulations blown with liquid blowing agents, e.g. The density distribution was calculated by a statistical
HCFC-141b and cyclopentane, a standardized procedure equation. It represents the sum of the density differences
was used to prepare handpours: raw materials are (taken from 13 samples each 10cm x 14cm x 3.5cm from
temperature controlled at 22°C; weighing the required the overpacked foam board) related to a foam volume o f
amount o f polyol component, including the catalysts, in a 1000 ml. The resulting figure should be as low as
550ml paper cup; adding the necessary amount of possible.
blowing agent and silicone surfactant; homogenizing the
components at 1000 rpm for 30 sec with a special plate Therm al Conductivity
stirrer; controlling/adding the blowing agent loss; adding
the necessary amount o f isocyanate; stirring A- and 13- To determine the thermal conductivity of the produced
side for 5sec at 3000rpm. foams, four test specimen were cut from the lance mold.
By taking four samples we get a first indication about the
High Pressure M achine Trials k-factor distribution along the foam length. The data
shown in the Tables are averaged numbers. The thermal
All foams blown with gaseous blowing agents were conductivities of the samples were measured with a
evaluated on a modified Edge-Sweets-High-Pressure- Hesto-Lambda-Control A 50 Heat Flow Meter (sample
Bench-Top three component foaming machine. The tanks size for this study was always 20 cm x 14 cm x 2.5 cm)
for the raw materials (polyol component and isocyanate) with a mean temperature of 23°C (gradient 36°C /1 0 °C).
are temperature controlled, pressurizable and equipped Initial k-factors were measured one day after foaming, the
with an agitator. The silicone surfactant was fed to the aged k -factors after six weeks o f storage at 25°C and 50%
mixing head as a separate stream. The tanks were kept humidity.
under 90psi nitrogen pressure to make sure the blowing
agent stayed in solution. After 45min conditioning time of Compressive Strength
the B -side calibration of all three streams was done.
The compressive strength, perpendicular to the rise of
the foams, of at least three foam samples taken from the
Pouring
overpacked lance mold were measured at 10%
The foams were poured into either an open box (27cm x compression with a Zwick 1445 machine. An averaged
27cm) or into a temperature controlled (45°C) lance mold number will be given.
(145cm x 14cm x 3,5cm) that was sprayed with a high
solid release agent. The parts were demolded after 15min. RESULTS AND DISCUSSION
Free rise foams were poured to determine the reactivity Compatibility/Solubility of HFC - 245fa
of the system and the free rise density and also to get an
idea about the stability and frothing behavior of the The compatibility/solubility of the different
system. components in a foam formulation is extremely important
The molded foams were used to determine and compare for both the processing and the final foam properties.
the flowability and the minimum fill density of the foam Systems based on polyols or additives designed to
systems as well as the k-factors, compressive strength, improve the overall compatibility of a foam formulation
core density, density distribution, and void formation. The allow to make foams with a significantly reduced number
degree o f overpacking was 10 % which was calculated of voids and a superior flowability and density

504 / Burkhart, Klincke, Willoughby

distribution [19,20]. As the raw material and system summarized. More than lOOpphp of HFC -245fa are
producers have their individual preferences in adjusting soluble in system 2, originally blown with HCFC-141b,
the compatibility of their polyols to a certain blowing followed by 88pphp in system 4, originally blown with
agent without loosing other essential processing and foam HFC-134a. About the same amount of HFC-245fa
characteristics, differences in the polyols can be expected. dissolves in the systems 1 and 3, roughly 45pphp. There is
How wide the spectrum of polyol adjustments actually is, a dramatic difference in the solubility o f HFC-245fa in the
was demonstrated when alkanes were first used as system 2 and 3, although both are developed for HCFC-
blowing agents. It was possible to achieve a sufficient 141b. Again, this confirms that there are many
solubility o f cyclopentane in the polar rigid foam polyols possibilities to formulate a polyol around a given blowing
without sacrificing too much in foam properties. agent and that there is quite some room to address the
Silicone surfactants are often considered as required processing and physical properties of the foam.
“solubilizers”, i. e. products that improve the solubility of The pressures in the glass tubes at the maximal soluble
insoluble materials. Our experience as well as literature HFC -245fa level were between 3.0 and 8.0psi. In all four
data don't confirm this. Surfactants indeed cause the systems the amount o f HFC -245fa, corresponding to the
opposite, e.g. they reduce the solubility of blowing agents amount of the original blowing agent was clearly soluble
in some rigid foam polyols or the solubility of water in and even with the different surfactants in the systems only
flexible foam polyols [21]. The real reason to use silicone with system 3 the solubility was not given anymore if
surfactants is the remarkably finer and more stable more hydrophobic surfactants were added. The pressure
emulsion that will be achieved above the solubility limits values are generally higher in the blends containing the
and those emulsions process similar to a real surfactants, between 4.5 and 22.0psi, although the amount
homogeneous system. of HFC -245fa is lower than in the first series. This clearly
Because the solubility of the blowing agent is so critical demonstrates that all surfactants reduce the solubility of
in the foaming process a lot of solubility data has been HFC -245fa in the different systems. There is no obvious
published, e.g. in [6,10]. Despite the low boiling point the correlation between the pressure data and the character of
solubility in most rigid foam polyols is quite good and the the silicone surfactants or the maximal solubility in the
pressure built up is acceptable. different systems.
In Table 2 the results o f our solubility study are The pressure data also indicate that a froth will be

Table 2: Results of the HFC-245fa Solubility Study

Max. solubility HFC-245fa- Surfactant 0) CO CO

Surfactant D

Surfactant E

Surfactant F
o f HFC-245fa equivalent level c c
Appliance without a a
(se e Tablel) o
foam system surfactant |
? sr
3 3
*+ 3.
[pphp] [pphp] [pphp]
> CD o
System 1 43 31 3
Appearance1’ cl cl cl d d d
Press. 1 [psi] 21.7 8.7 11.6 18.8 17.4 18.8
Press.2 [psi] 7.3 18.8 7.2 10.1 15.9 14.5 11.6
System 2 >100 36 2
Appearance1} cl cl d d d d
Press. 1 [psi] 13.0 15.9 5.8 21.7 5.8 8.7
Press.2 [psi] 5.8 11.6 14.5 5.8 17.4 5.8 7.2
System 3 45 45 2.5
Appearance1) cl d d tu tu sl. tu
Press. 1 [psi] 7.2 8.7 21.7 20.3 15.9 13.0
Press.2 [psi] 4.4 7.2 7.2 20.3 18.8 15.9 14.5
System 4 88 25 3
Appearance1} cl d d d d d
Press.1 [psi] 7.2 18.8 5.8 5.8 5.8 5.8
Press.2 [psi] 2.9 5.8 15.9 4.3 4.3 4.3 4.3

1) cl: clear; si. tu: slightly turbid; tu: turbid

Burkhart, Klincke, Willoughby / 505

formed when the pressure is released and that surfactants stabilization. The hydrophobic character of the surfactants
might even reinforce this. increases from surfactant A to surfactant F.
Under real foaming conditions the situation is of course
more complicated and more difficult to quantify. In the System 1
final foaming mixture, including the isocyanate, the
As indicated in Table 1 the appliance system 1 is
compatibilities will be much different and really dynamic
originally blown with 16pphp cyclopentane and stabilized
as the temperature and the polarity changes in the course
with 3pphp surfactant. The cream time, string time and
of the expansion. Additionally the transition of the
tack free time are >5sec, 21 sec and 45sec. The processing
blowing agent from the dissolved stage into the gaseous
and foam data are summarized in Table 3. Flowability is
phase takes place now and surfactants play a critical role
shown in percentage relative to surfactant D as its
in this transition [22].
character is similar to the surfactant being used in the
In general the solubility of the blowing agent in the
original formulation. Up to a certain extent the
polyol is more important for the process whereas the
flowability, as well as the density distribution and the k-
solubility o f the blowing agent in the final mixture and
factors improve with the more hydrophobic surfactants
how this is effected by the silicone surfactant is much
but the surfactants E and F are already insoluble in the
more critical for the foaming result.
polyols.
Replacing the cyclopentane by the equivalent o f
FOAM TESTS
31pphp o f HFC -245fa the data as shown in Table 4
Foam tests were done with the original systems blown resulted. Surfactants E and F were remarkably good in
with HCFC-141b, HFC-134a and cyclopentane and with flow resulting in the lowest overall densities and density
the equivalent amounts of HFC -245fa as well. The distributions as well as low k-factors. Almost as good in
silicone surfactants were chosen to cover a broad range of flow was surfactant A but the aged k-factors were much
possible structures especially considering our experience worse because of a coarser and more irregular cell
in emulsification of blowing agents and froth structure. Compared with the cyclopentane the HFC-

Table 3: System 1, Blown with Cyclopentane - Processing and Foam Data

Free rise Relative Overall Density C om pressive K-factor K-factor

Surfactant density flow density distribution strength perp. initial aged
type [BTU-inchl 'BTU-inchl
[kg/m2] [%] [kg/m2] [P«i]
h -fi-° F \

Surfactant A 21.1 -5.0 41.8 1.73 16.0 0.149 0.166

Surfactant B 21.9 -2.0 39.8 1.50 14.8 0.143 0.159
Surfactant C 22.6 -0.5 39.9 1.45 14.2 0.141 0.156
Surfactant D 22.3 0 39.4 1.41 13.8 0.137 0.151
Surfactant E 22.0 +1.5 38.6 1.05 13.0 0.139 0.151
Surfactant F 22.2 +1.0 38.9 1.18 13.2 0.140 0.154

Table 4: System 1, Blown with HFC-245fa - Processing and Foam Data

Free rise Relative Overall Density Com pressive K-factor K-factor

Surfactant density flow density distribution strength perp. initial aged
type 'BTU -inchl TBTU-inchl
[kg/m2] [%] [kg/m2] [P»i]
h- f f bf \ [h-jf^F]

Surfactant A 20.5 +6 36.3 0.81 17.8 0.132 0.146

Surfactant B 22.2 -1 39.8 1.11 16.4 0.136 0.144
Surfactant C 21.8 +1 38.5 0.90 16.8 0.135 0.141
Surfactant D 22.4 0 39.2 0.90 15.9 0.133 0.138
Surfactant E 21.0 +9 35.6 0.76 15.7 0.132 0.139
Surfactant F 21.5 +4 37.7 0.83 16.5 0.133 0.139

506 / Burkhart, Klincke, Willoughby

245fa blown foams had better flow properties, higher on the type o f surfactant. But the densities are somewhat
compressive compressive strength and up to higher in the HFC -245fa blown foams indicating a loss of
0.013BTU*inch/h ft2*°F lower aged k-factors. The only some blowing agent. The rather low reactivity o f the
negative aspect are the somewhat more pronounced system could explain this.
subsurface voids. The k-factors o f the HCFC-141b blown foams are up to
0.014BTU*inch/hft2 °F lower and the compressive
System 2 strengths reduced by more than 3.5psi, partially because
System 2 is blown with HCFC-141b and stabilized with o f the reduced densities.
2pphp of a surfactant similar to surfactant D. Cream time, This system has a superior solubility for HFC -245fa
string time and tack free time are 9sec, 48sec and 88sec. and therefore the excellent performance of the more
The foam results we found with the original and the HFC- hydrophobic surfactants was expected. Although the
245fa blown system are given in Table 5 and 6. judgement of the foam structure was better than with
The flow properties in both series do not depend much system 1 it still has to be improved.

Table 5: System 2, Blown with HCFC-141b Processing and Foam Data

Free rise Relative Overall Density C om pressive K-factor K-factor
Surfactant density flow density distribution strength perp. initial aged
type TBTUinchl \ BTUinchl
[kg/m2] [%] [kg/m2] [P*i]
Yh-fi-'F]
Surfactant A 23.5 0 38.0 0.38 14.8 0.137 0.155
Surfactant B 23.2 -1.0 36.4 0.24 14.2 0.131 0.148
Surfactant C 22.3 0 36.9 0.22 14.7 0.135 0.152
Surfactant D 22.4 0 36.9 0.55 13.2 0.133 0.150
Surfactant E 23.7 +1.5 37.9 0.44 15.1 0.131 0.147
Surfactant F 22.9 +1.0 36.8 0.36 14.4 0.129 0.149

Table 6: System 2, Blown with HFC-245fa - Processing and Foam Data

Free rise Relative Overall Density C om pressive K-factor K-factor
Surfactant density flow density distribution strength perp. initial aged
type [BTU inchl BTUinchl
[kg/m2] [%] [kg/m2] [P*i]
[ hj?°F\
Surfactant A 25.8 +1.0 39.5 0.36 17.4 0.150 0.159
Surfactant B 26.2 -1.0 39.9 0.72 20.3 0.144 0.156
Surfactant C 24.9 +1.0 38.2 0.37 18.0 0.145 0.154
Surfactant D 24.4 0 39.3 0.27 18.6 0.144 0.156
Surfactant E 24.7 +1.0 37.9 0.40 16.8 0.145 0.152
Surfactant F 25.3 +0.5 38.7 0.38 18.3 0.143 0.152

System 3 In Table 7 the processing and foam data of the HCFC-

System 3 was blown with the highest amount of 141b blown foams are shown. Surfactant E, being quite
blowing agent, originally 39pphp HCFC-141b equivalent hydrophobic, is the reference product in this series as it is
to 45pphp HFC -245fa and that was about the maximum superior in flow as well as in k-factor.
soluble concentration of HFC -254fa in this polyol. When we replaced HCFC-141b by HFC-245fa we were
Therefore with 2.5pphp o f more hydrophobic surfactants not able to make any decent, reproducible foam to
the blend was already turbid (see Table 2). evaluate the different surfactants. Obviously this polyol

Burkhart, Klincke, Willoughby / 507

was specifically optimized for the blowing agent HCFC- was just given, a smooth transition from solution to the
141b and eventually for a certain processing environment. gaseous stage during the reaction could not be achieved,
With HFC-245fa, although the solubility in the polyol no matter what surfactant we used. Our machine settings

Table 7: System 3, Blown with HCFC-141b - Processing and Foam Data

Free rise Relative Overall Density C om pressive K-factor K-factor
Surfactant density flow density distribution strength perp. initial aged
type \BTU*inch\ [BTUinchl
[kg/m2] [%] [kg/m2] [psi]
\ _ h - f? 0F \
Surfactant A 22.1 +2.0 40.1 0.88 16.1 0.139 0.157
Surfactant B 23.7 - 2.5 39.5 0.67 14.9 0.132 0.154
Surfactant C 23.3 -4.0 39.8 0.71 15.7 0.132 0.150
Surfactant D 22.8 -1.0 38.4 0.65 14.4 0.127 0.143
Surfactant E 23.0 0 38.1 0.59 13.8 0.126 0.141
Surfactant F 23.7 -1.5 38.6 0.62 15.1 0.129 0.145

were not suitable to deal with this less compatible system. in this series. Table 8 shows the data for the HFC -134a
In this case the difference in solubility between HCFC- blown foams. The higher free rise and molded densities
141b and HFC -245fa becomes clearly decisive for the indicate that we had a loss o f the low boiling blowing
foaming result. agent HFC-134a. The lowest k-factors and the best cell
structures with the fewest voids we got with surfactants D
System 4 and £ which are the preferred products for froth systems.
The results we got with system 4, when blown with
This system was originally blown with 19pphp HFC- HFC-245fa, are summarized in Table 9. The densities are
134a equivalent to 25pphp HFC -245fa the lowest significantly lower than in the original foams blown with
concentration o f blowing agent in all four polyols. The HFC-134a. Because o f a much less blowing agent loss in
creamtime, string time and tack free time were 5sec, this series. K -factors are also lower and much better aging
40sec and lOOsec. Surfactant D was our reference product properties were found.

Table 8: System 4, Blown with HF0134a - Processing and Foam Data

Free rise Relative Overall Density C om pressive K-factor K-factor
Surfactant density flow density distribution strength perp. initial aged
type f BTUinch\ [BTU -inert
[kg/m2] [%] [kg/m2] [psi]
F
F

\ _ h -J? aF \
Surfactant A 27.1 +2.0 43.3 0.49 20.1 0.151 0.177
Surfactant B 28.4 - 1.0 44.1 0.52 21.9 0.149 0.176
Surfactant C 29.3 - 4.0 44.7 0.45 20.7 0.146 0.173
Surfactant D 26.0 0 42.9 0.39 18.8 0.143 0.168
Surfactant E 27.5 +3.5 42.2 0.49 18.3 0.144 0.171
Surfactant F 28.7 +1.5 43.4 0.47 19.2 0.146 0.171

508 / Burkhart, Klincke, Willoughby

 Table 9: System 4, Blown with HFC-245fa - Processing and Foam Data
Free rise Relative Overall Density C om pressive K-factor K-factor
Surfactant density flow density distribution strength perp. initial aged
type \B TU inch\ TBTUinch\
[kg/m2] [%] [kg/m2] [psi]
\ _ h J ? 0F \ YhJ?°F\
Surfactant A 21.6 +6.0 32.9 0.59 10.2 0.138 0.153
Surfactant B 22.4 - 4.0 36.2 0.48 12.3 0.140 0.156
Surfactant C 22.4 -3.0 35.8 0.49 13.0 0.139 0.151
Surfactant D 22.3 0 35.2 0.90 9.8 0.141 0.155
Surfactant E 22.2 +6.5 33.6 0.86 11.8 0.138 0.149
Surfactant F 22.6 +2.5 34.7 0.62 12.7 0.140 0.153

Summarizing the processing and foam data o f all four identified could be confirmed with these trials. The results
systems with different surfactants leads to the following are shown in Table 10 as well as in Figure 1. In surfactant
conclusions concerning the choice of the silicone G we had further increased the hydrophobic character of
surfactant: the surfactant, whereas in surfactant H the organic portion
was more pronounced. A significant negative effect on
• The frothing is as critical as expected and therefore flow was observed as well as a coarser cell structure and a
the structural parameters in the surfactant proven to high k-factor compared to surfactant E.
support froth stabilization are important.
• In systems with a sufficient solubility for HFC 245fa FUTURE W ORK
more hydrophobic surfactants perform superior in
Up to now we have gained some basic experience how
terms o f insulation and flow properties as well as
to work with HFC -245fa in simple handpours but much
minimizing foam defects.
more important on our machines consuming a minimum
• When the solubility o f HFC 245fa is borderline
of HFC-245fa. The surfactant screening has given us
surfactants can’t overcome processing deficiencies. meaningful directions for the choice of the silicone
With this in mind we screened some surfactants in a depending on the type o f the system and the OEM
system which was already adjusted to the requirements o f requirements.
HFC-245fa. To save material we had to do handpours, We are prepared to cooperate with raw material and
cooling the polyol and the isocyanate down to 10 °C and system suppliers as well as with the OEMs to develop the
also using our L-shape mold. With this mold we have most suitable surfactant to support the transition to HFC -
slight overpacking in the horizontal part and basically free 245fa. Our experience will allow us to respond fast and
rise conditions in the vertical part. The trends which we sound.

Table 10: HFC-245fa Blown Appliance System - Processing and Foam Data
Free rise Relative Overall Density C om pressive K-factor K-factor
Surfactant density flow density distribution strength perp. initial aged
type I"BTU- incfA [ BTU-inch\
[kg/m2] [%] [kg/m2] [P*i]
l h f ^ 0F \
Surfactant D 28.3 +2.9 36.1 0.65 13,5 0.136 0.156
Surfactant E 28.0 0 34.6 0.57 12.3 0.131 0.151
Surfactant F 28.2 -1.0 37.5 0.63 13.7 0.135 0.155
Surfactant G 27.9 -8.8 40.7 1.43 17,6 0.164 0.235
Surfactant H 30.7 -4.0 38.6 0.82 14.8 0.142 0.163

Burkhart, Klincke, Willoughby / 509

 Figure 1: Foam sections from L -shape lance

CONCLUSIONS Cyclopentane”, Proceedings o f the SPI, 1995, pp.

In four different appliance polyols blown with HCFC- 448
141b, HFC -134a and cyclopentane the blowing agents 2. Wacker W., A. Cappella, W. Hoffinann, 1996,
“Evolution of Hydrocarbon Blown Formulations for
were replaced by HFC -245fa and a solubility study and a
Domestic Appliances Application”, 1996, pp. 18
surfactant screening were carried out. There are 3. Parenti V., G. H. Dean, P. Clavel, A. Selin, L.
significant differences in the HFC -245fa solubility in the Zancai, R. Del Perugia, 1997, “Lower Density
four systems nevertheless the amount of HFC -245fa Cyclopentane/ Butane Blown Rigid Foam: Industrial
equivalent to the original concentration of blowing agent Validation for Domestic Large Appliance Insulation”,
was soluble. In one system more hydrophobic surfactants Proceedings o f the Polyurethanes World Congress,
reduced the solubility of HFC -245fa and caused some 1997, pp. 308
haziness. 4. Decaire B. R., H. T. Pham, R. G. Richard, I. R.
A variety o f surfactants designed to address the frothing Shankland, 1992, “Blowing Agents: The Next
Generation44, Proceedings o f the SPI-34th Annual
behavior of the systems and the solubility of HFC -245fa
Technical/Marketing Conference, 1992, pp. 2
in the polyols has been investigated. Some structural 5. Knopeek, G. M., R. C. Parker, R. G. Richard, I. R.
parameters could be identified which reduce froth defects Shankland, 1993, “Evaluation of Next Generation
and k-factor and improve flow properties. Surfactant E Blowing Agents”, Proceedings o f the Polyurethanes
and F from our study will be our reference surfactants in World Congress, 1993, pp. 465
future developments with systems already adjusted to the 6. Knopeek, G. M., R. C. Parker, R. G. Richard, I. R.
HFC -245fa requirements. Shankland, 1994, “Status Report on the Development
of a Liquid HFC Blowing Agent”, Proceedings o f
ACKNOWLEDGEMENT 35th Annual Polyurethane Technical/Marketing
Conference, 1994, pp. 115
7. Williams, D. J., M. C. Bogdan, R. C. Parker, G. M.
The authors thank Mr James Alexander for the machine Knopeek, 1995 “Update on the Development of
work and Mr Torsten Metz for carrying out the solubilty HFC -245fa as a Liquid HFC Blowing Agent
studies. Proceedings o f the Polyurethanes, 1995, pp. 2
8. Logsdon P. B., R. C. Parker, D. J. Williams, 1997
REFERENCES “HFC -245fa as a Blowing Agent for the Appliance
Industry”, Proceedings o f the Polyurethanes World
1. Birch A. J.,V. Parenti, K.-J. van Duin, G. F. Smits, P. Congress, 1997, pp. 468
Clavel, 1995, “Iso-/n-pentane Blown Rigid 9. Doerge, H. P., 1995, “Zero ODP HFC Blowing
Polyurethane Foam for Appliances: A Realistic and Agents for Appliance Foam”, Proceedings o f the
Economically Attractive Alternative to Polyurethanes, 1995, pp. 432

510 / Burkhart, Klincke, Willoughby

10. Bogdan, M. C., R. C. Parker, D. J Williams, 1997 BIOGRAPHIES
“Utilization o f HFC -245fa in the Construction
Industry, Proceedings o f the Polyurethanes World Georg B urkhart
Congress, 1997, pp. 7 Dr. Georg Burkhart joined
11. Doerge, H. P., 1997, “New Developments in HFC Goldschmidt AG in 1981 after
Appliance Foam”, Proceedings o f the Polyurethanes receiving his Ph.D. in
World Congress, 1997, pp. 484 chemistry at the University of
12. Bogdan, M. C., P. Verbriest, D. J Williams, 1998 Bochum, working mainly at
“Utilization o f HFC -245fa in Spray Foam, the Max-Planck-Institut of
Proceedings o f the Polyurethanes Expo 1998, pp. 23 Coal research. He has been
13. Boze, C. D., K. J. Elsken, 1998, “Effects of involved in both polyurethanes
Processing Variables on a Polyurethane Foam System research and technical service
Blown with HFC 245fa”, Proceedings o f the and since 1986 he is the leader
Polyurethanes Expo 1998, pp. 351 of Goldschmidt’s poly­
14. Logsdon, P. B., D. J. Williams, 1998, “Third urethane laboratories.
Generation Blowing Agent Update for The Appliance
Industry”, Proceedings o f the Polyurethanes Expo
1998, pp. 337 M anfred Klincke
15. Bogdan, M. C., D. J. Williams, 2000, “HFC -245fa
Systems Co-Blown with C O 2 (Water): A Quality, Manfred Klincke obtained his
Cost Effective HFC Alternative for Construction engineering degree from the
Foam Applications ”, Proceedings of the Fachhochschule Niederrhein,
Polyurethanes Conference, 2000, pp. 163 Germany 1978. Afterwards he
16. Doerge, H. P., St. L. Schilling, 2000, “Appliance joint Goldschmidt, spending
Foams with Reduced Levels of HFC -245fa”, ”, nine years in R&D and
Proceedings o f the Polyurethanes Conference, 2000, technical service o f poly­
pp. 445 urethanes. He focussed on
17. Bogdan, M. C., D. J. Williams, P- B. Logsdon, R. C. flame retardants, polyether
Parker, 1996, “Status Report on the Development of polyols, catalysts and new
HFC -245fa as a Blowing Agent”, Proceedings o f the foam systems. 1989 he spend
Polyurethanes Expo, 1996, pp. 394 two years in PU technical service group of Goldschmidt
18. Burkhart, G., M. Klincke, 1993, “Innovative Silicone Chemical Corp., Hopewell, VA. Since 1991 he has been
Surfactant Technology for HCF -134a and working in the rigid urethane group of the technical
Cyclopentane BlownRigid Polyurethane Foams”, service department.
Proceedings o f the Polyurethanes World Congress,
1993, pp. 361
19. Altoe, P., P. Clavel, 2000, “Excellent Insulation Kenneth Willoughby
Properties AND Fast Demold AND Optimum
Process AND ...Applying the Power of the AND in Kenneth Willoughby graduated
PU Rigid Foam for Appliances”, Proceedings o f the from East Texas State
Polyurethane Conference, 2000, pp. 459 University in 1982 with a B. S.
20. Burkhart, G., M. Klincke, 1997, “Additives to in Chemistry. In 1983, he joined
Optimize Rigid Foam Processing and Performance”, Rmax as Laboratory Chemist.
Conference Papers o f the Utech Asia 1997, Paper 27 While there he gathered a
21. Stevens, R. E., G. P. Dado, M. J. Kimock, M. L. significant amount of
Listemann, 1997, “The Influence of Silicone experience with poly-urethane
Surfactants on the Dispersion and Reactivity of and polyiso-cyanurate foam
Water in Flexible Slabstock Polyurethane Foam”, insulation. In 1991, he joined
Proceedings o f the Polyurethanes World Congress, Goldschmidt Chemical Corporation as Technical Services
1997, pp. 118 Representative, being promoted to Senior Technical
22. Boinowitz, T., Burkhart, G., M. Klincke, Metz, T., Services Representative in 1996. His responsibilities
Willoughby, K., 1999, “Surfactants Supporting the include all areas of rigid polyurethane foams.
Use o f Hydrocarbons in US-Type Rigid Foam
Formulations44, Poster presented at the Polyurethanes
EXPO 1999

Burkhart, Klincke, Willoughby / 511

“LinFlex™”, A Proven Success, New-Innovative Refrigerator
Cabinet Polyurethane Machinery

DAVID R. HANNE a n d JAM ES WILLISON KURT PIEPER a n d BEATRIX GEIGER

Hennecke Machinery Hennecke GmbH
Polyurethanes Division Birlinghoven Strasse 30
Bayer Corporation D-53754 Sankt Augustin
100 Bayer Road Birlinghoven
Pittsburgh, Pennsylvania 15205 Germany

ABSTRACT Additionally, since the LinFlex system is a turnkey

The competitive pressure in the refrigerator industry is process, other enhanced technologies will be presented.
constantly increasing. Therefore, many manufacturers are These processes include a patented design to blend a
forced to reduce their costs while simultaneously variety of blowing agents and reduce the number of pour
enhancing production flexibility. Hennecke is once again stations from eight (8) or sixteen (16) to one (1) or two (2)
responding to the demands of the market by providing the respectively.
LinFlex™ system, an entirely new production line concept
for insulating refrigerator cabinets with rigid polyurethane INTRODUCTION
foam.
This paper will illustrate the new technologies developed The LinFlex ™ System (Figure 1) is designed for high
for the LinFlex system that are superior to the conventional performance, mass production and accommodates different
stationary refrigerator cabinet systems. These new LinFlex types of cabinets. In contrast to the stationary plant, the
technologies have been proven in many plant installations supporting molds are transported by means of a linear
to save up to 30% manufacturing space and subsequently
transport system. The following provides a brief
reducing capital cost for the equipment. In order to
explanation with the numbers that corresponds to the
achieve this success, innovative designs were implemented
numbers on the LinFlex System (Figure 1).
to allow rapid and easy change of fixtures/cores from one
1. Preheating Oven: The prefoamed cabinets are
refrigerator model to another with virtually no effect on
production (zero-time change over). Just as important, the conveyed to the preheat oven so the prefoamed
cycle time could be as low as 30 seconds depending upon cabinets can be heated at the proper temperature
production requirements. before foaming.
As noted, this LinFlex system can produce many 2. Infeed Transfer: This station transfers the heated
different models, but one more critical aspect needed to be prefoamed cabinets from the oven to the Central
solved. This was to supply variable curing times to Station.
support different models which was solved with innovative
designs to have variable curing times as low as 3.75
minutes up to 15 minutes.

POLYURETHANES EXPO 2001 , SEPTEMBER 30 -OCTOBER 3, 2001 513

 Figure 1 - The LinFlex System

3. Central Station: The Central Station is where the LINFLEX™ SYSTEM CONFIGURATION
prefoamed cabinet is loaded into a fixture, which
has the appropriate core. Once the fixture, core and The configuration of the LinFlex™ system is mainly
prefoamed cabinet are locked together as a unit, this driven by two factors: production requirements, which
unit is then transferred to the foaming station. Also, indirectly determine the cycle time, and maximum curing
depending on the configuration, this central station time. The cycle time determines the number of central
will be utilized to unload the foamed cabinet once stations since the central station has the longest cycle time.
cured. If the cycle time is about 60 seconds, one central station is
4. Foaming Station: The Foaming Station is where the required which performs the following functions;
mixhead fills the cabinet with foam. Also, at this unlocking the moveable foaming fixture, lifting the
station, the ventilation is used. The foaming supporting core with the mold cover, removal and loading
cabinets are transferred from the foaming station to of the cabinet from the foaming fixture; closing and
the curing station via the shuttle (6). locking of the foaming fixture. Table 1 illustrates the
5. Curing Station: At the curing station, the foamed LinFlex system configuration with a cycle time of 60
cabinets are cured to a predetermined time. The seconds. As mentioned previously, the production
number of stations will be determined in accordance requirements and cycle time complement each other. The
with the required cycle time. production capability is 60 cabinets per one (1) hour.
6. Shuttle: There are two (2) shuttles: one shuttle to However, if the customer requires a higher production rate,
transfer the foamed cabinet and its corresponding the cycle time must be decreased as shown in Table 1, a
fixture/core unit to the curing station, and another LinFlex system with 30-second cycle time.
shuttle to transfer the cured foamed cabinet and its Subsequently, the production rate increases to 120
corresponding fixture / core unit to the central cabinets per hour and the number of central stations needed
station. becomes two (2). One of the central stations will perform
7. Outfeed: The foamed cabinets are transferred from the same functions as mentioned in the 60-second cycle
the central station to the customers post foam LinFlex system. The second Central Station, apart from
assembly. the basic functions, is needed for the 30-second cycle time
8. Changing Station: The changing station allows other LinFlex system, which will unload the foamed cabinet
model fixture/core units to be transferred in the after curing. The last step to the configuration is knowing
system and to remove a fixture / core unit, by the maximum curing time, which indirectly determines if
utilizing the shuttle. the number of required fixture / core units, which should
increase with curing time. For example, if the maximum
curing time is 6 minutes (360 seconds), a minimum total of
nine (9) fixture / core unties for curing and 1 fixture / core
units are needed for each of the three work stations

5 1 4 / Hanne, Willison, Pieper, Geiger

 Table 1. The LinFlex System Cycle Times
Process 60-Second Cycle Time 30-Second Cycle Time
Cabinet Production / Hour 60 120
Central Station 1 2
Foaming Stations / Mixhead 1 1
Shuttle 2 2-3
Curing Time 3.75 - 1 5 minutes 5 minutes
Pre-heating Oven 1 1
Number of Curing Stations 13 10
Work Stations 3 4

handling equipment, the constructive design of the

Zero Time Change Over LinFlex fixtures are less complex
♦ Filling holes can be conventional (at the compressor
The uniqueness of the LinFlex system is the ability to end) or Top Flow (at the middle-back of cabinet)
change from one refrigerator cabinet model to another with ♦ One or two pouring stations
virtually no effect on production. This change can be ♦ Flexibility to increase capacity
performed semi- or fully automatically at the changing ♦ Adaptability to a variety of curing times
station. Once the fixture is adjusted and a new core is in ♦ Zero-time change over for the fixture / core unit.
place, the fixture / core unit is transferred into the system.
Figures 2 and 3 show a drawing and a picture of the CONCLUSION
changing station respectively.
With over 50 years of experience, Hennecke continues
LINFLEX SYSTEM SUMMARY to lead the way with new innovative equipment for
producing polyurethane. The experiences and proven
To summarize, the LinFlex system has many features technologies continue to captivate the market place to
such as: provide solutions for the everyday problems and especially
♦ In some instances, up to 30% less space is needed as for the future requirements.
compared to conventional production lines.
♦ Simple foaming fixtures: compared to fully
stationary foaming fixtures fitted with their own

Figure 2 - Changing Station

Hanne, Willison, Pieper, Geiger / 515

 Figure 3 - Changing Station

BIOGRAPHIES Beatrix Geiger

David R. Hanne Beatrix Geiger is currently a Project Manager at

Hennecke GmbH, Sankt Augustin-Birlinghoven and is
David R. Hanne is currently an Account Manager for responsible for refrigerated appliances. Graduated in
Hennecke Machinery, a unit of the Bayer Corporation, process engineering from Cologne technical university, she
Polyurethane Division in Pittsburgh, PA. David has been has been employed by Hennecke GmbH since 1985.
employed with Hennecke since 1984. He has an MBA
from Duquesne University, Pittsburgh, PA and a BS in
Electrical Engineering from Penn State University, State
College, PA.

Jam es Willison

James Willison is Sr. Technical Sales Representative for

Hennecke Machinery, a business unit of Bayer
Corporation, Polyurethane Division in Pittsburgh, PA. He
began his career with Hennecke Machinery in 1993. He
has a degree in robotics from the University of Kentucky /
CCAC. Jim has held various positions at Hennecke
Machinery, including Programmer and Project Manager.

K urt Pieper

Kurt H. Pieper is currently a Sales Manager at Hennecke

GmbH, Sankt Augustin-Birlinghoven, and is responsible
for refrigerator plants. He has been employed by Hennecke
GmbH since 1987 and is a diploma engineer from
Fachhochschule Flensburg.

516 / Hanne, Willison, Pieper, Geiger

POSTER SESSION
CASE
A New Anti-hydrolysis Ether-Based PU System for
Uni-sole Application

CHARLES C. T. HUANG a n d SHARON C. P. LIU

Pou Chien Chemical C o L t d .
No. 265 Sec. 1, Yeong Fu Rd.
Yenog Ging, Chang Hua
Taiwan, R.O.C.

ABSTRACT Charles C.T. Huang graduated from Taiwan Tung Hai

University in 1985 and obtained his M.S. in chemical
Instead of commonly used polyester based outsole system, engineering from National Tsing Hua University in 1987.
the newly developed polyether based system is providing He worked for Union Chemical Laboratories since 1988 and
excellent hydrolytically property. The data show even after joined BARITS Corporation (now named as Pou Chien
4 weeks of hydrolysis test (70°C, 95% RT Humidity) the Chemical) in 1995 as R&D director. His major research
tensile strength will keep in same level without dropping topics included polymer synthesis and characterization,
from the original value, and test result for the tear strength, industrial adhesive application, water-base polyurethane and
elongation and flexibility are the same as well. microcellular elastomer. Now he is managing director of
In this paper, we also conducted a series of testing for real Pou Chien Chemical.
molded part of uni-sole including process ability, low
temperature flex, slip resistance, abrasion resistance, Sharon C.P. Liu
solvent resistance, mold release agent sensitivity and
appearance evaluation. Sharon C.P. Liu graduated from Providence University
All the data show excellent performance making this Taiwan in 1993 and obtained her M.S. in applied chemistry
system becoming a wise choice of outsole material for from Providence University in 1995. She joined BARITS
shoemaker. Corporation (now named as Pou Chien Chemical) in 1995
and is responsible for R&D programs related to application
BIOGRAPHIES development in Footwear.

Charles C.T. Huang

POLYURETHANES EXPO 2001 , SEPTEMBER 30 -OCTOBER 3, 2001 519

Investigation of Polyurethane Adhesives for Use in the Apache
Aviator NBC Protective Mask

GERALD P. YOUNG, STEVEN CIBOROWSKI, DAVID MAY, LARRY OSWALD a n d STEVEN GROSS
United States Army Soldier Chemical-Biological Command
Edgewood Chemical-Biological Center
Berger Laboratory
Aberdeen Proving GroundM aryland 21010-5424

Abstract

Polyurethane adhesives are currently used in Apache abraded with sand paper and treated with Lord 7701
aviator masks to secure polycarbonate lenses to a adhesion promoter.
bromobutyl-natural rubber face piece. This study examined The rubber-adhesive-polycarbonate rupture strength
the rupture behavior of polycarbonate(pc)components results, loads reported in pounds are listed below(average
bonded to butyl-rubber with polyurethane adhesives. of all specimens with error in measurement reported as
Composites were fabricated and destructively tested to 95% confidence limit).Loads were applied perpendicular
reveal differences in failure mode and rupture strength as a to the bond surface plane (1 inch diameter right circular
function of surface preparation, curing regime, and cylindrical surface) in an Instron Load cell. Grips of the
polyisocyanate-polyol formulations. load cell were moving apart at a constant rate of 0.15
Alternative polyurethane adhesives packaged in pre ­ inches per minute. Load versus displacement results were
measured assemblies offer greater ease of application and recorded for seventy -five specimens and correlated with
were investigated to determine if the replacement adhesive failure mode. Distinct failure modes for heat-cured versus
could be used in lieu of the original adhesive employed ambient-cured composites are evident in the
during the manufacturing process. Data on the rupture specimens(photographs are included in the body of the
strength of polycarbonate-adhesive-rubber joints was report.)
collected to compare LORD 7500 and Lord 7540(pre- Load measurements indicate both adhesives are
packaged) adhesive. Rubber components of the prototypes adequate for securing optical lenses within the aviator
were abraded with 80 grit ,100 grit, or no surface abrasion mask. To improve polyurethane adhesive bonding in the
and bonded to polycarbonate with adhesive. Specimens aviator mask application, curing at 63 degrees Celsius for
were cured at temperatures of 63 Celsius or ambient 18 to 24 hours is recommended.
temperature(23 C). All polycarbonate surfaces were

Load to Induce Rupture

Adhesive Surface Curina Separation Mode Load at Failure

7500 100 grit 63 C rubber tearing 261+/-31lbs
7540 100 grit 63 C adhesive tear from pc 239+/-41bs
7500 100 grit 23 C adhesive peeled from rubber 195+/-21lbs
7540 80 grit 23 C adhesive peeled from rubber 177+1-36 lbs
7540 100 grit 23 C adhesive peeled from rubber 170+/-19 lbs
7500 80 grit 23 C adhesive peeled from rubber 149+/-15 lbs

Key Words; polyurethane, rupture strength, adhesives,

polycarbonate-polyurethane-bromobuytl rubber composite.

POLYURETHANES EXPO 2001 , SEPTEMBER 30 -OCTOBER 3, 2001 521

INTRODUCTION polyethers); and flexible, alcohol-terminated polyesters
such as poly-1,4-butylene adipate:
Adhesives have a long history of use in human artifacts.
Studies by the Weizmann Institute of Science reveal that in
the stone age, humans were putting organic compounds to
use to join dissimilar materials together. As far as 8,000-
years-ago Neolithic cavemen who lived in the area of the
Dead Sea were producing a form of collagen glue, as an
adhesive to hold together tools, utensils, and ceremonial
artifacts that included decorated skulls.1Natural glues and The alcohol-terminated polyethers and polyesters are
resins derived from animal and plant sources were used known as polyols. Isocyanates commonly used to prepare
extensively from antiquity up to the early 20th century. It polyurethanes are toluene diisocyanate (TDI), methylene-
is known that the ancient egyptians used collagen in its 4,4 -diphenyl diisocyanate (MDI), and a polymeric
gelatinous form as the basis for the ’’carpenter's glue" to isocyanate (PMDI). These isocyanates have the following
hold together furniture.With the advent of the industrial structures \
revolution and the establishment of a chemical process
industry, adhesives began to be derived as adjuncts to a
wide range of compounds derived from petrochemical
sources.
In the Apache Aviator Mask, Polyurethane is used as a
material to bond polycarbonate lenses to a face piece that is
composed of a blend of natural and bromobutyl rubbers.
There is probably no more versatile material for so many
high-performance applications than polyurethane. With a
molecular structure composed of four of the most common
elements - carbon, hydrogen, oxygen, and nitrogen - During the late 1930s Otto Bayer, manager of the IG
polyurethane has a combination of properties which are not Farben laboratories in Leverkusen, Ger., prepared many
available in any other thermoplastic material including polyurethanes by condensation reaction of dihydric
excellent resistance to ozone, weak bases, fuels and alcohols such as 1,4-butanediol with difunctional
oils.Urethane was first introduced in 1849 by Germany's diisocyanates. A major breakthrough in the commercial
Wurtz and Hoffman, when they presented the chemical application of polyurethane did not occur until 1941, when
reaction between Isocyanate and Hydroxyl.2 The a trace of moisture reacted with isocyanate to produce
precursors were initially limited to Diisocyanate and carbon dioxide. The production of this gas resulted in
Polyester; however, the industry began looking for many small empty areas, or cells, in the product (which
different Hydroxyl Polyols due to problems with was subsequently called "imitation Swiss cheese"). In 1953
processability, cost, foaming physical properties, etc. By Bayer and the Monsanto Chemical Company (now
1957, various Polyether Polyols were being produced Monsanto Company) formed the Mobay Chemical
,offering not only cost effectiveness but also anti-foaming Corporation to produce polyurethane in the United States.3
physical properties. Thereafter, the Urethane Industry was Urethane is a product of Hydroxyl compounds such as
able to make leaping advances with the development of Polyether Polyols and polymerized derivatives of
silicone surfactants. Polyurethanes are a class of Isocyanates. Urethane’s unique chemical structure is the
extremely versatile polymers that are made into flexible result of supplemental reactions between the
and rigid foams, fibres, elastomers, and surface coatings. Polyisocyanate and Polyol(activation hydrides).
They are formed by reacting an isocyanate (a compound The external characteristics of the polyurethane end-
having the functional group NCO) with an alcohol (having products vary, resulting in diverse characteristics such as
the functional group OH).Polyurethane molecules can marshmellow -like softness to concrete-hardness, or rubber­
adopt a linear or a network architecture. Linear like stretch-ability that allows the product to stretch to 10-
polyurethanes are formed by reacting a dialcohol with a times its size to materials that offer almost no change in
diisocyanate, whereas network polyurethanes are formed form. Additionally, Urethane, as a comfort and
from polyfunctional alcohols or isocyanates. Dialcohol convenience-providing material, is used in a wide range of
monomers include ethylene glycol (HOCH2CH2OH); applications in automobiles, electronics, furnishings,
diethylene glycol (HOCH2CH2OCH2CH2OH); 1,4- architecture, etc. because they offer unique qualities
butanediol (HOCH2CH2CH2CH2OH); 1,6-hexanediol relative to other general high molecular compounds.
(HO[CH2]6OH); alcohol-terminated polyethers such as In the Apache Aviator Mask, polyurethane adhesives are
polyethylene oxide and polypropylene oxide (see Aliphatic critical materials for joining dissimilar materials together
and providing chemically resistant, mechanically tough

522 / Young, Ciborowski, May, Oswald, Gross

bonds and joints. In this study, two polyurethane polyurethane adhesive-polycarbonate plastic joints. The
adhesives cured at elevated and ambient temperature and rupture strength data will serve as a basis for the
with different surface preparations are compared. As a subsitution of adhesives in planned mask maintenance,
measure of the effectiveness of the adhesives, the bond upgrade, and repair operations where lens replacement is
rupture strengths were measured for the various rubber- required.

MATERIALS AND METHODS Approximately 0. 1 cubic centimeter of adhesive is

applied to each face of the polycarbonate and then spread
Test items consisting of one inch diameter right circular between the two-polycarbonate mating surfaces.
cylindrical blocks of polycarbonate, stamped circular Once both surfaces were covered with the adhesive the
rubber coins and polyurethane adhesive were assembled. prepared rubber coins were inserted between the
The polycarbonate substrates were machined from polycarbonate cylinders, compressed together by hand,
polycarbonate 1 -inch diameter rod with a 1.5inch stem excess adhesive removed and the joined components set in
(see figure A) using ASTM 429 as a general guide to the a 1 inch inner diameter plastic tube which acted as a jig to
configuration of test specimens. keep the parts from slipping apart or sliding out of
Once machined, the polycarbonate pieces were sanded alignment during curing. Ten Specimens prepared with
with either with 80 or 100 grit aluminum oxide sand paper, 7500 adhesive were then cured in a 63 degrees Celsius
and cleaned with isopropyl alcohol solution (99% by convection oven for 24 hours . Ten assemblies prepared
volume) to remove any residual polycarbonate dust or grit with 7540 adhesive were also cured in the convection oven
left on the surface from the sand paper. The surfaces were at 63 degrees Celsius for 24 hours. All the remaining
coated with LORD 7701 adhesion promotion solution assemblies were allowed to cure for a minimum of 48
(consisting of a mixture of ethyl acetate and alcohols), hours at room temperature.
followed by drying for several minutes while resting on
absorbent paper. The stamped rubber coins were also
sanded with 80 or 100 grit sand paper, followed by
cleaning with isopropyl alcohol solution and application of
the LORD 7701 adhesion promoter. The individual
components were then assembled in groups, which were
differentiated by the use of either LORD 7500 adhesive or
LORD 7540 adhesive. In any assembly the polycarbonate
surfaces were all either 100 grit sanded or 80- grit sanded.
Some rubber coins were not sanded to determine the
impact of no rubber surface preparation on joint rupture
strength.
Lord 7500 adhesive is a two part adhesive consisting of a
black„extraordinarily viscous polyol which must be mixed
with a cream colored isocyanate mixture in the proportion
of 1 part by weight polyol to 1.7 parts isocyanate. The
resulting mixture then appears a light gray color and
depending on degree of mixing will sometimes display
variegated strands of lighter and darker colors. Good
mixing of the components results in a visually uniform
color without dark and light striations. The useful life of
this mixture is exceedingly short, on the order of 10
minutes or less. The reaction between polyol and
isocyanate quicly proceeds to the point where it is
impractical to apply the adhesive since it becomes
extraordinarily viscous and too difficult to apply.
Lord 7540 adhesive is also a two part polyurethane
system, however it is manufactured in pre-measured
applicator tubes, which are designed for use in a glue gun .
A static mixing element is attached to the outlet of the
cartridge to blend the two components. In this application
the ratio of components appears to be 1 to 1 as both
preloaded tubes are of equal diameter. The 7540 adhesive,
when well mixed, is also a light gray color.

Young, Ciborowski, May, Oswald, Gross / 523

 Figure 1: Configuration of P rototyp es

524 / Young, Ciborowski, May, Oswald, Gross

RESULTS

The follow ing rupture load results were obtained when the test specim ens were placed under
tensile stress in the Instron apparatus.

Loadto Induce Rupture

Adhesive Surface Curina Separation Mode Load at Failure

7500 100 grit 63 C rubber tearing 261+/-31 lbs

7540 100 grit 63 C adhesive tear from pc 239+/-41 bs
7500 100 grit 23 C adhesive peeled from rubber 195+/-21 lbs
7540 80 grit 23 C adhesive peeled from rubber 177+/-36 lbs
7540 100 grit 23 C adhesive peeled from rubber 170+/-19 lbs
7500 80 grit 23 C adhesive peeled from rubber 149+/-15 lbs

GROUP SURFACE CURING RUPTURE MODE Load a t R upture

£Lbs)
7500 100 Grit 63 C elsius R ubber T earing 261+/-31

Young, Ciborowski, May, Oswald, Gross / 525

7540 100 Grit 63 Celsius Adhesive Tearing 239+ / - 41 lbs
from Polycarbonate

7500 100 Grit Ambient Rubber Delaminates 195+ / - 21 lbs

from Adhesive

526 / Young, Ciborowski,May, Oswald, Gross

7540 80 Grit Ambient Rubber Delaminates 17 7 + / - 3 6 lbs
from Adhesive

7540 100 Grit Ambient Rubber Delaminates 1 7 0+ / - 19

from Adhesive

7500 80 Grit Ambient Rubber Delaminates 149+ / - 15

Young, Ciborowski, May, Oswald, Gross / 527

 Rupture Strength Comparison
13 Feb 2001 Data

528 / Young, Ciborowski, May; Oswald, Gross

Young, Ciborowski, May, Oswald, Gross / 529
 Rubber Delamination: Specimen in Instron Machine

C o n c lu s io n s surface preparations and curing regimes are equivalent.

Curing the specimens at 63 degrees Celsius for a period of
One-inch diameter right circular cylindrical Polycarbonate- 18 to 24 hours appears to strengthen the bonds. Surface
Rubber joints can be effectively bonded with polyurethane treatment and abrasion of the rubber in specimens cured at
adhesives. Preparation of the polycarbonate surface elevated temperature seems to improve the bond strength,
includes sanding with 100 or 80 grit aluminum oxide sand but only minimally from 3% to 6%.
paper, cleaning and removal of grit and abraded Surface treatment of rubber seems to have a stronger
polycarbonate with commercially available over-the- effect on improving bond strength in specimens that are
counter isopropyl alcohol followed by treatment with an not subjected to elevated temperature curing and are
ethyl acetate-based adhesion promoter such as Lord allowed to cure under ambient conditions for several days.
Product 7701. Rubber surfaces can also be pretreated with The average increase in bond strength for surface treated
100 or 80 grit papers or no sanding as long as adhesion and abraded rubber in ambient cured specimens ranges
promoter is applied to the rubber prior to application of from 38% to 58%. The overall bond strength of ambient
uncured polyurethane adhesive. cured (20 degrees Celsius) specimens irrespective of
The measurements show that joints of rubber and surface treatment is significantly less than those cured at
polycarbonate bonded with Lord Polyurethane Adhesive 63 degrees Celsius, on the order of 38% weaker than oven
7500 fail under loads that are approximately equivalent to cured samples.
the loads, which cause failure in rubber-polycarbonate
joints bonded with Lord Adhesive 7540, as long as the The following table summarizes the results obtained:

530 / Young, Ciborowski, May, Oswald, Gross

Summary of Results

Rubber Predominant
Surface Cure Joint Failure Stress applied
Adhesive Preparation Condition Mode at joint failure Range
Rubber cleanly delaminates from
7540 none Ambient adhesive with little rubber tearing 146psi +/-18%
Ambient Rubber delaminates from adhesive with
7500 80Grit Small areas of Rubber Tearing 174psi +/-26%
7500 None Ambient Rubber delaminates from adhesive 182psi +/-9%
7540 100 grit Ambient Rubber delaminates from adhesive 217psi +/-11%
7540 80Grit Ambient Rubber delaminates from adhesive 231 psi +/-15%
7500 10Ogrit Ambient Rubber delaminates from adhesive 249psi +/-11%

7540 None 63C Cure Large area of rubber tearing 315psi +/-14%
7500 None 63C Cure Rubber tearing 333psi +/-12%

7540 100 grit 63C Cure Adhesive tearing from polycarbonate 333psi +/-20%
7500 100 grit 63C Cure Rubber tearing 343psi +/-8%

2 Organic Chemistry. Hendrickson J.B.; Cram

LITERATURE CITED D.J.;Hammond G.S., McGrow Hill Series in Chemistry,
1970,.

^ h e se findings, reported in the Hebrew -language 3Encyclopaedia Brittanica Online. Polyurethane Entry.
Archeologia Umada'ei Hateva ("Archaeology and Natural
4. Modem Adhesives. McGraw-Hill Science Press, 1998
Sciences"), add a new chapter to the history of adhesives
and to humankind’s technological history in general. (An
English translation of the paper is available, as are color
slides of the ancient decorated skulls).

Biography highest civilian award, the Superior Civilian Service

Medal, for contributions to the development of
Gerald P. Young is an Engineer at the technologies to safely destroy the Nation’s obsolete
United States Army Chemical stockpiles of Mustard and Nerve Agents without the use of
Biological Center located at Aberdeen incineration. Jerry recently traveled the country providing
Proving Ground, Maryland. With a chemical and biological terrorism awareness training under
Chemical Engineering Degree from the auspices of the Nunn-Lugar-Domenici Domestic
the University of Delaware, Jerry has Preparedness Act, providing realistic consequence
contributed to the design, management exercises for Firefighters, Police, and
development and production of Emergency Medical Workers in 20 major U.S. Cities.
Pyrotechnics, Explosives, Chemical
Detection and recently, Individual
Respiratory Protection Equipment.
Jerry was awarded the Army’s second

Young, Ciborowski, May , Oswald, Gross / 531

Low Gloss, Low Build-Up Water Based Release Agents for
Microcellular and Elastomer Systems

MILTON JOHNSON
Chem-Trend, Inc.
1445 West McPherson Park Drive
Howell, Michigan 48843

A bstract Biography

Polyurethane elastomers and microcellular systems have a M ilton Johnson

wide range of applications including footwear, skate
wheels, rollers, hydraulic seals and bumpers. Chem-Trend Milt Johnson is currently the Technical Director -
has made significant advances in the development of Marketing & Technology at Chem -Trend Inc., as well as
water based release agents for these applications, with being the Technology Leader of Polyurethanes. He
particular emphasis in footwear. Polyurethane elastomers received his B.S. in Chemistry from Iowa State University
are unique materials that require that the release agent be and is a Certified Quality Engineer. Milt is currently
tailored to meet thier wide range of application and responsible for overseeing and coordinating the global
property/performance requirements. One of the key product and technical development of polyurethane
properties is the skin formation and overall gloss of the release agents.
finished part.

Recently the market place has demanded more

environmentally friendly, solvent-free parting agents and
Chem -Trend has accepted the challenge.

We have developed a family of novel water

based release agents that are designed to impart low gloss
while maintaining minimal build-up. These polymers
produce a uniform low gloss surface over a wide range of
processing parameters specifically designed for
microcellular and elastomer systems. The detailed
performance of these novel and innovative release agents
will be detailed in the poster.

POLYURETHANES EXPO 2001 , SEPTEMBER 30 -OCTOBER 3, 2001 533

Lower Durometer TDI Elastomers “Plasticizer Free” for
Solvent Resistance

RAJ. B. DURAIRAJ JO SEPH D e ALMEIDA

INDSPEC Chemical Corporation INDSPEC Chemical Corporation
1010 William Pitt Way 411 Seventh Avenue, Suite 300
Pittsburgh, Pennsylvania 15238 Pittsburgh, Pennsylvania 15219

ABSTRACT ARTW ORK

Until recently, the greatest use of the HER

HER™ chain extender family, in particular (Hydroxy Ethyl E ther of Resorcinol)
HER-HP, had been noted in MDI
applications where higher hardness, tougher
tear, higher tensiles and resistance to
fatigue were necessary properties. The
most recent evaluation of HER™ products in
the chain extension of TDI prepolymers (low
free isocyanate and conventional systems)
suggests there are new opportunities for
longer chain HER™ in the production of
“lower durometer” TDI elastomers without
the use (or with reduced use) of plasticizers;
depending upon the target hardness.
The HER™ product line is comprised of
various grades that include a powder at
room temperature, waxy solids, and liquid at
room temperature consistencies. The
information within presents the physical
properties and resistance to solvents that
can be attained with HER™ extension of TDI
prepolymers. Much of the data readily
pertains to roll cover applications (e.g. inks
& coatings) but other uses where durable
low durometer “plasticizer free" properties
are needed will also find this data to be of
value. The results of this study provide
TDI’s an opening to commercial applications
requiring physical performance at low
hardness with resistance to solvent attack.

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 535

 BIOGRAPHIES

Joseph de Almeida Raj B. D urairaj

Joe de Almeida is Raj B. Durairaj,

INDSPEC Chemical Ph.D. is INDSPEC
Corporation’s (Pitts­ Chemical Corpora­
burgh, PA) Director, tion’s (Harmarville,
Resins and Commer­ PA) Director of
cial Development, a Research. His re ­
position he has held sponsibilities include
since m id - 1996 fol­ the development of
lowing INDSPEC’s resorcinol-based aro­
affiliation with Occi­ matic liquid diols for
dental Chemical Corporation. Previously, he polyurethanes and other polymers. He has
was manager of commercial development at authored 25 technical papers and holds 13
Occidental’s Designed Products Custom patents in the United States and abroad.
Manufacturing Group and marketing manager of
Raj joined INDSPEC in 1986, when it was
ACL® chlorinated isocyanurates following
acquisition of that business group from Koppers Company’s Industrial Specialty
Monsanto. Chemicals Division, and has been involved in all
aspects of the company’s resorcinol chemistry
Joe joined Occidental Chemical in 1980 as a activities since that time. His work includes the
chemical engineer. During his 16 years at development of a variety of resorcinolic
Occidental he held various positions of products, including resorcinol-formaldehyde
resins, epoxy resins, bis-maleimide resins, flame
increasing responsibility in technology support,
retardants and resorcinol manufactured by
manufacturing, technical services, marketing,
hydroperoxidation processes.
and business management. Throughout Joe’s
career, he has dedicated his efforts to new
product commercialization, business growth and Prior to his career in industry, Raj was a
product line expansion. Product areas in which researcher and instructor in chemistry
Joe has expertise include halogenated ag. departments at the University of Connecticut,
Drexel University and Case Western Reserve
intermediates, metal plating, detergent chemistry,
and rubber additives. University. His work included the synthesis and
study of liquid crystalline polymers from
That breadth and depth of experience has isocyanate monomers; synthesis of deuterated
enabled Joe to lead INDSPEC’s efforts to polymers; polymeric surfactants and chain
dynamics; and water-soluble polymers.
commercialize Resorcinol diols Hydroxyethyl
Ether of Resorcinol (HER) and Hydroxypropyl
Ether of Resorcinol (HPR). Raj earned a Ph.D. in polymer science from the
University of Madras, India. His doctoral work
Joe earned an M.B.A. from the State University involved the synthesis and characterization of
of New York at Buffalo and holds a B.S. in new polymers using isocyanate monomers in the
chemical engineering from the University of synthesis of polyurethanes and polyimides.
Massachusetts.

536 / Durairaj, de Almeida

New Developments in Light and Thermal Stabilisation Packages
for Flexible Polyurethane Integral Skin Applications

PATRICK JAC O BS, BRIAN JO HNSON, PETER ORIZONDO

MONICA d e la CRUZ a n d Great Lakes Chemical Corporation
HILDE GOYVAERTS
Great Lakes Technology Belgium
Toekomstlaan 13
2200 Herentals
Belgium

ABSTRACT

The UV resistance and thermal stability have always been

key performance requirements of flexible polyurethane
integral skin foam.

A combination o f these two properties, along with low

fogging characteristics, is especially important in the wide
and fast growing field of (interior) automotive applications.
In contrast, for outdoor uses such as shoe soles, the thermal
resistance is of lesser importance, with the main
performance criterion being the UV resistance

This paper examines the light stability performance of

different stabiliser packages in flexible integral skin foam
under standard weathering conditions (PV1303).

From the results available in this work it is possible to

identify two stabiliser packages, one which shows
performance similar to traditional stabiliser systems and a
second which gives substantially better performance in
both shoe sole and automotive applications.

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 537

RimSpray™: A Unique Method for Applying Polyurethane
Foam to Molds and Substrates

MICHAEL F. BARTENSTEIN
Gusmer-Adtniral, Inc.
305 West North Street
Akron, Ohio 44303

A B STRA CT

Gusmer-Admiral, Inc. has developed a unique spray method of applying polyurethane foams and
coating to mold or substrate surfaces. The concept combines technology from time -proven
Gusm er Corporation spray and Gusmer-Admiral RIM processes into a mechanically self-cleaning,
recirculating RimSpray™ system. This method allows covering large areas of mold or substrate
surfaces by dispersing the media into wide patterns directly related to tip orifice configuration and
height off the work surface. This method has applications throughout the polyurethane industry
from mold coating and filling to reinforced substrate composition.

Figure 1: 6mm Spray Head

D E V ELO P M E N T H IS TO R Y

Gusmer-Admiral was approached by a large, first tier automotive supplier to investigate the
possibility of dispersing polyurethane foam in an open mold NVH application. Drawing from forty
years experience with spray - foam patterns at G usm er Corporation, a prototype w as developed to
fit the client’s specific needs of delivering around 150 grams per second with an approximate
width of twelve inches. The mixhead was required to fit on an existing robot and adapt to the
daylight allowed by the press and mold (Figure 2).

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 539

 Figure 2: 6mm spray head prototype

Due to the successful result of this endeavor, several other applications where spraying could be
more efficient than pouring were discovered through sales and marketing efforts. Development of
other sizes soon followed based on these applications. As development progressed, certain
obstacles had to be overcome.

PRODUCT DEFINITION

Characteristics: The mixheads are mechanically self-cleaning, hydraulically-operated, double-acting,

recirculating mixheads developed to spray polyurethane compounds by dispersing the mixed material
through a spray tip (nozzle) configured to create a fan or round pattern on a mold or substrate. Mechanical
Self - cleaning: To succeed in expunging all the mixed material out the tip of the mixhead upon closing, the
male spherical radius on the end of the mix piston must fit exactly into the mating concave ball socket in the
spray tip. To facilitate line-to-line contact and not damage parts through impact, a semi-resilient washer is
compressed between the tip clamping surface and the tip-retaining nut (Figure 3).

Figure 3: Tip Washer

Basically, the rod and valve block dimensions are tooled to the sam e exacting tolerances a s
Gusmer-Admiral standard mixheads. The thermoplastic bearing material em bedded in the
grooves to separating the A and B chemicals and supporting the rod is a GAI standard and also
characteristic of the spray head design.

R ecirculation: The chemicals are routed through grooves machined in the sides of the non ­
rotating mix piston allowing the chemicals to recirculate back to their source continuously for
precise temperature control. When retracted (head open) the piston covers and seals the return
port causing the two chemicals mix in the open cham ber and exit out the spray tip mounted on
the end of the head. Both ingredients flow through adjustable orifices used to create pressures for
impingement mixing. This is the case whether in recirculation or dispensing.
S pray Tips (Nozzles): The spray tip profile is based on the desired output of the mixhead, the
resultant orifices required, and the width of pattern desired. The criterion for tip development
includes minimizing the cross - sectional tip angle to allow directing the majority of material to the
extremities of the fan slot
(Figure 4).

540 / Bartenstein
 Figure 4: Tip Angle
After deciding the orifices required to develop pressures needed for mixing, normally 1800 -2200
psi, the area of the slot is determined. This is usually slightly less than the orifice areas combined
allowing for sufficient backpressure to disperse the mixed material. Realizing the difficulty of
pinpointing all the tip combinations to be used, certain sizes were determined using a 15 - degree
tip angle cutter and various cutter depths, which developed slot areas where the cutter breaks
through (Figure 5). The tip is then hardened and the concave ball surface polished to fit the rod
end.

Figure 5: Slot Area

Tip C leaner (Figure 6): To insure no distortion appears in the pattern, it is imperative that the
spray tip slot be kept clean. For this reason, a tip cleaner is available providing a mechanism that
employs two bronze or nylon wire wheels spinning in opposite directions. The mixhead tip is
moved parallel to the slot over both wheels by manipulator or other m eans providing foam
cleaning in two directions thus avoiding build-up in one slot corner or the other.

Figure 6: Tip cleaner

O ver Spray: Due to the dispersal of uncured fluid from the tip, there is a tendency to create over ­
spray at the extremities of the fan pattern. Selecting the proper tip based on the output being
used can minimize this phenomenon. There will be a requirement for masking areas local to the
work, however, due to som e over - spray and the ricochet of particles at the edge of the work.

SPRAY HEAD SPECIFICATIONS & CAPABILITIES:

6mm S pray head:

Mix piston diameter: 6mm (.236")

Bartenstein / 541
 Capacity: 302 gram s/sec. (40 Ibs./min.) @ 1:1 ratio
Maximum Viscosity: 1500 cps
Pattern Width, max: 762mm (30”)
Maximum pressure: 206 Bar (3000 psi.)

13m m S pray head:

Mix piston diameter: 13mm (.511”)
Capacity: 1664 gram s/sec. (220 Ibs./min.) @ 1:1 ratio
Maximum Viscosity: 1500 cps
Pattern Width, max: 1524mm (60”)
Maximum pressure: 206 Bar (3000 psi.)

All param eters are dependent on tip configuration.

A P P LIC A TIO N S :

The following are actual and possible applications for sprayed polyurethane processing:

A utom otive: All NVH (Noise, Vibration, and Heat) components such as carpeting, firewalls, trunk
surfaces, door panels, and headliners (Figure 7) are excellent candidates for rapid mold or
substrate coating and filling.

Figure 7: Headliner Demonstration

Furn itu re: Moveable office panels, carpet underlay, seating, and table surfaces.

B uilding P roducts: Modular roofing, garage door laminar panel manufacturing, moldings,
modular housing and mobile homes, and slab insulation (Figure 8).

Figure 8: 48” wide fiberglass panel demonstration

Leisure Item s: Pool float items, travel trailer or RV manufacture, and toys.

Tanks and P ipes: Thermal insulation, barriers, external and internal coating, and condensation
control.

542 / Bartenstein
A erospace and tra n s p o rta tio n : Fuel system insulation, fuselage structure and insulation, and
refrigeration trailer insulation.

SUM M ARY:

In other unique applications, Gusmer-Admiral, Japan has run tests in their laboratory coating
firewalls and other components without upper mold halves thereby gaining full coverage purely
through spray - up.

The process allows more complete “wet-out” of porous substrates such a s fiberglass mats or
carpet sw atches than conventional pouring methods. The recirculation feature allows improved
tem perature control over deadheaded, heated hose spray system s. No cold shots after dormant
periods are evident as experienced with non -recirculating systems.

Due to the ability to cover a wide area rapidly, use of faster reacting chemical system s are
possible that allow de -molding sooner than conventional RIM systems.

C O N C LU SIO N S :

Developments at Gusmer-Admiral are underway to use recirculating system s in the realm of

super low -output elastom er spray coating applications employing miniature components to mimic
the RIM devices used today.

These unique products and processes will offer the industry new applications previously not
accessible by stream pouring alone. Large surface coating (Figure 9) a s well as small mold
contours can be covered with varying thickness of polyurethane with standard RIM metering
system s and the new Gusmer-Admiral spray heads.

Figure 9: G u sm er -E urope dem onstration

B IO G R A PH Y

M ichael F. B artenstein

Michael F. Bartenstein has been involved in the production and engineering of polyurethane
spray foam and RIM equipment for 21 years at Gusm er Corporation in
Lakewood, New Jersey prior to becoming president of Gusmer-Admiral,
Akron, Ohio in January 1998 where he also augm ents R&D engineering and
sales efforts. Mike has been responsible for and assisted in the development
of improved CDC pumping system s, high -output plural component spraying
system s, innovative RIM metering system s, associative dry - side components,
and complete turnkey projects during his career. Mike is currently Director of
Global Engineering Services, Gusm er Machinery Group, parent of Gusmer -
Admiral, Inc.

Bartenstein / 543
 POSTER SESSION
FLEXIBLE FOAM/AUTOMOTIVE
Amine Emission Free Catalyst for Polyester Slabstock

R. VAN MARIS K. M. GAY

TOSOH Europe B. V. TOSOH USA Inc.
Crown Building South 1100 Circle 75 Parkway, Suite 600
Hullenbergweg 359 Atlanta, Georgia 30339-3097
1101 CP Amsterdam
The Netherlands

H. NAKAMURA a n d Y. TAMANO
TOSOH Corporation
Chemical Research Laboratory
4560 Kasei-cho, ShinNanyo City
Yamaguchi
ABSTRACT become an integral part of the polymer structure. It is
commonly understood that reacted amines will not emit
This poster will discuss the European automotive from foams. An earlier paper presented by Tosoh [1]
industry requirement to reduce the emission levels of illustrated that some reactive amine catalysts are indeed
amine PU catalyst. It will be revealed how by replacing the detected in foam emissions.
currently used catalysts (DMBA) by a new catalyst system Tosoh has previously introduced the reactive amine
based on reactive amines (Toyocat RX30) one can reduce catalysts Toyocat RX20 and RX21 for High Resilience
the amine emission and maintain good foam properties. Moulded Foam Formulations, based on either TDI/MDI
blend or all MDI, respectively [2]. This poster presents a
INTRODUCTION new reactive amine catalyst, Toyocat RX30, recommended
for polyester slabstock flexible foams.
Recently, European automotive manufacturers have
begun to regulate the use of conventional amine catalysts EXPERIMENTAL
in flexible foam systems. The goal in doing so is to
eliminate amine catalyst emissions from foams, thereby The new catalyst system Toyocat RX30 has been
reducing problems such as odour, PVC -staining and evaluated in polyester slabstock flexible foams. The
fogging, and contributing to the reduction of total Volatile formulation used for evaluation is given in Table 1.
Organic Compounds (VOC’s).
To overcome emission problems, the following special Table 1. Polyester Slabstock Foam Formulation
amine catalyst systems are proposed: Ingredient Amount (pbw)
Polyester polyol 1) 100
Low Volatile Catalysts (LVC) Silicone surfactant 1.1
Silicone emulsifier i.5
These are amine catalysts with high boiling points. Water 3.8
Higher boiling points implement lower emissions, and Catalyst Varied
hence a reduction in fogging is expected. However, LVC's, TDI 45.8
being non-reacted, remain in the free state and there exists Polyester polyol, OH value = 61.7 mgKOH/g.
therefore a possibility of catalyst release from the foam,
The foaming conditions are given in Table 2.
thus preventing the achievement of zero emissions.
Table 2. Foaming conditions.
Reactive Amine Catalysts Liquid temperature 25 °C
Mixing (5 seconds) 6000 rpm
Reactive amine catalysts contain hydroxyl or amino Mould dimensions (Wooden Box) 40 °C
functional groups, which react with isocyanate groups and Mould dimensions (Wooden Box) 25 x 25 x 25 cm
POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 547
RESULTS AND DISCUSSION

As mentioned in the introduction, the requirement is to

reduce amine emissions drastically. In order to make a
good selection from an emission point of view, the foam
samples were extracted with methanol for 48 hours at 65°C
and checked for emission. In Figure 1 the emission levels
of several non-reactive and reactive catalysts systems are
presented. As can be seen from the figure, most of the
evaluated catalysts emit from the foam after the heat
treatment was applied. The lowest emission of all however,
was shown by Cat B and also the newly developed
catalysts systems.
In the next stage, the foams were further evaluated on
reaction profile, physical properties and fogging/emission.
Figure 1. Emission of several evaluated amine catalysts in
The results of this evaluation are given in Table 3. Polyester Slabstock Flexible Foams

Table 3. Evaluation of reaction profile, physical properties and fogging/emission of reactive amine catalysts in Polyester
Slabstock Flexible Foam.

Catalyst DMBA NEM DMP NMP NP DM70 RX20 Cat A Cat B RX30
(pbw) 1.50 1.80 0.50 1.50 0.50 0.50 0.63 1.19 1.04 0.75
Reaction Profile (sec)
CT 13 13 13 12 13 12 12 13 13 13
GT 61 64 63 62 62 61 62 59 61 60
RT 72 76 83 75 80 77 76 77 82 75
Foam Properties
Core density (kg/m3) 30 31 29 31 30 30 28 30 29 31
CLD (N/m2)
25% 3.4 3.0 2.9 3.4 3.0 2.8 2.5 3.2 3.2 3.3
65% 10.0 8.2 7.5 10.6 6.8 8.5 7.0 7.6 8.0 8.5
Elongation (%) 174 162 153 212 155 210 166 210 189 198
Tensile strength (kN/m2) 127 120 103 142 110 129 120 138 128 129
Tear strength (kN/m) 0.86 0.83 0.62 1.08 0.70 0.86 0.90 0.73 0.73 0.83
Compression set (%)(*1) 22 23 20 21 20 16 43 21 28 21
Fogging value of amine (pg/g foam) (2) 180 200 60 180 60 50 10 35 10 14
Extraction (3) 8540 3450 300 3150 - - <1 220 <1 250
PVC Staining (4) none none none none - - none none none none
PVC discoloration: AE(%) 1.3 0.55 0.60 0.10 0.20
' '

1) 95% RH, 50% compression, 22 hr at 50°C, 50x50x25mm. 2) 100 degrees C, 48 hr. 3) Extraction in MeOH, 65 degrees C,
48 hr. 4) VW PV3937, 100 degrees C, 78 hr.

As can be seen from Table 3, the biggest differences The same trend was observed for the tensile strength and
occur in CLD, tensile properties and fogging value of tear strength, though the effect was less pronounced. As for
amine. compression set, it was discovered that all evaluated
Regarding CLD at 25% compression, similar values are amines came close to reference foam, with the exceptions
obtained for all non-reactive and reactive catalysts. At of RX20 and Cat B. No visual PVC discoloration was
65% compression however, a reduction in hardness of circa observed and therefore the relative colour change was
30%, especially for NP was obtained. With reference to the measured with a colorimeter. The results showed, even
tensile properties, such as elongation, DMP and NP though the absolute values were low, that RX20 and RX30
showed more than a 10% reduction of the value of the displayed a still lower level of discoloration.
reference foam. NMP and the reactive amines RX30 and In Figure 2, the emission of several amine catalysts are
Cat A even showed a 15% increase in value. compared.

5 4 8 / van Maris, Nakamura, Tamano, Gay

 type amine catalysts DMBA and NEM indicated very high
emission values compared to the reactive type amine
catalysts, which still give emission, although lower. In the
case of Toyocat RX20 and RX30 however, very low
emi ssion was detected.

CONCLUSIONS

TOYOCAT RX30, unlikely many reactive amine

catalysts, shows:
• high activity,
• good foam properties,
• no PVC discoloration,
• highly reduced amine emissions.

REFERENCES
Figure 2. The emission of several reactive amine catalysts,
analysed from the fogging deposit
1. H. Kometani, Y. Tamano, T. Masuda, K. M. Gay,
R. van Maris. Proceedings o f the API Polyurethanes
The evaluation o f amine emission was carried out by the
Conference 2000, Boston, 11-21.
fogging test method, DIN 75201 Gravimetric method. The
2. R. van Maris, H. Kometani, Y. Tamano, H. Nakamura,
deposit on the Al-foil was collected and analysed on the
K. Tokumoto. Poster Presentation at Utech 2000, The
specific amine. As can be seen in Figure 2, the standard Hague.

van Maris, Nakamura, Tamano, Gay / 549

DMAMP-80™: A Superior Reactive Amine Catalyst for
Polyurethane Foams

K. LITTEL a n d R. J . SWEDO F. M. CASATI, A. FANGET a n d H. MISPREUVE

ANGUS Chemical Company Dow Europe S.A.
1500 East Lake Cook Road International Development Center
Buffalo Grove, Illinois 60089 13, rue de Veyrot—P.O. 3
CH-1217 Meyrin 2 (GE)
Switzerland

ABSTRACT (polyvinyl chloride) and PC (polycarbonate) degradation

[2-4].
DMAMP-80, dimethylaimnomethylpropanol (DMAMP- Urethane foam catalyst suppliers have responded by in­
80 is a trademark of ANGUS Chemical) is an amino- troducing a variety of new, non-fugitive (sometimes called
alcohol with a structure similar to that of DMEA (N, N- reactive) amine catalyst compositions. However, the most
Dimethylethanolamine) indicating its potential as an alter­ widely used amine catalyst containing a reactive hydrogen
native to this compound. Compared to DMEA, DMAMP- in use today is still DMEA. This in spite of evidence that
80 is a catalyst with lower volatility and is not a lachryma- this product does not fully bind into the polymer network
tor. when used in foam production [5]. This amine catalyst is
When evaluated in four representative foam technologies also a known Iachrymator.
(flexible slabstock, HR molding, NVH and semi-rigid Data presented hereafter demonstrate that the substitu­
packaging foams) DMAMP -80 was shown to be a viable tion of DMEA by DMAMP -80 in flexible slabstock foam,
alternative to DMEA. Foams produced with DMAMP-80 HR (High Resilient) molding, NVH (Noise, Vibration,
had properties that were similar to those made with DMEA, Harshness) and semi-rigid packaging foam is straightfor­
but without odor at demold, as were the DMEA foams. No ward and yields several advantages that will be outlined.
PVC staining of NVH foam by DMAMP-80 was observed.
Overall, DMAMP-80 appears to be a viable alternative BACKGROUND
to DMEA as an amine catalyst for polyurethane foams.
Most commercially manufactured polyurethane foams
INTRODUCTION and elastomers are made with the aid of at least one tertiary
amine catalyst. Molecular structure of the amine and the
The Polyurethane industry is continuing its legacy of resulting steric hindrance impact how easily the active ni­
successfully reducing the amount of volatile organic trogen group can approach the isocyanate and polyol reac­
chemicals released from its operations. However, there is tants to facilitate the desired reactions. As commercially
still a need to eliminate amine vapor emissions. Indeed
most amine catalysts used today are fugitive products be­
cause of difficulties in finding new molecules with low
volatility, which are effective in catalyzing the various
polyurethane reactions.
The industrial hygiene and safety issues related to han­
dling and use of amine catalysts in polyurethane foams and
elastomers production is complex [1]. In addition, there is a
growing concern about emissions from manufactured arti­
cles releasing amine vapors due to potential fogging, PVC Figure 1. DMAM P and DMEA Molecular Structures

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 551

available DMAMP-80 is an 80% by weight solution of Table 2. Slabstock Foam Evaluation
DMAMP (2-dimethylamino-2-methyl-l-propanol) in water. Components DMEA DMAMP-80
The structure of DMAMP presented in Figure 1 shows
VORANOL 3040 100 100
that the minimal steric hindrance of the nitrogen group al­
lows it to be readily available to promote the polyurethane Silicone stabilizer 0.8 0.8
reaction. DMEA 0.12 0
A potential advantage of DMAMP-80 versus DMEA DMAMP 80 0 0.12
would be the low volatility of DMAMP with a vapor pres­
sure of ~ lmm Hg at 20°C [6]. The vapor pressure of NIAX A1 0.04 0.04
DMEA is 4 mm Hg. DMAMP-80 may also demonstrate Water 4 4
higher catalytic activity given its higher pKa 10.20 versus DABCO SO 0.2 0.2
9.31 for DMEA [7].
VORANATE T 80 52.4 52.4
The CAS registration and EINECS numbers of
DMAMP-80 are 7005-47-2 and 230-279-6, respectively. TDI Index 110 110
This product finds use as a synthesis intermediate, emulsi­
fying amine, and vapor-phase corrosion inhibitor. In addi­ Rise Time ( sec) 87 100
tion, DMAMP-80 has a low toxicity profile [7], is not a
lachrymator, and is FDA approved for indirect food contact
[8]. Physical Properties
Density (kg/m3) 25.3 24.5
EXPERIMENTAL CFD 40% (kPa) 5.04 4.32

The protocol for foam preparation and foam testing has IFD 40% (N) 195 181
been described in the literature [9]. Slabstock foams were SAG Factor 2.4 2.3
produced on a POLYMECH machine. Other foams were IFD Guide Factor 7.71 7.39
made by manual mixing. Formulations are reported in each
Air flow (cfm) 3.1 3.7
section of the paper.
The list of raw materials used in this study is reported in Tensile Strength (kPa) 103 109
Table 1. Elongation at Break (%) 104 132
SPECFLEX, VORANOL, VORANATE are trademarks
of The Dow Chemical Company. In addition, various
catalysts and surfactants were used as indicated in each
RESULTS AND DISCUSSION
section of this report. NIAX* is a trademark of Crompton
Corporation; DABCO* is a trademark of Air Products and
The objective of this work was to assess the performance
Chemicals Inc. of DMAMP-80 when substituted for DMEA in representa­
Foam properties were measured according to A.S.T.M.
tive foam formulations. Several foam technologies were
and D.I.N. test methods. PVC staining testing was carried
evaluated and each is reported hereafter separately.
out according to VW’s PV 3937 test procedure.
SLABSTOCK FOAM
Table 1. List of Raw Materials
The following data in Table 2 shows the effect of re­
Product Chemical Type Supplier placing DMEA by the same amount of DMAMP-80. De­
VORANOL 3040 Polyether polyol DOW spite the fact that the amount of DMAMP-80 catalyst is
u ii
lower on a molar base than the DMEA catalyst all foam
VORANOL CP1421 physical properties are virtually identical. Small differences
u it
VORANOL CP6001 in hardness (4 % of softening as expressed in guide factor
VORANOL 9815 II M
i.e. Hardness corrected for density) are not significant in
II ii view of the slight differences in airflow. The only notice­
VORANOL CP4755
II ii
able effect is the longer rise time that can be explained by
VORANOL CP450 the relatively lower amount of active molecules in the total
SPECFLEX NC700 SAN Copolymer DOW formulation. If needed, the amount of catalyst can be varied
polyol to fine-tune the reactivity of the system.
VORANATE T80 TDI Isocyanate DOW
SPECFLEX NS540 MDI Isocyanate DOW
VORANATE M229 DOW

552 / Littel, Swedo, Casati, Fanget, Mispreuve

 Table 3. H R Molded Foam Evaluation Table 4. N VH Fomtulation Evaluation
Components DMEA DMAMP-80 Components NIAX + DMEA DMAMP-80
DABCO
VORANOL 9815 76 76
VORANOL CP 6001 92.2 91.20 90.8
SPECFLEX NC 700 24 24
VORANOL CP 1421 3 3 3
Water 2.5 2.2
Water 3.80 3.80 3.2
Silicone stabilizers 0.65 0.65
Silicone stabilizer 0.50 0.50 0.50
Crosslinker 0.5 0.5
NIAX A1+DABCO 0.50 0 0
TDI/MDI Blend 47.7 47.7 33LV
DMEA 0.90 0 DMEA 0 1.50 0

DMAMP-80 0 1.50 DMAMP-80 0 0 2.50

SPECFLEX NS540
index 70 70 70
Cream time (s) 17 16
Gei time (s) 130 120 Gel Time (sec) 53 51 53
Rise time (s) 230 160 Rise Time (sec) 69
Free rise density (kg/m3) 45.7 45.4 Blow-Off (sec) 65 55
Mold fill time (s) 122 92 Settle (cm) 1 2
Molded density (kg/m3) 61.3 61.1 Free Rise Density 40.7 40.3 42.6
Amine Odor At Demold Strong Strong Weak

FLEXIBLE HR MOLDING
frame. The use of DMAMP -80 combined with Dow's ac­
tive polyol [11] and some Dabco T has allowed producing
Flexible HR molded foam pads were prepared on the
foams with densities below 10 kg/m3 with good properties
bench with formulations presented in Table 3. Demolding
as shown in Table 5.
time was 4 minutes with a mold heated at 55°C. When
DMEA was used a strong and unpleasant smell could be
CONCLUSION
noticed when taking the pad out of the mold, indicating that
amine vapors were still present in the freshly made foam.
Data presented in this report, based on four representa­
When DMAMP-80 was used no amine odor was observed.
tive foam technologies and with formulations, containing
Skin of foam based on DMEA was still tacky at demold
very low, medium and extremely high water levels and
while foam based on DMAMP-80 was well cured.
different isocyanates, have demonstrated that DMAMP-80
is a viable alternative to DMEA. The properties of the test
NOISE VIBRATION HARSHNESS (NVH) FOAM
foams were similar and even improved when made using
DMAMP-80.
Flexible polyurethane foam is the material of choice in
Given the superior environmental and handling profile of
automotive noise, vibration and harshness control due to its
DMAMP-80, there is a clear incentive to the foam formu-
ability to be molded in thin and intricate sections (10).
lators to consider switching to DMAMP-80 in their polyu­
Comparative NVH foams produced with either conven­
rethane manufacturing plants when looking for hygiene
tional catalysts, DMEA or DMAMP -80 and demolded one
improvements and for lower amine emissions.
minute after pouring showed that the latter is the catalyst of
choice due to low emission of amines at demold. Full data
REFERENCES
are reported in Table 4 below.
Furthermore PVC staining tests confirmed that
1. Polyurethane Amine Catalysts: Guidelines for Safe
DMAMP-80 was not released during foam heat aging
Handling and Disposal; Technical Bulletin AX173,
while the PVC of the control foam was heavily stained.
November 2000; Alliance for the Polyurethane Indus­
try: Arlington, VA.
PACKAGING FOAM
2. Williams K.P.; J. Gerrard, D.L. ‘T he Degradation of
Poly (Vinyl Chloride) Studied using Fourier-
Packaging foams, used to protect goods during transpor­
Transform Ramam Spectroscopy” (1990), Eur. Polym.
tation, are low-density foams produced with formulations
J.,26(12)-1355-8.
based on very high water levels. These polyurethane prod ­
3. Wilson A.S.; Gerrard D.L.; Bowley H.J. “Problems of
ucts require a high concentration of a catalyst package
Interaction Between Plasticised PVC and Polyurethane
since the foam has to rise and cure in a very short time

Littel, Swedo, Casati, Fanget, Mispreuve / 553

 Table 5. Packaging Foam Evaluation
Components DMEA DMAMP-80
VORANOL CP 4755 34
Experimental Polyol B 34
VORANOL CP 450 14 14
Compatibilizer Additive 14 14
PEG 600 10 10
Water 25 25
DMEA 1.50 0
DMAMP-80 2.0
DABCOT 0.50 0.50
DABCO 33LV 0.50
Silicone stabilizer 0.50 0.50

VORANATE M229 100 200 300 400 100 200 300 400
Index 25 50 75 100 25 50 75 100
Free Rise Density (kg/m3) 7.5 9.2 7.0 7.0
Shrink No No Yes Yes No No Yes Yes
Blend Stability @72 hours, 23°C Clear Clear

11. Casati F.M.; Herrington R.; Sonney J.M.; Tu Johnson;

Foam” (1987). Presented at P V C 87, Brighton, UK., Mispreuve H.; Fanget A.; (2001) “Elimination of
April 1987. Amine Emissions From Polyurethane Foams: Chal­
4. Christfreund A.; Huygens E.; Eling B. “Amine-Free lenges and Opportunities”; Proceedings of API
Catalyst Systems for Automotive Instrument Panels” 2001Conferences.
(1991), Proceedings of Polyurethane World Congress,
Technomics: Lancaster, PA., 272-277. BIOGRAPHIES
5. Nutt A. R.; Skidmore D.W. (1987) “Recovery of Chlo­
rofluorocarbon 11 by Activated Carbon Scrubbing on Kenneth J. Littel
a Polyurethane Foam Slabstock Plant”, Cellular Poly­ Kenneth Littel graduated from Iowa
mers, Volume 6, 1987,62 -77. State University with a Master of
6. ANGUS Chemical Company Technical Bulletin, TB Science degree in Microbiology. He
56, DMAMP-80 An Amine Solubilizer For Water- also holds a Master of Business
Borne Polymers. Administration, Marketing degree
7. ANGUS Chemical Company Technical Data Sheet, from DePaul University. He has
TDS 18, DMAMP-80. worked as a research microbiologist
8. ANGUS Chemical Company Marketing Benchmark, developing rapid microbial
ABM 4, FDA Clearances For AMP and DMAMP. contamination detection systems. At
9. Herrington R., De Genova R., Casati F., and Brown ANGUS Chemical Co. he has over 10 years of experience
M. (1997) “Molded Foam, Chapter 11”, in Polyure­ as a product and market development manager in industrial
thane Flexible Foams; Herrington, R. and Hocks, K. biocides. In addition he also served as country manager for
eds. The Dow Chemical Company, Form No 109- sales of ANGUS products into India. He is currently a New
01061, 177-182. Business Development Director for ANGUS.
10. Broos R.; Sonney J. M.; PhanThanh H.; Casati F. M.;
“Polyurethane Foam Molding Technologies for Im ­
proving Total Passenger Compartment Comfort”
(2000) Proceedings API 2000 Conferences, p.31.

554 / Littel, Swedo, Casati, Fanget, Mispreuve

Raymond J. Swedo Alain Fanget

Raymond J. Swedo received a Ph.D. Alain Fanget joined the Geneva-

in organic chemistry in 1977 from based Polyurethane Research and
the University of Arizona. He has Development group of Union
over 20 years of industrial Carbide Europe in 1978 and was then
experience in the synthesis and transferred to BP Chemicals in 1979
applications of monomers, polymers, and to the Dow Chemicals Company
membranes, coatings, chemical in 1989 where he worked in various
intermediates, and catalysis with PU molding and ACES applications.
DeSoto, UOP, AlliedSignal, and He is currently a Technical Service
ANGUS Chemical Co. He is currently a Senior Product Specialist in the Automotive Thermosets R&D Group of
Research Specialist in the Product Development group of the Meyrin International Development Center, with specific
ANGUS. responsibilities in seating and NVH molding applications.

Francois M. Casati

Francois M. Casati graduated from

ICPI (F), now CPE-Lyon (Ecole
Superieure de Chimie Physique
Electronique de Lyon), in 1967. He
has over 30 years of experience in
Polyurethanes, Industrial Amines
and Biocides, with S.N.P.E.,
Recticel, Abbott Laboratories, BP
Chemicals and The Dow Chemical
Company. During that time he has held various positions in
Manufacturing, Marketing and Research & Development.
He is currently a Product Development Leader, R&D, for
the Flexible Foam business of Dow.

H enri M ispreuve

Dr. Henri Mispreuve graduated

from the State University of Mons
(Belgium) in 1975 and received his
Ph.D. in Physical Organic
Chemistry from the same University
in 1979. In 1984, he joined Dow
Europe S.A. in Horgen
(Switzerland). He is currently
responsible for Technical Service
and Development of Polyurethane Slabstock Foams at the
company’s International Development Center in Mey-
rin/Geneva (Switzerland).

Littel, Swedo, Casati, Fanget, Mispreuve / 555

Effect of Nano and Micro-Silica Fillers on Polyurethane
Foam Properties

I. JAVNI, W. ZHANG, V. KARAJKOV a n d V. DIVJAKOVIC

Z. S . PETROVIC University o f Novi Sad
Kansas Polymer Research Center
Pittsburg State University
Pittsburg, Kansas 66762

ABSTRACT complex. It consists of discrete hard -segment domains

dispersed within a continuous soft polymer phase.
Two series o f polyurethane foams both rigid and
flexible were prepared with two types of silica fillers. Mineral fillers are sometimes used as reinforcing
The density of the flexible foams was 60 kg/m3 and agents in polymeric materials. In polyurethane foams
that of rigid 30 kg/m3. The fillers were micro silica, of their role is much more complicated. They can affect
the average particle size of 1.5 pm and nano-silica of the chemical reaction of the polyurethane formation,
the average particle size of 12 nm. The concentration macroscopic cell geometry or interact with the soft
of fillers varied from 0-20%. The micro-silica filler segments of flexible foams giving a higher modulus.
did not show any significant effect on density o f both Reactive groups on the surface can react with
rigid and flexible foams. Nano-silica increased the isocyanate and change the balance in the isocynate-
density of both types of foams only at concentration polyol reaction. At higher concentrations they increase
above 20%. Nano-silica lowered the compression viscosity o f the reaction mixture and affect the cell
strength of both types of foams, at all concentrations growth process changing the cell geometry.
while micro -silica exhibited the same effect at
concentrations above 10%. It can be expected that the nano-size fillers with an
extremely large surface per unit volume would affect
The hardness and compression strength in flexible PU foam properties much more than regular fillers.
polyurethane foams with nano-silica was increased The objective of this project was to examine the effect
and the rebound resilience decreased. Reduced of nano-silica on the properties of rigid and flexible
density of foams filled with nano-silica was not foams. For comparison we have also selected a filler
changed up to 20% of the concentration. It is assumed with an average particle size of 1.5 micron.
that the nano-filler as an additional physical
crosslinker increased modulus of the flexible segment
in the polyurethane matrix, resulting in increased EXPERIMENTAL
hardness and compression strength. The micro-filler in
flexible foams lowered hardness, and compression M aterials
strength, but increased rebound resilience.
Flexible polyurethane foams were prepared from
Wide angle x -ray scattering (WAXS) showed Multranol 3400 (triol from Bayer, MW =3000) and
amorphous morphology of both flexible and rigid crude diisocyanate Mondur 448 (Bayer). The rigid
foams filled with nano-silica. WAXS of the micro­ foam samples were prepared from soybean oil polyol
silica filled foams showed the presence of randomly Soypolyol 206. This polyol was synthesized in our
oriented crystalline quartz particles and the amorphous laboratory according to the proprietary procedure. The
structure o f the polymeric matrix. functionality of the polyol was 4 and molecular weight
1090. The isocyanate component was Mondur MRS-5
INTRODUCTION (Bayer). Water was used as a blowing agent in all
cases. Nano-silica, having a particle diameter of about
Polyurethane foams are solid-gas composites in 12 nm, was supplied by Nissan Chemical Co. as 30%
which the continuos solid phase provides structural dispersion in methylethylketone. Crystalline silica
mechanical characteristics and, in closed cells of rigid (micro-silica) Min-U-Sil 5 (average particle size was
foams, containment for the gas phase[l]. The 1.5 pm (1500 nm) was supplied by US Silica.
mechanical properties of polyurethane foams depend
essentially on the matrix structure and shape
(morphology) of the cells[2]. The polyurethane matrix
is based on a polyol and an aromatic isocynate.
Morphology of the flexible foam matrices is rather
POLYURETHANES EXPO 2001 , SEPTEMBER 30 -OCTOBER 3, 2001 557
 Methods

Polyurethane foams were prepared by mixing both

The polyol and isocynate components with a
mechanical stirrer. In the first step the nano-silica
solution was mixed with polyol and the solvent was
removed by vacuum distillation. In the second step all
other ingredients were added to the polyol (surfactant,
catalysts and water as the blowing agent) and
thoroughly premixed. Then the polyol was mixed with
isocyanate and poured into an open mold to produce
the free raised foam. The filler concentration varied
from 0-20%. The structure of the foams was analyzed
by WAXS. The mechanical properties were measured
on the Q-Test2 machine from MTS.The cell
dimensions and the micro-silica filler dispersion in the
polymer matrix were determined by optical
microscopy.
Figure 2. Effect of the micro- and nano-silica fillers on
RESULTS AND DISCUSSION compression strength of rigid foams at 10%
strain
Rigid foams
Because of that, the higher concentrations of the
The rigid PU foam matrix is a single phase, highly surfactant and catalyst were used during foam
crosslinked polymer. In order to compare density of preparation.
the foams with the different concentration of fillers we
subtracted the weight of the filler to obtain a so called Compression strength of the rigid polyurethane
“reduced density”. The micro-silica filler did not foams decreased with increased filler concentration.
change the reduced density of the foams significantly, Micro-silica showed a slight reinforcing effect at the
Figure 1. The nano-silica filler at a concentration concentrations of 5 and 10% but at higher
above 10% increased the reduced density. It increased concentrations the strength of the foams was lower,
viscosity much more than micro-silica filler and Figure 2. The loss of the strength at higher filler
showed a strong inhibition effect on the foam growth. concentrations could be assigned to the cell
irregularity caused by imperfections of the struts due
to the micro-silica presence. Compression strength of
the samples with nano-silica was decreased much
more than the strength of the foams with the micro­
silica filler, which may have been caused by both
chemical and physical factors.

Flexible foams

The polymer matrix of flexible polyurethane foams

has a segmented structure with separated flexible
polyether segments and hard urea segments. Hard
segments form rod-like domains with an average
length of 2-5 nm and an average diameter of 3-5 nm.
At higher concentrations these segments are
agglomerated to polyurea balls with an average
diameter of about 300 nm, playing the role of the
physical crosslinkers[2]. The fine nano-silica particles
of an average diameter of 12 nm, are probably
Figure 1. Reduced density of rigid P U foams with randomly dispersed across the polyurethane matrix
micro- and nano-silica along with polyurea hard segments. They interact
with the soft segment decreasing their mobility and
increasing modulus and Tg, playing the role of an
active filler. It is interesting that the addition of nano­
silica had the opposite effect, Figure 3. The reduced

558 / Javni, Zhang, Karajkov, Petrovic, Divjakovic

density of the foams filled with nano-silica filler was indicating a higher energy dissipation in nano-silica
not changed in up to 20% of the concentration, Figure filled foams.
4. The compression strength, Figure 5, at the
beginning decreased, probably as a result of the cell
geometry imperfection and then increased as a result
of the higher density and thicker struts. Rebound
resilience also decreased, Figure 6.

Micro-silica as the filler contributed to a lower

hardness of the flexible foam samples, Figure 3. The
reduced density of the samples filled with micro-silica
was not affected by the presence of the filler, Figure 4.
The micro-silica filler with an average particle of 1400
nm can be considerably larger than the thinner part of
the struts, which is about 500 nm. During the rising of
the foam the filler particles are probably located in the
Plateau border area, which has an average diameter of
50,000 nm. This is an area or space between several
cells and it is formed as a result of the redistribution of
the reacting polymer during the cell growth. In the
initial phase of foaming the cells have a spherical
shape. Once they occupy more than 74% of the
volume they are distorted into multi-sided polyhedra.
The liquid phase is redistributed between the Figure 4. Reduced density o f flexible PU foams with
tetrahedral interstices and the bubble surfaces. With micro- and nano-silica
continued expansion of the cells, the wall-thinning
process takes place and the reacting polymer is
drained from the windows and struts into the Plateau
border area. Presence of the micro-filler in this area
can cause an additional drainage.

Figure 5. Effect of the filler type on compression

strength of flexible PU foams at 10% of
strain

WAXS difractograms of the both PU foam samples

flexible and rigid with nano-silica as the filler, shows
Figure 3. Effect of the filler concentration on hardness an amorphous morphology of the foam samples with a
of flexible foams strong but diffuse maximum of the dissipation at 20
=19 degrees of the Brag’s angle, Figure 7.
This could cause further thinning of the cell struts
that decrease the hardness, Figure 3, and compression
strength, Figure 5.Figure 6 shows that the rebound
resilience is higher in foams with micro-silica

Javni, Zhang, Karajkov, Petrovic, Divjakovic / 559

 Figure 6. Effect of the filler type on rebound resilience Figure 8. WAXS scan of the PU foam sample with
of flexible foams 20% of the micro-silica filler

CONCLUSIONS

Micro- and nano-silica fillers, although of the same

chemical origin, showed considerably different effect
on the rigid and flexible polyurethane foam properties.

Hardness and compression strength of flexible

polyurethane foams with nano -silica were increased
and rebound resilience decreased. Reduced density of
the foams filled with nano -silica filler was not
changed up to 20% of the concentration. It is assumed
that nano-filler as an additional physical crosslinker
increased modulus of the flexible segment in the
polyurethane matrix, resulting in a higher hardness
and compression strength.

The micro-silica filler in flexible foams decreased

the hardness and compression strength, but increase
rebound resilience of the foams.

LITERATURE

Figure 7. WAXS difractogram of PU foam with 15% of 1. D. Klempner, K.C. Frisch, Handbook of
the nano-silica filler Polymeric Foams and Foams Technology, Hanser
Pub., N. York, 1991
Rigid and flexible PU foams with micro-silica as a 2. R. Herrington, K. Hock, Flexible Polyurethane
filler show difractogrames with relatively broad peaks Foams, Dow Plastics, 1997
at Brag’s angles at 20 = 20.9° i 26.7° (with spacing of
0.34nm and 0.33 nm respectively). These angles BIOGRAPHIES
corresponds to the maximums of crystalline quartz,
Figure 8. Ivan Javni, Ph.D.

The peaks broadening of S i0 2 is caused by low Dr. Ivan Javni received his B.Eng. and M.Sc. at the
density of PU foams and thus by small x-ray Faculty of Technology, Tuzla, and Ph.D. at the
absorption of the sample. University of Novi Sad, Yugoslavia. He held a

560 / Javni, Zhang, Karajkov, Petrovic, Divjakovic

 teaching position at the Vladimir K arajkov
Faculty of Technology,
Tuzla, from 1971-1981. Vladimir Karajkov received his B.Eng. at the
In 1981 he joined Soda-So Faculty of Technology, University o f Novi Sad,
Co., a major producer of Yugoslavia and M.Sc. at the University of Pittsburg,
urethane components and USA. His special interests are application of synthetic
systems in Yugoslavia, and resins in production of foams, coatings and lacquers.
was the Head of the He is currently production manager at Hemiprodukt, a
Polyurethane Institute. He paint and coating company from Yugoslavia.
was also selected UNESCO
expert for polyurethanes. Z oran S. Petrovic,
He is currently a Research Ph.D.
Scientist at KPRC, Pittsburg State University,
Pittsburg, Kansas. He has published about 30 papers, Dr. Zoran Petrovic
and has several patents and patent applications. His received his B. Eng. at
special interests are polyurethanes and elastomers in the Faculty of
general. Technology, Belgrade
University, Yugosla­
Vladim ir Divjakovic via, and Ph.D. at the
Strathclyde Univer­
Vladimir Divjakovic received his B.Sc. in Physics, sity, Glasgow. He
Faculty o f Sciences, University of Belgrade, worked in the electrical
Yugoslavia and Ph.D. at the University of Bern, industry for 12 years
Switzerland. He is currently Professor at University of on problems of appli-cation of polymers in electrical
Novi Sad, Yugoslavia. His special interests are crystal insula-tion. From industry he moved to the Faculty of
structures of minerals, some transition metal Technology, University of Novi Sad, Yugoslavia,
complexes and morphological investigation on where he taught Physical Chemistry o f Polymers. He
crystalline polymers. He has published over 70 papers spent three years at the Polymer Science and
and made over 50 oral and poster presentations. Engineering Department, University of Massachusetts,
at Amherst, working on different research projects. He
Wei Zhang, M.Sc. is currently Director of the Kansas Polymer Research
Center (KPRC) and Professor of Plastics Technology
at Pittsburg State University, Pittsburg, Kansas. He
Wei Zhang received his B.Sc. has published over 100 papers, 3 books, and has seven
degree from Zhejiang patents. His special interests are polyurethanes, bio­
University, P. R. China and based polymers, networks and gels, blends and
M.Sc. degree in chemistry polymer characterization.
from Pittsburg State
University, Pittsburg, Kansas.
He held a teaching and
research position at the
Capital Normal University,
Beijing, P. R. China for 12 years where he taught
Organic Chemistry and Polymer Chemistry. He also
worked on organic and polymer synthesis. He is
currently a Research Associate at KPRC, Pittsburg
State University, Pittsburg, Kansas.

Javni, Zhang, Karajkov, Petrovic, Divjakovic / 561

Effect of Particle Size Distribution on Performance of
Copolymer Polyols

RAYMOND NEFF a n d WASSANA APICHATACHUTAPAN

BASF Corporation
1609 Biddle Ave.
Wyandotte, Michigan 48192

ABSTRACT BIOGRAPHIES

Model poly(styrene-acrylonitrile) (SAN) disper­ Raymond Neff

sions were prepared in ethylene oxide-capped polyether
triols by two different methods, and evaluated in a high Ray Neff received his B.S. in Chemical Engineering from
solids, high resilience (HR) molded foam formulation. the University o f Illinois in 1990, and his Ph.D. in Chemi­
The first preparation method involved blending of two cal Engineering from the University of Minnesota in 1995.
nearly monodisperse dispersions of different average par­ After a post-doctoral assignment at the National Institute
ticle size. The second method involved preparation of a of Standards and Technology, Dr. Neff joined the BASF
polydisperse dispersion, and fractionating particles of Corporation in 1997. Presently he is a Research Chemist
various sizes by ultracentrifugation, thus eliminating pro ­ in the Urethanes R&D Applications Development group,
cess chemistry as a variable. However, while the particle where he develops technology for both flexible and rigid
size distribution of the centrifuged dispersions varied sig­ polyurethane foam.
nificantly, the method did not cleanly isolate large and
small particles. Foam tightness was measured using a W assana Apichatachutapan
force to crush test. A standard set of physical properties
was also measured. Wassana Apichatachutapan received her B.S. in Materials
Foam tightness strongly correlated with the par­ Science from Chulalongkom University, Thailand, in 1991
ticle size distribution, and specifically with the relative and her Ph.D. in Polymer Science and Engineering from
number o f large (> 5 pm) particles. Dispersions contain­ the University of Southern Mississippi, MS, in 1999. She
ing predominately small particles produced relatively joined BASF Corporation in July 1999, working as a
open foam (low force to crush), while even a relatively Product Specialist in the Styrenic Copolymers group and
small proportion (< 20 vol. % of solids population) of as a Research Chemist in the Fibers R&D group. Cur­
large particles resulted in significantly tighter foam (high rently, she works in the Urethanes R&D Applications De­
force to crush). Tear, elongation and compression sets of velopment group as a Research Chemist responsible for the
foams prepared with the blended dispersions correlated development of isocyanate raw materials and polyurethane
directly with the blend composition, while hardness var­ systems for flexible and rigid foams. She has been in­
ied only slightly with the particle size distribution. Only volved in establishing structure-property relationships for
slight variations o f foam physical properties with particle molded seating foam products. Her responsibilities also
size distribution were found using the centrifuged disper­ include providing quality assurance and technical support
sions. to manufacturing and business groups.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 563

Flexible Polyurethanes Using a Castor Oil Based
Polyester Polyol

PIOTR PENCZEK, RYSZARD OSTRYSZ, S . V. UDESHI

DOROTA WIECZOREK a n d Jayant Oil Mills
ZBIGNIEW BONCZA -TOMASZEWSKI 13 Sitafalwaldi, Mazgaon
Industrial Chemistry Research Institute Mumbai 400 010
ul. Rydygiera 8, 01-793 Warszawa India
Poland

ABSTRACT The scope of this work comprised:

1. Preparation of urethane elastomers obtained from the
A polyester polyol from castor oil, adipic acid and castor oil based polyester polyol, 1,4-butanediol and toly-
polyhydric alcohols for flexible polyurethane foams and lene diisocyanate (TDI) at different molar ratios of the
cast urethane elastomers was developed. polyester polyol to 1,4-butanediol and a stoichiometric
Flexible polyurethane foams were prepared using the amount of TDI.
polyester polyol, TDI and auxiliary additives. Properties 2. Preparation of urethane elastomers obtained from the
of the foams such as density, stiffness, hardness, perma­ castor oil based polyester polyol, Poles 60/20 (a polyester
nent set, tensile strength, elongation at break, resilience polyol from adipic acid and ethylene glycol), 1,4-
and permeability were determined. The foams may be butanediol and TDI at different molar ratios of polyester
applied as sealing, damping, packaging and pipe insulat­ polyols to 1,4-butanediol and stoichiometric amount of
ing materials. They can also be used for laminating with TDI.
fabrics and as upholstery but to a limited extent only. The
3. Preparation of urethane elastomers of high hardness
castor oil based polyester polyol can also be added to
using glycerol as the low molecular chain extender and
formulations containing standard polyester polyols to
the branching (crosslinking) agent.
lower the costs.
4. Determination of the application properties of the ure­
Cast urethane elastomers of various Shore hardness
thane elastomers obtained, such as: hardness, tensile
were prepared using the castor oil based polyester polyol,
strength, elongation at break, abrasion resistance, and
a diisocyanate (TDI or MDI), and various amounts and
resistance to solvents (distilled water, isopropyl alcohol,
types of low molecular chain extenders (ethylene glycol,
1,4-butanediol, toluene, butyl acetate).
1,4-butanediol, glycerol). Formulations containing mix ­
tures of the castor oil based polyester polyol and a com ­ EXPERIM ENTAL
mercial polyester polyol obtained from adipic acid and
Materials
ethylene glycol were also tested. Mechanical properties of
1. The developed castor oil based polyester polyol
the elastomers such as tensile strength, elongation
at break, abrasion resistance and Shore hardness were a) raw materials, g :
determined. Chemical resistance to various media (water, castor oil 2115.5
toluene, isopropanol, 1,4-butanediol, butyl acetate) of the adipic acid 1700.0
elastomers was also studied. The elastomers are intended diethvlene glvcol 1387.0
for use as printing rollers and paint applicators. Total 5202.5

STATEMENT OF OBJECTIVE b) reaction temperature 205-210 °C

The aim of these studies was the optimization of formu­ c) reaction time 38 h
lation of cast urethane elastomers for printing rollers and d) properties of the polyester polyol:
evaluation of the application properties related to this acid number 1.75 mg KOH/g
application and actualized costing. The optimization was hydroxyl number 64.0 mg KOH/g
performed for elastomers of different Shore A hardness: viscosity at 25 °C 4200 mPas
a) in the range of 20 - 30 density at 25 °C 1.070 g/cm3
b) in the range of 40 - 50
c) in the range of 60-70
d) above 80 (for the rice-mill roller application)
POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 565
possible the preparation of urethane elastomers on the Testing methods
laboratory scale. a) Shore A hardness determination according to Polish
The polyester polyol of lower viscosity in comparison Standard PN-80/C-04238 (ISO 868)
with the polyester polyol for flexible polyurethane foams b) Tensile strength and elongation at break according to
was obtained in order to lower the viscosity of the pre­ ISO 527-1, section 4
polymer mixture to the level suitable for filling the molds c) Abrasion resistance (Schopper) according to ISO 4649
used. d) Chemical resistance (weight changes) according to
2. Poles 60/20 - (Chemical Works „Organika Zachem”, Polish Standard PN-78/C -89067.
Bydgoszcz, Poland ) - adipic acid and ethylene glycol RESULTS AND DISCUSSION
polyester diol
Table 1 presents the formulations and properties of the
acid number 1.8 mg KOH/g elastomers prepared from TDI, the castor oil based poly ­
hydroxyl number 59.6 mg KOH/g ester polyol and 1,4-butanediol.
viscosity at 75 °C 568 mPas The urethane elastomer obtained from TDI and the
melting point 54 °C castor oil based polyester polyol without 1,4-butanediol
3. 1,4-Butanediol (Hyperlast, Great Britain) (Sample No. I) exhibits relatively poor mechanical proper­
ties. Addition of 1,4-butanediol, a low molecular chain
4. Glycerol, 98 % (POCh, Gliwice, Poland) extender, causes an increase in hardness, tensile strength,
5. Tolylene diisocyanate (TDI) - Izocyn T-80 (Chemical elongation at break, and resistance to solvents (Sample
Works „Organika Zachem”, Bydgoszcz, Poland ), a mix­ No. II-V). The higher amount of 1,4-butanediol the higher
ture of 2,4-tolylene diisocyanate (80 %) and 2,6-tolylene increase in the mechanical parameters and chemical resis­
diisocyanate (20 %); the diisocyanate content - at least tance. Increasing the NCO/OH ratio causes a distinct
99.7 % improvement of the application properties of elastomers
(compare Sample No. IV to XI, and V to X). We recom ­
mend the formulations IV and XI as the application for­
Preparation o f elastomers mulations for elastomers of Shore A hardness in the range
a) Dehydration o fpolyols
of 20-30 and the formulation X as the application formu­
The mixture of the polyester polyol(s), low molecular
lation for elastomers of Shore A hardness in the range of
chain extender and orthophosphoric acid, 0.05 wt. % in 40-50.
relation to the polyester polyol(s), is heated up to 110-115
Urethane elastomers of higher hardness were obtained
°C under reduced pressure (below 50 hPa) with continu­ using glycerol or glycerol/1,4-butanediol as the low mo­
ous stirring for 1 hour. Next, the mixture is cooled down lecular chain extender (Table 2, Sample No. VI and XII -
below 40 °C, maintaining reduced pressure and stirring XIV). Elastomers with Shore A hardness above 80 and
during cooling. distinctly improved chemical resistance were obtained
Orthophosphoric acid is used to lower the reactivity of (Sample No. VI, XII and XIV). For elastomers of Shore A
polyester polyols. Lowering of the reactivity is not neces­ hardness in the range of 60-70, we recommend the formu­
sary when urethane elastomers are prepared using profes­ lation XIII, and for the elastomers of Shore A hardness
sional mixing-dosing equipment (e.g., SECMER type).
above 80 (e.g. for the rice-mill roller application) - the
b) Synthesis o fprepolymer formulations XII and XIV.
TDI is added to the dehydrated mixture of polyols, the
Urethane elastomers using TDI and mixtures of the
whole mixture is stirred for 10 minutes and then the pres ­
castor oil based polyester polyol with Poles 60/20 (adipic
sure in the reactor is reduced to the minimum possible
acid and ethylene glycol polyester diol) were obtained.
value. After 5-10 minutes, the stirring is turned off. Addi­
Their formulations and properties are presented in Table 2
tion of TDI causes the rise of the temperature of the reac ­
(Sample No. VII-IX). Addition of Poles 60/20 improves
tion mixture up to 70-80 °C. When the reaction mixture
the mechanical properties and chemical resistance of ure­
(the prepolymer) is degassed, it is poured into molds lu­
thane elastomers in comparison with the elastomer ob ­
bricated with a silicone wax (SILPASTA A, „Organika- tained from TDI and the castor oil based polyester polyol
Sarzyna”, Poland) and preheated to 70-80 °C. only (Table 1, Sample No. I). Increasing amount of Poles
c) Curing 60/20 causes an increase in hardness and tensile strength,
The curing was performed at 90 °C for 1 hour, then at
but the elongation at break practically does not change (it
115 °C for 2-4 hours. decreases for the sample IX with the highest amount of
The time needed for complete curing depends on the Poles 60/20). The elastomers obtained with an addition of
dimensions of the cast sample, the composition, NCO/OH
Poles 60/20 are not completely transparent because they
ratio and reactivity of the reagents. The curing time should
contain an amorphous phase formed by the castor oil
be determined for particular articles (dimensions and
based polyester polyol and a crystalline phase formed by
composition). Usually, the curing is carried out until the
Poles 60/20. The higher amount of Poles 60/20 the less
hardness of the elastomer stops increasing. transparent elastomers were obtained.

566 / Penczek, Ostrysz, Wieczorek, Boncza-Tomaszewski, Udeshi

Table 1. Formulations andproperties of the elastomers obtainedfrom TDI, the castor oilbasedpolyesterpolyol and 1,4- butanediol
Sam ple No. I II III IV V X XI
P aram eter
/. Composition
1. C astor oil - based polyester polyol, mole OH 1.00 1.00 1.00 1.00 1.00 1.00 1.00
2 . 1,4 - Butanediol, mole OH - 0.33 0.66 1.00 2.00 2.00 1.00
3. TDI (Izocyn T - 80), mole NCO 1.00 1.33 1.66 2.00 3.00 3.15 2.10
4. NCO/OH 1.00 1.00 1.00 1.00 1.00 1.05 1.05
//. Mechanical properties
1. H ardness, Shore A 18.3 27.0 20.1 24.3 37.8 40.5 24.7
2. Tensile strength, MPa 0.61 0.69 0.63 0.69 1.25 1.88 1.23
3. Elongation at break, % 94.3 146.0 217.0 220.0 227.0 240.0 203.0
4. Abrasion resistance, mm3 (Schopper) *) *) *> *) 491 322 *)
III. Chemical resistance, weight increase (%)
1. Distilled w ater
after 6 h 0.33 0.30 0.29 0.27 0.18 0.19 0.28
after 24 h 0.54 0.53 0.48 0.45 0.37 0.37 0.51
2. iso-Propyl alcohol
after 6 h 3.10 2.94 2.86 2.46 1.86 1.54 2.13
after 24 h 6.47 5.73 5.11 5.11 3.89 3.39 4.79
3 . 1,4 - Butanediol
after 6 h 16.15 12.85 12.96 11.90 9.26 7.44 10.10
after 24 h 31.17 24.88 27.10 23.94 18.26 15.34 20.71
4. T oluene
after 6 h 85.87 69.12 72.08 61.60 45.77 36.80 51.41
after 24 h **) **) **) **) **) 74.50 106.89
5. Butyl a cetate
after 6 h 63.89 44.82 52.87 43.84 30.00 23.00 33.38

after 24 h **) **) **) **) **) ** **)

Table 2. Formulations and properties of the elastomers obtained from TDI, the castor oil based polyester polyol or its mixture with
Poles 60/20, extended with 1,4- butanediol and/orglycerol._____________________________________________________
Sam ple No. VI XII XIII XIV VII VIII IX
P aram eter
1. Composition
1. C astor oil - based polyester polyol, mole OH 1.00 1.00 1.00 1.00 1.00 1.00 1.00
2. P o les 60/20, m ole OH - - - - 0.28 0.50 0.90
3. Glycerol, mole OH 3.37 1.75 0.83 2.30 - - -

4 . 1,4-Butanediol, m ole OH - 1.15 2.27 1.50 - - -

5. TDI (Izocyn T - 80), m ole NCO 4.37 4.00 3.89 4.80 1.25 1.50 2.00
6. NCO/OH 1.00 1.02 0.95 1.00 0.975 1.00 1.05
II. Mechanical properties
1. H ardness, S hore A 80.9 86.9 65.7 86.0 29.5 41.5 62.3
2. Tensile strength, MPa 8.59 7.23 4.22 - 1.04 1.62 1.96
3. Elongation at break, % 50.4 85.0 271.0 - 201.0 200.0 162.0
4. Abrasion resistance, mm3 (Schopper) 373 292 285 272 *) *) 1041
III. Chemical resistance, weight increase (%)
1. Distilled w ater
after 6 h 0.23 0.18 0.22 - 0.31 0.28 0.15
after 24 h 0.46 0.37 0.43 - 0.49 0.55 0.38
2. /soPropyl alcohol
after 6 h 1.11 0.17 1.62 2.39 1.95 1.31
after 24 h 2.32 2.61 3.67 - 4.77 4.27 2.89
3 . 1,4 - Butanediol
after 6 h 9.26 4.17 8.52 - 9.88 8.75 6.80
after 24 h 18.26 8.67 17.78 - 20.56 17.74 12.50
4. T oluene
after 6 h 17.02 20.58 40.23 - 63.51
after 24 h 41.72 45.34 83.03 - **) **) **)

5. Butyl a cetate
after 6 h 10.07 12.70 25.11 47.22 37.75 24.96
after 24 h 27.18 27.87 **) **) **) **)

*)*) determination by this method is not possible because the abrasion resistance is too low; **) damaged

Penczek, Ostrysz, Wieczorek, Boncza-Tomaszewski, Udeshi / 567

CONCLUSIONS BIOGRAPHIES
Optimizedformulations: Piotr Penczek
a) elastomers of Shore A hardness in the range of 20 -30
Piotr Penczek joined Industrial Chemistry Research Insti­
(Formulations IV and XI)
tute, Warsaw in 1958. He is Head of Department Polyes­
Castor oil-based polyester polyol, mole OH 1.00 ters and Epoxy Resins. His Research activities have in­
1,4-Butanediol, mole OH 1.00 cluded projects with Polish manufacturing companies in
TDI, mole NCO 2.00 - 2.10 the field of polyester polyols, polyurethanes, unsaturated
NCO/OH 1.00 - 1.05 polyester resins, polyestrimide varnishes and epoxy resins.
Dr. Penczek is co-author of two books. He has contributed
b) elastomers of Shore A hardness in the range of 40 -50 160 published papers and holds 200 Polish and interna­
(Formulation X)_________________________________ tional patents. He was active as UNIDO expert.
Castor oil-based polyester polyol, mole OH 1.00 Dr P. Penczek was elected member of the Academy of
1,4-Butanediol, mole OH 2.00 Engineering in Poland in 1993.
TDI, mole NCO 3.15
Ryszard Ostrysz
NCO/OH 1.05
Ryszard Ostrysz, M.Sc., Ch.Eng. joined Industrial Chemis­
c) elastomers of Shore A hardness in the range of 60 -70 try Research Institute, Warsaw in 1960. He is Head of
(Formulation XIII)_______________________________ Laboratory Polyester Resins. He has contributed 20 pub­
Castor oil-based polyester polyol, mole OH 1.00 lished papers and holds 75 patents. Ryszard Ostrysz devel­
Glycerol, mole OH 0.83 oped and implemented numerous manufacturing processes
1,4-Butanediol, mole OH 2.27 in the field of polyester polyols, polyurethanes and unsatu­
TDI, mole NCO 3.89 rated polyester resins. Chemical utilization of
NCO/OH 0.95 poly(ethylene terphthalate) [PET} scrap for the manufac­
ture of polyurethanes, polyester resins and phenolics be­
longs to his main achievements.
d) elastomers of Shore A hardness above 80 (e.g., for the
rice-mill roller application) Zbigniew Boncza -Tomaszewski
(Formulations XII and XIV)_______________________
Dr. Zbigniew Boncza-Tomaszewski was bom in Warsaw,
Castor oil-based polyester polyol, mole OH 1.00 1.00
Poland. He graduate at Warsaw University, Department of
Glycerol, mole OH 1.75 2.30 Chemistry in 1970. Ph.D. in 1977 („The Application of
1,4-Butanediol, mole OH 1.15 1.50
Oxidation Reaction in Degradation of Dienes”). 1979 -
TDI, mole NCO 4.00 4.80
1981 Post-doctoral Fellow at Universite Laval (Quebec,
NCO/OH 1.02 1.00 Canada); 1981 -1988 Research Associate at Department of
Chemistry, Warsaw University; 1988-1989 Senior Re­
The amount of the low molecular chain extender should search Associate at University of Illinois at Urbana Cham ­
be experimentally adjusted in dependence on the proper­ paign (USA); 1989-1991 Visiting Professor at Department
ties of the polyester polyol used. of Medicinal Chemistry Purdue University (USA). Since
joining ICRI (Industrial Chemistry Research Institute,
Warszawa) Department of Epoxies and Polyesters Resins
in 1993, he has been working as the head of Laboratory of
Epoxies and Polyesters Resins. His particular interest in­
clude formulations and curing techniques for coatings.

568 / Penczek, Ostrysz, Wieczorek, Boncza-Tomaszewski, Udeshi

Innovative Polymers for Optimized Release Performance for HR
Molded and Flexible Foam

RICHARD ROSSIO
Chem-Trend, Inc.
1445 West McPherson Park Drive
Howell, Michigan 48843

A bstract Biography

Chem -Trend has recently made significant R ichard Rossio

advances in the development of innovative polymers for
high resilience molded flexible foam both automotive and Rich Rossio is Senior Research Scientist for
non-automotive. Polyurethanes in the Marketing and Technology division
of Chem -Trend, Inc. He received his B.S. in Chemical
In general mold release is a significant factor in Engineering from Michigan Technical University in 1985
the ease of processing, skin formation and overall He is currently responsible for developing release agents
appearance of molded flexible foam. As the foam for the global HR molded flexible foam markets, as well
manufacturers are continually driven to reduce density as all related technical services. Prior to joining Chem-
while improving foam durability properties, the release Trend in 1995, Rich was a Group Leader in automotive
agent’s performance becomes more critical. These novel technical service and product development at BASF.
polymers have been uniquely formulated to offer a high
degree of processing flexibility while maintaining
outstanding foam aesthetics. These release agents are
designed to promote foam bondabilty as well as reduced
mold build-up.

All of these proprietary polymer blends of parting

agents are water based, environmentally friendly and non-
VOC. Chem -Trend will cover in the poster, the detailed
performance of various families of innovative polymers
that are particularly suitable for the automotive and
furniture industries.

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 569

New Styrene-AcryloNitrile Polymer Polyols for Cold Cure
Moulding Automotive Applications

JO S CENENS, ELS VAN EETVELDE, JEAN -CLAUDE VANDICHEL

TED SCHRODER a n d BENOIT FLAMMANG Shell Chemicals
Shell Chemicals Shell Centre
Monnet Centre International Laboratory London SE1 7NA
Avenue Jean Monnet 1
B-1348 Louvain-La-Neuve
Belgium

ABSTRACT BIOGRAPHIES

The trend in flexible moulded automotive seating foam Jos Cenens

towards the production of high quality but lower weight
foams has been with us for many years. To meet this Jos Cenens has worked for Shell Chemicals for the past 13
continuing need for extra customer comfort at reduced cost, years. He is currently department manager for the moulding
reactive polymer polyols containing solid particles have and CASE applications development group.
been developed for use in the various moulding
technologies. Polymer polyols are based on the Styrene- Els Van Eetvelde
AcryloNitrile (SAN) particle technology or polyols based
on polyurea or polyisocyanate polyaddition technology Els Van Eetvelde joined Shell Chemicals in 1998 and is
(PHD or PIPA) whereas Shell Chemicals has relied on its working on flexible automotive moulding seating
POSTech technology (a stable dispersion of polystyrene applications.
particles) for its polymer polyol grade range. Competitive
polymer polyols based on SAN technology can contain Ted Schroder
higher solids contents compared to those achievable via the
POSTech technology whilst maintaining acceptable Ted Schroder has worked for Shell Chemicals for the past
viscosities. Shell Chemicals has now made significant 11 years. In 1999 he joined the moulding group where he
advances in the design of its own SAN polymer polyols works on flexible automotive seating applications.
using a proprietary particle stabilising technology.
Benoit Flammang
Following the technical approval of polymer polyol
samples produced on laboratory scale, Shell Chemicals has Benoit Flammang has worked 9 years for Shell’s
recently evaluated the potential of our new SAN grade on a polystyrene business. In June 2000 he joined the moulding
semi-commercial production scale. The so-produced applications development group.
polymer polyol was evaluated on a 6 component high
pressure moulding machine in low-density TDI-based Jean-Claude Vandichel
automotive seating applications. The new Shell SAN
polymer polyol was compared to both competitive SAN Jean-Claude Vandichel has worked for Shell Chemicals for
polyols and Shell's POSTech polyol. The results obtained the past 13 years. He is currently Global Marketing
show that this new product offers the desired degree of Development Manager for Shell’s Urethanes business.
foam design flexibility within a wide load-bearing and
density range. The possibilities to optimise durability and
comfort characteristics with the SAN polymer polyol
significantly adds to Shell Chemical's product offering in
the automotive seating market.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 571

Spray Polyurethane Elastomer for Use In Automotive Dash
Insulator Applications

ANTHONY J . MESSINA a n d GIRMA GEBRESELASSIE

Collins and Aikman Corporation
Global Product Development Division

ABSTRACT from Wayne State University in Detroit, MI. He is a

member of the Society of Plastics Engineers and the
The automotive industry is always looking for ways to Society of Automotive Engineers.
reduce weight. However, in the past, it has been difficult
to optimize mass for NVH performance. Collins & G irm a Gebreselassie
Aikman’s new technology will allow for the production of Girma M. Gebreselassie is
lighter weight parts that will not sacrifice acoustic Manger of Advance materials
performance by putting the mass where it is needed. & process development for
Collins & Aikman Corporation
Currently, there are 3 major ways to produce Dash Global product development
insulator barriers: 1) PVC sheet, 2) EVA sheet, and 3) division in Plymouth ML He
Injection molded TPO. The limitation with the PVC and holds BS in Plastics & Telecom
EVA sheeted materials is that they do not have the ability munication engineering. His
of variable thickness, therefore material and cost is wasted industrial career includes principal project leader with Lear
on the areas where sound transmission is a minimal Corporation, program manger with MASCO Tech.,
problem. Injection molded materials, like TPOs, have research & development engineer with US Plastics, project
varying thickness, however they have a significant tooling & quality engineer with engineering plastic products. He
cost. has worked in the automotive industries for over 16 years.
He has presented a number of technical papers to various
Using a polyurethane elastomer that is robotically sprayed trade shows. He received 1st place award from Interior
will give a variable thickness product with low cost Automotive Exp. 1998 and 1st place award from
tooling. Once Sound Intensity Mapping (SIM)of the Polyurethane Exp. 1998. He holds a number o f US &
vehicle’s dash area is complete, the part can be engineered International Patents in interior trim applications, advance
to place more material in the acoustical ‘hot spot’ areas and materials & process. He is a senior member of Society of
minimize material in the areas that will be cut out for pass Automotive Engineers / Society of Plastics Engineers /
throughs, etc. This will produce a part that reduces weight Society of Mechanical Engineers.
by using the mass efficiently without losing NVH
performance.

AUTHORS

Anthony J. Messina
Anthony J. Messina is currently
a Development Engineer in the
Advanced Material and Process
Group for Collins & Aikman’s
Global Product Development
Division. He has worked in
Plymouth, MI since 1997 where
his main focus has been polyurethane foams and
elastomers. Anthony has a B.S. in Chemical Engineering

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 573

Synthesis of PU High Resilience Flexible Foams with Dispersing
Liquid Prepared by Polyether Grafting Starch Copolymer

CHEN SU a n d XIA XUE-WEI

Nanjing University o f Chemical Technology
Dept, o f Chemical Engineering
Xin Mo Fan Ma Lu 5#
Nanjing, 210009
China

ABSTRACT EXPERIMENTAL

The effects of components on properties of Main Raw M aterials

polyurethane high resilience foams synthesized by
starch grafting polyether dispersing liquid were N-1618 polyether polyols were obtained from the
investigated. It is shown that this dispersing liquid has second chemical factory of Jinling petrochemical
improved greatly load-bearing capacity of polyurethane company.
flexible foams. The optimum formulation was given. hydroxyl number 33 d : 3mgKOH/g,
The major properties of the polyurethane high resilience primary hydroxyl content: over 68%,
foams reached the level of ASTM criteria for water content(%): max. 0.1.
polyurethane high resilience flexible foams. The dispersing liquid (ND) was prepared by polyether
polyols ( M n - 5000) grafting starch under coupling
INTRODUCTION
agents and catalysts (solid content 25 -35%, hydroxyl
number 29mgKOH/g, water content max. 0.15%,
The problem o f improving comport and load-bearing viscosity at 25 °C, (3000~4000Mpa • s). Isocyanates such
capacities of polyurethane (PU) flexible foams was as TDI, MD1 or TDI/MDI (4:1) system was obtained
concerned by PU industry. Adding inorganic fillers or from Olin Company.
adding organic fillers is traditional methods. Firstly
directly adding inorganic fillers such as CaC03, B aS04, Foaming Procedure
can make their products low cost. However, properties
of these foam products are often of inferior quality. The raw materials including polyether polyols,
Their compound system is unstable too. It caused much catalyst I, II, water and crosslinking agents are poured
difficulty to technical operation for forming. Secondly, into a tank and mixed completely. The temperature of
directly adding organic fillers such as polymer polyols the tank was kept at 20-25 °C. Then the isocyanates (TDI
or polyurea polyols dispersing liquids can make their or TDI/MDI) were poured into the tank and mixed
foam products high cost. However, Properties of those quickly. Mixing time is from 3 to 15 seconds,
products are of excellent quality. Polymer polyols were depending on the desired cell size, and the stirring
often prepared by polyether polyols grafting styrene and system. As a result of the mixing process a certain
acrylonitrile. Polyurea polyols were often prepared by amount of air is stirred in, which affects the finer cell
polyether polyols reacting with isocyanate and binary structure. Stirring which is too intensive may results in
amine.
extremely fine, irregular cells, and splits.
In this paper, a new type of dispersing liquid made by
After this last stirring operation, the mixing equipment
Polyether polyols grafting starch copolymer was used
is raised up by means of the lifting device, swung away
for high resilience polyurethane foams(HR PURF).
from the mould and cleaned in a solvent bath by
These dispersing liquids o f good stability and lower cost
activating the stirrer for a short time.
comparing with organic filler. Meanwhile, properties of
About 10 minutes after the foam has reached its final
foam products are the same as those foam products
height, de-molding can take place. The reactions which
prepared by organic fillers respectively.[l]
occur during foam formation continue to generate heat
for some hours after expansion is completed, and this
heat is essential for the development of optimum foam

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 575

 Tablel. Effects o f water content on foam properties
Water content Foam properties
phr Density Kg/m3 Resilence % Tensile strength Kpa Elongation %
1.5 56.16 64 102.1 168
2.0 42.35 61 71.0 152
2.5 39.08 60 61.5 145

Table2. Effects o f crosslinking agent on foam properties

Crosslinking agent Foam properties
Phr Density Kg/m3 Elongation % Tensilie strength Kpa Resilience %
0.3 56.16 168 102.1 64
0.5 56.55 169 78.0 58
0.7 56.70 161 53.2 59
0 55.20 172 109.2 63

properties. For this reason, no post-treatment should be

undertaken until about 24 hours curing time has elapsed.

RESULT AND DISCUSSION

Effect on Factors of Foam Properties

I. WATER

Table 1 shows that the density, elongation and tensile

strength o f foams decrease with water content increasing.
It is the result that rigid urea chain content in foams with
water increasing.

II. CROSSLINKING AGENT

Crosslinking agents were generally not only used to

increase the crosslinking density of foams, but also used Figurel. Relation o f ND dispersing liquid withfoam
to increase hardness of foams and loadbearing capacity. density
Table 2 shows that tensile strength and elongation of
foams decrease with crosslinking agents (01) content
increasing. So crosslinking agent (01) must be
controlled under 0~3phr. Crosslinking agents are usually
diethanolamine, triethanolamine, 1,4-butanediol and
dimethylethanolamine. In our formula of HR PU foams,
diethanolamine is used as the crosslinking agent.

III. POLYETHER POLYOLS REACTIVE GRAFTING

STARCH DISPERSING LIQUID (ND DISPERSING
LIQUID)

Effect o f ND dispersing liquid content on foam

density (see Fig.1)

Figure 1 shows that the density of foams changes a

little with the ND dispersing liquid content increasing.
The elongation increases with the content of ND
Figure 2, Relation ofN D dispersing liquid content
dispersing liquid increasing according to the figure 3.
with foam resilience
The tensile strength increases when the ND content
increases until reaching 50%, decreases when the ND
content increases from 50% to 80%. Effect of ND dispersing liquid content on tensile
strength and elongation of foams (see Fig 3)
Effects of ND dispersing liquid on foam (see F ig.2)

576 / Su, Xue-Wei

 Table3. Formulation JS-A ofH RPU foams
Raw materials Weight content
N -1618 polyether polyol 60
ND dispersing liquid 40
Water 1.5 -2.3
Crosslinking agent 0.2 - 0.3
L-5307 stabilizer 0.8-1.0
Cat-1 0.09 -0.1
Cat-2 0.4 - 0.5
Cat-3 0.015 -0.02
TDI/MDI index 100 -110

Table 4. HR PU flexible foam properties and comparison

Properties JS-A ASTM3770-7911
Density Kg/m3 56 40
Elongation % 163 60
Tensile strength Kpa 168 70
Resilience % 63 100
Sag factors 65%/25% 2.96 >2.4

B IO G R A PH Y

C H EN SU

He obtained his bachelor degree in

Chemical Engineering from
Nanjing Forestry University in
1987. He then spent 4 years in
Jinling Petrochemical Plastics
Research Center on study o f
polyurethane foam plastics,
elastomer and synthesis of
polyether polyols. In 1994 , he
obtained his master degree in polymer material from
Nanjing University o f Chemical Technology. Then he
worked as associate professor and director o f polymer
Figure 3. Relation ofN D dispersing liquid content engineering institute in Nanjing University o f
with elongation and tensile strength Chemical Technology. His research interests are
IV. FOAM FORMULATION AND FOAM synthesis and modification of polyurethane,
PROPERTIES (see table3 and table4) ring-opening polymerization and emulsion
polymerization. In 2001, he obtained his Ph.D. in
The optimum formulation (see table3) was given industry catalysis from Nanjing University of Chemical
through studying on the factors of preparing HR PU Technology. Currently, he is the team leader o f DMC
foams. The foams’ properties (see table4) were also catalysts for preparing polyether and nanometer
given. polyurethane emulsion.

C O N C L U SIO N X IA X U E W E I

The dispersing liquid prepared by polyether polyols He obtained his bachelor degree
grafting starch was used to prepared high resilience cold in Nanjing University of Chemical
cure PU flexible foams. It can increase greatly load Technology, majored in chemical
bearing and comport capacities of HR PU foams. The engineering. Now he is a postgraduate
properties o f the HR PU foams prepared by this ND in Nanjing University of Chemical
dispersing liquid can reach ASTM criteria. Especially, Technology. His main research subject
its foam is of excellent resilience. is ring-opening polymerization.

REFEREN CE

[1] Z.Wirpsza. 1993. Polyurethanes Chemistry,

Technology and Applications. PA: Ellis Horwood
Limited, pp. 226-246. Su, Xue-Wei / 577
Whiplash Countermeasures: SafeGuard™

ED KUCZYNSKI PAUL FARKAS a n d ROMEO STANCIU

Product Development Group Chemical Research and Development
Woodbridge Foam Corporation Woodbridge Foam Corporation
2400 Meijer Drive 8214 Kipling Ave.
Troy, Michigan 48084 Woodbridge, Ontario L4L 2A4
Canada
ABSTRACT
O The injury source for whiplash is the sudden
differential movement of the head, neck, and torso
The Visco-Elastic behavior of polyurethane polymers
(Figure 1).
provide for a delayed and impact rate dependent response
■ Extreme hyperextension of the neck leading to the
to an applied stress and makes them desirable in parts
straining or tearing of the muscles, tendons, or
designed for specific energy management (EM).
ligaments in the neck.
Polyurethane Visco-elastic (VE) foams have not found,
■ Damage to the nerves entering the spinal column can
till now, a significant success in automotive interior
occur with rapid changes in spinal column pressure
applications due to a notable loss of some essential
and as a result o f any shape change.
properties like softness and general "friendliness” towards
■ Flattening and extension of the spine that may cause
contact with human body at low temperature and
the soft intervetrebral disks of the elongated,
substantially reduced VE property at elevated
weakened neck to shear in series.
temperatures.
O The symptoms range from soreness lasting several hours
The Woodbridge Group® has developed a new visco ­ to headaches, pain, nausea, decreased range of motion ...
elastic polyurethane foam (SVE foam) that demonstrates to tingling in the arms and legs that can last several
some truly unique capabilities. As an important part of a months or more.
complete assembly, SVE foam, marketed under
IntelliSense™ helps meet the energy absorption
requirement in head restraint regulation (ECE 25), and may REAR IM PACT DUMMY (BIO RID)
help mitigate the impact forces and momentum associated
with whiplash injuries. This new material has VE The more “human -like” BioRJD is proposed to replace the
properties and consequent impact-rate sensitivity over a previous Hybrid III dummy used for impact tests.
greater temperature range, while providing good energy O It quantifies a measure of neck injury risk, the Neck
absorption and suitable recoverability as compared to Injury Criteria (NIC) [3].
traditional HR. ■ NIC indicates the head motion is delayed behind the
The poster is intended to complement the information torso motion.
provided in our, paper entitled: ’’Automotive, Moulded ■ NIC compares the speed and acceleration o f the head
Visco-Elastic Foam” P. Farkas at al., Polyurethanes Expo at the top o f the neck with the speed and acceleration
2001) and to underline the key aspects o f this advancement of the first thoracic vertebra at the bottom o f the neck.
for both the polyurethane and automotive industries.
This novel approach to visco-elastic foams will benefit
energy management automotive applications like head
restraint, seat occupant safety, sound or vibration
attenuation, etc.

BIO - MECHANICS

Whiplash injuries are the relatively minor neck injuries

that typically occur in rear-end collisions [1,7].

Figure 1. Rear-End Collision

Typical Head Movement

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 579

 ■ NIC value of 15 m2/s2 is the threshold below which
the chances of whiplash injury with prolonged
symptoms are very low.

COUNTERMEASURE DESIGN

1) Reduce occupant acceleration.

2) Minimizes forces and rebound.
3) Minimize relative movements between adjacent
vertebrae.
a) Seat back rest and head restraint should
geometrically support the curvature of the back
and neck.
b) The head restraint should be located closer to the
head during impact events.
c) No local hard or soft structures in the seat back
that would force the spine into localized bending Figure 2. BioRid Dummy [11]
HEAD RESTRAINT
CRASHWORTHINESS REGULATIONS • Dynamic Sled Test Compliance Option (U.S.
only) where the HICmax of 150 within a 15
The estimated cost of whiplash-related injuries in the millisecond window using a 50th percentile
United States is $5.2 billion per year. Hybrid III dummy at 15 mph.
O NHTSA issues NPRM for FMVSS 202 upgrade
January 4th, 2001 [2,7]. INTELISENSE™
• Harmonization with European ECE 25. ENERGY - DISSIPATING FOAM
• Requirement for Head Restraints in rear outboard
seats, with retraction option. O Visco-elastic foam padding may provide maximized
• Design mandate of maximum backset of 1.97” at energy dissipation and suitable recoverability
top heights of 29.5” - 31.5”. O A material is Visco-Elastic (V E) when the relation
• Minimum width for bench seat Head Restraints is between an applied stress and the resultant strain is
10”, and 6.7" for single seats. time dependent [4].
• Maximum gap (Halo) within the Head Restraint is O Visco elasticity is a characteristic o f long chain
2.36". structures which may behave simultaneously as elastic
• Maximum displacement of headform of 4” under solids and viscous liquids [9] (see Figure 4).
3,300 inch-pound moment. O Other common names for visco-elastic foam include:
• Energy-Absorption - The deceleration of the 15 memory, dead foam, shape-memory , low recovery,
mph Free Motion Head (FMH) must not exceed slow recovery, pressure relief and zero-resilience
80 g’s continuously for more than 3 milliseconds. foam [5].

Figure 3. Typical Head Restraint Setup

580 / Kuczynski, Farkas, Stanciu

 Table 1. Typical Properties - IntelliSense“ VEFoam
T y p ic a l P ro p e rtie s In te lliS e n s e ™ A B
Foam

Density [kg/m3] 65-70 67-73

Tg [deg.C] -25 -43
Ball Rebound [%] 23 21
ILD(50%) [N] 190-300 370-540
Tensile Strength [kPa] 160 155
Elongation [%] 150 128
Tear Strength [N/m] 260 290
Compression Set (50%) [%] 25 -30 88
Figure 4. Modulus vs Temperature Fogging 100 100
Flammability [mm/mln] 43 42
(pass) (pass)
O The relative energy-dissipating ability, measured as
the ratio between the Loss Modulus (E”) and the
Elastic Modulus (E5) [6,9,10] is the trigonometric O An approximate constant tan-delta, as function of
tangent value of the angle of two vectors, known as temperature, means that a product suitable for energy
“tan -8”, a characteristic o f the polymer backbone (see management applications has expected and known
Figure 5). behavior within die considered temperature range.
• E” is the loss modulus (viscous dissipation
energy) - a non-recoverable entity, being Typical High Resiliency (HR) Foam
dissipated as heat.
• E ’ is the storage modulus (stored elastic energy) O Tg is ~ -40°C, and T rf is +58°C.
a recoverable entity. O The tan-8 curve (The E”/E’ ratio) is approximately
• The vectors (strain and stress) are out of phase by parallel to the temperature axis only between ~ -20°C
an angle “delta”, where: tan-8= E”/E \ to ~ +45°C (see Figure 5 b).
O Markedly, there is a substantial softening by HR
foams at elevated temperature visible in the dynamic
impact test, while the IntelliSense™ foam has little
change from its behavior at ambient temperature (see
Figure 6 -no curve shown for HR foam @80°C as the
sample was bottoming-out).

Figure 5. IntelliSense ™(a) vs. HR Foam (b) Figure 6. IntelliSense m vs. HR Foam
DMTA Analysis Impact Behavior at Various Temperatures

Kuczynski, Farkas, Stanciu / 581

 Table 2. Effect of Energy-Absorbing IntelliSense on
Neck Flexion
M aterial F le x io n B e n d in g
M o m e n t (N m )

C o n v en tio n a l H ea d 291
R estrain t

S a feG u a rd ™ with 87
In telliS en se™ F oam

IntelliSense™ foams demonstrates the increased

hardness of the latter with increased impact velocity
(Figure 7a,b).
O Typical IntelliSense™ foam properties are presented in
Table 1.

SAFEGUARD™ HEAD RESTRAINT

Figure 7. Effect of Strain-Rate on
R ear Im pact Test - Effect of Energy Absorbing
Material Load-Resistance
IntelliSense on Neck Flexion
IntelliSense™ V -E Foam
Test Setup:
O Conventional V -E foam has typical a high Glass Honda Odyssey
Transition Temperature (Tg) of +20 to + 25° C. • Hybrid III head and neck in calibration set-up
O IntelliSense™ VE polyurethane foam extends the • Back set of 2”
effective operating range of the material from above • 15 mph impact speed
>+80°C, to <-20°C, while maintaining the relative Significant reduction of the Neck Flexion Bednding
energy-dissipation values at the designed levels Momentum was achieved (see Table 2).
(Figure 6).
• Tg is ~ -43°C, and T rf is +108°C to below 0°C., Validation to FMVSS 202 (NPRM) and ECE 25
• Reasonable range of predictable performance is Energy - Absorption C riteria
between - -20° C to - +100°C (see Figure 5 a).
By using the SafeGuard™ head restraint with
O All IntelliSense™ foams show a good corroboration of
their thermal behavior with the impact results under IntelliSense™ foam the maximum deceleration durring
dynamic testing done with temperature-conditioned the 15 mph FMH test is reduced to less than 40g. (see
samples (Figure 6). Figure 8 & 9)
All Samples met 80 -g’s Regulatory requirement
O The expected visco-elastic behavior of IntelliSense™
foam is evident in its impact rate sensitivity [8]. The • Conventional Head Restraints did not meet 64-g (max)
impact curves at different velocities for the HR and for OEM “Due-Care” requirements - 20% buffer.

0 S 10 1S 20 25 30 35 40 45 SO
Time (ms)
Figure 8. IntelliSense ™ vs. HR Foam Figure 9. IntelliSense ™ vs. HR Foam
Commercial Design #1. Commercial Design #2.
582 / Kuczynski, Farkas, Stanciu
CONCLUSIONS

SafeGuard™ Head R estraints with IntelliSense™

Visco -Elastic Foams offers:

• The Comfort and Feel of a conventional “Head Rest”

materials
• Energy-Dissipation - Optimized Energy Absorption
over extended range in a fully-recoverable foam
• Temperature Stability Consistent Performance within
a wide temperature range
• Strain-Rate Synergy - Increased energy absorption
with increase impact force
• Product Performance Reducing loads to help meet
Regulatory and Due-Care requirements

REFERENCES

1. Insurance Institute for Highway Safety (IIHS) Report

April 1997 (Updated version in April 2001).
2. Stoffer, H., Jan.8, 2001, “NHTSA seeks tougher head
restraint rules”, Automotive News,
3. Bostrom O. at all. 1999, “Neck Injuries in Car
Collisions - A review Covering a Possible Injury
Mechanism and the Development of a New Rear-
Impact Dummy”, WAD’99 Congress, Whiplash-
Associated Disorders Compendiu / Traffic and Auto
Engineering, February 7-11,
4. Birley, A. W. at al. 1991. Physics o f Plastics. New
York, NY: Oxford University Press, pp. 260-261.
5. Herrington, R. July 20, 2000. “Low Resiliency
Flexible Polyurethane Foams”, Tech Note /Report to
WFC, Dow Plastics/Vortex.
6. 4.Riande, E. at al. 2000. Polymer Viscoelasticity. New
York, Basel: Marcel Decker, Inc. pp. 238-240
7. NHTSA, Jan. 2001. 49 CFR Part 571 [Docket No.
NHTSA-2000-8570] “Notice of Proposed
Rulemaking”; MVSS; Head Restraint.
8. Davies, O. at al. 1999. “The rate dependence of
Confor Polyurethane Foams”, Cell. Pol. 18(2): 117-
136
9. Whitman, R.D. at al. 1963. “Urethane Foam Properties
related to The Fundamental Polymer Molecular
Structure”, Proceedings o f the 1963 SPI Polyurethane
Conference, Technomic, Lancaster, PA. 1963, p. 1
10. Rieger, J., 2001. ”The Glass Transition temperature Tg
o f polymers-Comparison of the values from different
thermal analysis (DTA, DSC) and dynamic
mechanical measurements (torsion pendulum).,
ELSEVIER, Polymer Testing, 20:199-204.
11. Biomechanical Research Testing, December 12, 2000
“PMRF Seat Evaluation Project Summary of Results
/ Report to WFC.

IntelliSense™ & Safeguard ™ are trademarks of

Woodbridge Foam Corporation

Kuczynski, Farkas, Stanciu / 583

 POSTER SESSION
FUNDAMENTALS/EQUIPMENT
Compatibility Study of Poly bd® Resin with Other Polyols and
Property Evaluation of Polyurethanes Prepared Therefrom

HERBERT S . CHAO
Sartomer Co., Inc.
502 Thomas Jones Way
Exton, Pennsylvania 19341

ABSTRACT

All the polyester or polyether polyols tested are

compatible with poly bd R45HTLO in low proportions
ranging from 10 to 50% of the total weight of the
resulting blend. Compatibility of those polyols with poly
bd resin is affected by polyol type, functionality and
molecular weight. Polypropylene glycols are in general
more compatible with poly bd resin than
polytetramethylene glycols and those polyols containing
aromatic groups. Among polypropylene glycols, it
appears that diols are more compatible than triols of the
same molecular weight. For the same type of polyols, the
solubility in poly bd resin increases with the molecular
weight of polyols.
Tensile and tear strength of the polyurethanes made
from the polyol blends and Isonate 143L (MDI) decrease
with increased amount of poly bd resin. However, poly bd
resin improves the water resistance of the final material.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 587

Dielectric Polarization Measurement with CMD Provides New
Insight into the Foaming Process

BERND H. W. HOFMANN D. JEFFREY G RO SS

Format Messtechnik GmbH Electrotech Industrial Inc.
D -76187 Karlsruhe Northfield, Ohio 44067
Germany

ABSTRACT

Standard laboratory methods for foam testing are mainly polyurethane matrix. The effects of catalysts, water, and
concerned with the measurement of mechanical and thermal additives including formulation changes can be seen most
parameters. The linear expansion of foam samples are effectively in the polarization curves (figs. 3,4).
determined by rise height tests. Pressure measurement has The negative slope of the polarization curve represents
been significantly improved by replacing normal test cups the curing behavior of the foam which can be correlated to
with the FPM system. The temperature is commonly rise height, shrinkage, pressure and other data provided by
measured by thermocouples inserted into the foam. the Foam Qualification System FOMAT.
The dielectric polarization is a new key parameter that
provides more insight into the electro-chemical processes Patent pending
during the transition from the liquid mixture of polyoles,
additives, and isocyanates to finished PU foam with the
desired properties. The dielectric polarization of liquids is
measured simply by inserting surface capacitors into the
sample. Though the foam components polyoles and BIOGRAPHIES
diisocyanates have a strong electric dipole moment, this
method cannot be applied to foaming materials because of Bernd H. W. Hofmann
their increasing volume and the formation of a solid sticky
bulk. Format Messtechnik GmbH, Karlsruhe, Germany Dr. Hofmann received his
developed a new approach in polarization measurement Ph. D. in solid state physics from
technique. Introducing the Curing Monitor Device (CMD) the Technical University of
by meeting both simple handling and high polarization Karlsruhe, Germany, in 1976. He
sensitivity, the invention combines polarization has worked in the field of
measurement with routine foam testing methods avoiding conventional and nuclear
any additional recovery needs for the testing equipment. physical engineering at the
The CMD comprises two comb shaped electrodes on a Karlsruhe Research Center, for
plane circuit board forming the polarization sensor. This ABB Reaktor and other
sensor is integrated into the pressure measurement device companies. Dr. Hofmann is
FPM, where it completely covers the base plate at the director of Format Messtechnik
bottom of the test volume (fig. 1). The rise pressure makes GmbH since 1994.
close contact between the foam sample and the CMD
sensor, thus ensuring unspoiled penetration of the electric
field. As the CMD sensor is covered completely by the D. Jeffrey Gross
liquid mixture, polarization data can be obtained from the
very beginning of the foaming reaction. Mr. Gross is the Sales Manger
Measurements made with flexible and rigid foams have with Electrotech Industrial, Inc.,
shown a significant correlation between the water content of the exclusive distributor of
the formulation and the resonance structure in the Format Messtechnik GmbH
polarization curve (fig. 2). This is caused by polyurea equipment in North America.
generation as an intermediate before it is bonded to the

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 589

 Figure 1. Cross section of the CMD sensor integrated into the foam pressure measurement device FPM

Figure 2. Comparison of the rise height and the dielectric polarization for two flexible foam samples with different
isocyanate content Sample B was made with 5/3 of the isocyanate amount of sample A.

590 / Hofmann, Gross

Figure 3. Dielectric polarization measured with CMD, and other physical data of a semi rigid foam obtained by FOMAT.
The polarization curve shows a significant resonance structure and curing.

Figure 4 . Dielectric polarization measured with CMD, and otherphysical data of a rigid foam obtained by FOMAT.
The polarization curve shows an early curing after about 200 seconds.

Hofmann, Gross / 591

1C Fluid Power for Fluid Measurement and Control

TONY MARSAGLIA
IC Fluid Power

ABSTRACT

IC Fluid Power was founded in 1989 as a North American higher accuracy, lower pressure drops and a much wider
distributor for companies; VSE, Beinlich, and DST. These turndown range than the conventional positive
companies have extensive experience with a focus on displacement flow meter.
special applications for metering, transfer and containment
of fluids such as those used in the polyurethane industry. With the broad range of pumping and metering products IC
The main goal of IC Fluid Power is to provide customers Fluid Power can provide the hardware and technical
throughout North America unique solutions while backing information to be sure all variables are accounted for in
each product with engineering and technical support. installations. Having experts available who understand the
dynamics of the fluid transfer, measurement and control is
The key products utilized in the polyurethane industry that an asset to the customer looking to improve a process,
IC Fluid Power carries are pumps, precision flow meters eliminate scrap, reduce downtime and contain chemicals.
and magnetic couplings. These products are frequently
used to provide the main function of fluid transfer, and IC Fluid Power invites all PUR Expo visitors to inquire
measurement. Added benefits are reduced down time, about how these products can show you today ways to
scrap elimination, improved products and a safer work improve your process and equipment.
environment for employees.

Beinlich Pump is a manufacturer of internal gear, external

gear and radial piston pumps for the hydraulic and
chemical industry. Special seals for the Polyurethane
industry are available such as the 3 seal system, face seal
for abrasives and the leak free magnetic coupling
assembly. Flow ranges from 0.25cc/rev up to 750 cc/rev
can be achieved with the Beinlich range.

DST is a manufacturer of magnetic couplings specializing

in custom couplings for almost any application. Magnetic
couplings have seen rapid utilization in the chemical
market to eliminate the need for shaft seals. This non-
contact coupling provides excellent power transfer and a
maintenance free, leak free, shaft seal configuration. It is
possible to provide retrofit packages for standard pumps to
eliminate shaft seal problems and create a completely
contained pumping system.

VSE was founded in 1989 as an affiliated company of the

well-established manufacturer of hydraulic components,
HBE. The flow meter series VS and VHM were developed
by experienced engineers and are manufactured in
compliance with the highest quality standards (DIN/ISO
9001). The patent gear design of the VS series allows for

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 593

Mixing Equipment for the In-Line Addition of Scrap Regrind and
Other Fillers into PUR

DARIO RAMAZZOTTI
Edge Sweet Company

ABSTRACT

The potential of recycling scrap has generated more

interest recently and the need for a continuous in-line
blending system has become apparent. The traditional
method of adding fillers into the polyol stream would be to
pre-blend the filler into a polyol holding tank and agitate
until blended. This system has many drawbacks and does
not give the best results due to saturation of a porous filler
with a polyol resulting in a polyol rich filler.

The system offered by Edge-Sweets eliminates this

problem by utilizing a continuous feed system which
meters the filler into the polyol at the time it is to be used.
Pre-blending and saturation of the filler with the polyol is
eliminated. Due to the inherent design characterists of the
equipment, the scrap ratio can be altered on-the-fly and in
fact, blends of scrap can be implemented. The system was
originally designed to incorporate powdered flame
retardants, however it has been adapted to this new
requirement and offers the slabstock producer and the
molder a system which can be used as an in-line blender or
as a final mixer, depending upon the process involved. The
system can take advantage of finely ground powder as well
as pelletized material in varying ratios up to 50% in some
cases.

The poster session includes photographs of the equipment

and samples of the product produced.

AUTHOR

Dario Ramazzotti

Dario Ramazzoti is currently Vice President, Sales &

Engineering at Edge-Sweets Company in Grand Rapids,
Michigan.

POLYURETHANES EXPO 2001 , SEPTEMBER 30-OCTOBER 3, 2001 595

 Structure and Properties of Polyurethanes Based on
Model Triglycerides

ZORAN S . PETROVIC a n d ALISA ZLATANIC

Pittsburg State University
Kansas Polymer Research Center
1501 S. Joplin
Pittsburg, Kansas 66762

effect of structure on properties of polymers from vegetable

ABSTRACT oils [1-6]. However, the structure of natural oils is highly
variable even in the same sample. Thus selecting model oils
Polyols based on vegetable oils are increasingly with just one type of fatty acids would help resolve some of
becoming viable alternatives to the petrochemical polyols. the problems. In this study we used two synthetic oils -
However, their structure varies from molecule to molecule triolein and trilinolein (triglycerides of oleic and linoleic
introducing an additional factor, which might affect the acids). Oleic acid has 18 carbon atoms and a cis - double
properties. In order to examine the effect of structure on bond between the 9th and 10th carbon, while linoleic has an
properties and avoid this cause of heterogeneity we have additional double bond between the 12th and 13th C atoms. If
used model oils having three (triolein) and six (trilinolein) one OH is introduced per double bond in the oil the triolein
double bonds as well as linseed oil (which contains more will give a triol and trilinolein hexol. Trilinolenin with three
than six double bonds) and converted them to polyols by double bond per (linolenic) fatty acid is difficult to prepare
inserting an OH group per double bond. While this and we have used linseed oil instead, which contains a high
conversion was almost quantitative with triolein, the percentage of linolenic acid. Linseed oil typically has 22%
hydroxylation of trilinolein and linseed oil was not oleic, 16% linoleic and 52% linolenic acid. Crosslinking
complete, giving lower functionality than predicted. The density will be in the trilinolein based polyurethanes twice as
structure of the polyols was characterized by spectroscopic, much as that in the triolein ones as shown in Fig. 1. The
physical and chemical methods. They were reacted with pendant chain is not participating in bearing the stress and
MDI to give polyurethanes with different crosslinking may act as a plasticizer. The pendant chain in the linoleic acid
densities, which were determined from equilibrium is shorter than that in the oleic acid. Thus we will compare
swelling in toluene. While trifunctional oil (triolein)-based properties of two polyurethanes with two different
polyol gave a rubbery material at room temperature, crosslinking densities and different pending chain lengths.
trilinolein and linseed oil based polyurethanes had glass
transition temperatures at 60 °C and 76 °C, respectively.
Increasing crosslinking density caused an increase in EXPERIMENTAL
modulus and tensile strength and decrease in elongation at
break. M aterials

Triolein and trilinolein (99%+) were purchased from NU-

INTRODUCTION CHEK PREP, inc. Epoxidized linseed oil Drapex 10.4 was
obtained from CK Witco Corp. Toluene (HPLC grade) and
Vegetable oils are triglycerides of fatty acids. They have a methanol (99 %) were purchased from Fisher Scientific. Ion
number of excellent properties, which could be utilized in exchange resin Amberlite IR-120 was supplied by Supelco.
producing valuable polymeric materials such as Ion exchange resin Lewatite MP-64 was purchased from
polyurethanes. Apart from being a renewable resource, Bayer. 4,4’-diphenylmethane diisocyanate (MDI), Rubinate
vegetable oils are highly hydrophobic, imparting high 44, obtained from ICI, was used after distillation under
moisture resistance, good thermal and hydrolytic stability. It reduced pressure.
is of great scientific and practical interest to examine the

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 597

Methods
Table 1. Properties of polyols
The polyols were synthesized from epoxidized oils by
oxirane ring opening with boiling methanol. The final P olyol OH# Eq. Functio Mn D en sity, V isco sity ,
products were analyzed by FTIR, GPC, steady shear MW nality g /c m 3 P a .s
(2 7 ° C)
viscosity measurements and OH number determination.
Triolein 157 357 2 .8 5 1123 0 .9 7 1 .7
Polyurethanes were synthesized using 2% mole excess of
Trilinolein 200 281 3 .7 7 1394 0 .9 7 23
isocyanate groups (Index 1.02). The polyols and MDI were L in seed 221 254 4 .2 6 1204 0 .9 7 31
preheated separately, then mixed and poured into molds.
Curing was carried out at 100 °C overnight. The crosslinking much higher extent in the cases of trilinolein- and linseed
density of the polyurethane networks was estimated from oil-based polyol (conversions 69% and 79 % respectively).
swelling measurements. Physical and mechanical The side reactions are mainly due to polymerization, i.e.
characterization of polyurethanes was carried out by formation of dimers and trimers. The degree of
differential scanning calorimetry (DSC), thermogravimetry intermolecular reactions can be estimated from GPC
(TGA), dynamic mechanical analysis (DMA) and chromatographs, Fig 2. The presence of higher molecular
mechanical methods. weight species significantly increases viscosity of the
product, in spite of the lower than expected OH content.
RESULTS AND DISCUSSION The density of the polyols is closely related to their
structure.
Characterization of polyols The number average molecular weight of polyols was
determined by vapor pressure osmometry (VPO), using
Three polyols based on model triglycerides were obtained Vapor Pressure Osmometer, Osmomat 070, UIC, Inc.
by ring opening reaction of the oxiranes in the epoxidized Measurements were carried out at 60 °C in toluene. Benzil
oils using methanol and acid catalyst. The properties of the was used as the calibration standard. The values of the
polyols are given in Table 1 (OH#, equivalent molecular number average molecular weights were 1123 g/mol for
weight, functionality, Mn, density and viscosity at 27 °C). triolein-, 1394 g/mol for trilinolein- and 1204 g/mol for
The conversion of the triolein to a polyol was carried out linseed oil-based polyol (Table 1). These values are
with better success than the conversion of the other two oils, considerably higher than theoretical Mn calculated from the
resulting in a much lower functionality than predicted of the
structure of polyols, experimental functionality and
latter. Nevertheless the three polyols show different assuming no polymerization. Intermolecular reactions,
functionalities, which resulted in different crosslinking however, occurred in our systems. The content of higher
densities of polyurethanes based on them. Also the viscosity molecular weight products was estimated from GPC
increased with increasing OH number from 1.7 Pa.s for the chromatograms of polyols. The content of dimmers and
triolein based polyol to 23 for trilinolein and 31 for the trimers were 13, 45 and 26 % for triolein-, trilinolein- and
linseed oil based polyol. It is highly desirable in the linseed oil based polyol, respectively. Based on that, the
technology of polyurethanes that polyols used have high OH number average molecular weights of polyols were
content and low viscosity. The best conversion of functional calculated to be: 1092 g/mol for triolein-, 1372 g/mol for
groups and the lowest viscosity are given in the case of trilinolein- and 1271 g/mol for linseed oil- based polyol.
triolein based polyol (96 %). The loss of functional groups These values are in reasonable agreement with VPO results.
during hydroxylation due to the side reactions occurred to a

Oleic acid

Figure 1. Reactive sites in oleic and linoleic acids

598 / Petrovic, Zlatanic

 Figure 3. Dependence of storage moduli of the three poly­
urethanes on temperature
Figure 2. GPC chromatograms of the oils-based polyols

Physical properties of the polyols are related to the structural

order and the physical state. The differential scanning
calorimetry thermograms of the model triglycerides-based
polyols exhibited several melting peaks; the values are listed
in Table 2. Multiple peaks in polyols are ascribed to
different crystalline polymorphs that are also present in
starting oils.

Characterization of polyurethanes

Measurements of the equilibrium swelling degree of model

polyurethane networks in toluene allows the determination
of their crosslinking densities. Triolein based polyurethane
increased the weight in toluene (Gs/G0) of 95%, trilinolein-
based PU had increase of 51.9% and the linseed oil PU had
43.5% weight increase. The experimentally determined Figure 4. Dependence of the loss moduli of the three poly­
functionalities of polyols were 2.85 for triolein-, 3.77 for urethanes on temperature
trilinolein- and 4.26 for linseed oil-based polyol. The
molecular weight of network chains in triolein-based The change of storage (E’) and loss modulus (E”) with
polyurethanes was calculated to be 1085 but it was lower for temperature of the three polyurethanes is presented in
the other two. Such molecular weight produces Figures 3 and 4.
polyurethanes with relatively low Tg (rubbery at room Tensile and flexural properties as well as notched Izod
temperature). The glass transition of the model triglycerides- impact strength of polyurethane networks are shown in
based networks is affected by their crosslinking densities. Table 3. While trilinolein- and linseed oil-based
Glass transition temperatures measured by DSC were 25 °C, polyurethanes exhibited strengths of 49 and 57 MPa,
58 °C and 76 °C for the triolein, trilinolein- and linseed oil respectively, the tensile strength of the triolein-based sample
based polyurethanes. DMTA at 5 °C/min and 10 Hz gave was only 7.9 Mpa, but it displayed a high elongation at
somewhat higher values of 42, 74 and 85 °C, respectively. break of 135%. The other two samples have very low
elongation at break characteristic of a brittle fracture. The
Table 2. Melting peak temperatures for different polyols main reason for the lower mechanical strength of the
triolein-based polyurethane was its low glass transition
P e a k # /P o ly o l> Triolein Trilinolein L in se e d oil temperature, which is close to the temperature at which
1 -4 5 -2 7 -2 2 mechanical tests were carried out. Practically, triolein-based
2 15 11 0 polyurethane was tested in its leathery region, while other
3 41 5
two samples were tested in the glassy region. The rubbery
4 17
nature of the triolein-based polyurethane explains more than

Petrovic, Zlatanic / 599

 REFERENCES
Table 3. Mechanical properties of oil-based polyurethanes
1. Pride, E. H., Gast, L. E., Frankel, E. N., Carlson, K.D.,
S a m p le T e n sile E lo n g a ­ Y o u n g Flexural Izod Polymer-Plastics Technology and Eng. 1976, 7, 1-26.
stren gth tion, M odulus, m o d u lu s, Im pact,
2. Petrovic, Z., Guo A., Javni I., Zhang W., Proc.
M Pa % M Pa M Pa J/m
Triolein 7 .9 136 - - 446
Polyurethanes EXPO 2000, API, Boston, MA, 2000.
PU 3. Guo, A., Cho, Y., Petrovic, Z.S., J. Polym. Sci., Part
Trilinolein 49 5 1289 4971 17 A: Polym.Chem. 2000, 38, 3900-3910.
PU 4. Petrovic, Z. S., Guo, A., Zhang, W., J. Polym. Sci.,
L in seed 57 6 1324 591 1 14 Part A: Polym.Chem. 2000, 38, 4062-4069.
oil PU 5. Javni, I., Petrovic Z. S., Guo A., Fuller R., J. Appl.
Polym. Sci. 2000, 77, 1723-1734.
an order of magnitude higher value of the impact strength 6. Javni, I., Petrovic Z. S., Guo A., Fuller R., SPE
than in the other two. Technical Papers (ANTEC ’99) 1999, Vol 45, 3801.
3805.
Thermogravimetric measurements were performed using a
TGA 2050 module of TA Instruments thermal analyst
system with controller 3100. Thermal degradation of the BIOGRAPHIES
polyurethane samples was followed in air and nitrogen, at
the heating rate of 10 °C/min. Initial weight loss is similar in Zoran S. Petrovic, Ph.D.
all polymers, suggesting that degradation starts with the
dissociation of urethane bond and carbon dioxide formation. Zoran Petrovic received his B.
Considering the initial stages of degradation, there were no Eng. at the Faculty of Technolo­
significant differences in the thermal stability of the samples gy, Belgrade University, Yugo­
in air and nitrogen. At higher conversions polyurethanes slavia, and Ph.D. at the Strath­
exhibited higher stability in air than in nitrogen, which could clyde University, Glasgow. He
be explained by crosslinking rather than depolymerization worked in electrical industry for
that occurs in the presence of an inert gas. 12 years on problems of appli­
cation of polymers in electrical
CONCLUSIONS insulation. He started his acade-
mic career at the Faculty of Tech­
Three polyols based on model oils were obtained by ring nology, University of Novi Sad, Yugoslavia, where he
opening reaction of the oxiranes in the epoxidized oils. taught Physical Chemistry of Polymers. He spent three
Corresponding polyurethane networks were prepared from years at the Polymer Science and Engineering Department,
methoxylated epoxidized oils and MDI. University of Massachusetts, at Amherst, working on
The best conversion of functional groups (96 %) and the different research projects. He is currently Director of the
lowest viscosity are given in the case of triolein based Kansas Polymer Research Center (KPRC) and Professor of
polyol. The loss of functional groups due to the Polymer Technology at Pittsburg State University,
intermolecular reactions occurred to a much higher extent Pittsburg, Kansas. He has published over 100 papers, 3
during preparation of the trilinolein- and linseed oil-based books, and has seven US patents. His special interests are
polyols (conversions 69% and 79%, respectively). polyurethanes, bio-based polymers, networks and gels,
The glass transition temperatures of polyurethanes (DSC) blends and polymer characterization.
increased with their crosslinking densities from 25°C for
triolein- to 58°C for trilinolein- and 76°C for linseed oil-
based polyurethane. Increasing crosslinking density caused Alisa Zlatanic, M.Sc.
an increase in modulus and tensile strength and decrease in
elongation at break. Thermal stability of all three samples, Alisa Zlatanic received her
both in nitrogen and air, were comparable as judged by the B.Eng. and M.Sc. degree from
initial weigh loss. Faculty of Technology and Me­
tallurgy, Belgrade University,
Yugoslavia. She is currently Re­
search Associate at Kansas Poly­
Acknowledgment: This material is based upon work mer Research Center (KPRC),
supported by the Cooperative State Research, Education Pittsburg, Kansas, working on
and Extension Service, U.S. Department of Agriculture, thermoresponsive gels and poly ­
under Agreement No. 99-35504-7873. urethanes based on natural oils.

600 / Petrovic, Zlatanic

Three-Way Thermal Conductivity Instrument Comparison

BRAD CANNEY, CRAIG DIXON a n d NANCY MATHIS

Mathis Instruments, Ltd.
8 Garland Court
Fredericton N.B.
Canada

ABSTRACT EXPERIMENTAL
Three thermal conductivity techniques have been Instruments
compared on polyurethane foam: guarded hot plate (GHP),
transient plane source (TPS) and modified hot wire Three instruments were used to measure thermal
(MHW). Trends and technique differences will be conductivity: Guarded Hot Plate (GHP), Hot Disk (HD)
discussed in the areas of contact resistance, homogeneity and TC Probe (TCP). The GHP method is steady state and
influence, variation through the thickness and mounting follows ASTM D518 [1]. The GHP results were
pressure. Strengths and weaknesses of the techniques will conducted independently and the results were received by
be discussed in the areas of test time, sample the authors. The other two instruments systems (TCP and
preparation/size, ASTM acceptance and calibration. HD) were interfacial and transient in operation rather than
the traditional steady state technique in GHP.
INTRODUCTION Both of the transient techniques are based on a principle
Thermal conductivity is a critical property in the that involves supplying a known amount o f energy in the
insulation industry. It measures the ability of the material form of electrical current to an initially isothermal sample
to transmit heat based on the Fourier equation Q=kAAT/L via a contact sensor while monitoring the resulting
where: Q is the heat in W, k is the thermal conductivity of temperature rise by using the sensor also as a resistance
the material in W/mK, A is the cross sectional area thermometer. The sensors act as both a heat source and a
(m2)through which Q flows, L is the distance (m) through dynamic temperature sensor for measuring the thermal
which Q flows and AT is the temperature gradient (K). transport properties of a material. The sensors measure the
Insulation, by definition, are materials with low thermal temperature rise representing blocked heat flow from the
conductivity. These include not only polyurethane, but instrument sensor into the material tested. The lower the
polystyrenes, fumed silica, fiberglass, glass wool, cork and conductivity, the more heat is trapped and the higher the
a variety of other materials. In high temperature interfacial temperature rise.
applications, ceramic is the insulator of choice, although The heat from the element penetrates into the sample, but
the thermal conductivity at 50 mW/mK would not be never reaches the far side. The length of the test determines
considered low by those in the appliance or construction the penetration depth. Typically for polyurethane a
markets. On the lower end of the insulation spectrum, penetration depth of 2 mm would be expected during a 10
vacuum insulation panels can achieve conductivities as low second test. Results from testing correspond only to the
as 4 mW/mK. Over this range, it is key to be able to volume of material penetrated by the heat wave; therefore
characterize the thermal conductivity as a necessary design it is basically a surface measurement
property. The goal of this paper will be to compare
existing and innovative methods to measure thermal
conductivity.

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 601

Figure 1 (above): The transient techniques only penetrat
a few mm so the sensors must be in contact with the foam.

Figure 2 (right): The HD sample holder (right) stack

two samples with the sensor between. The TC Prob
requires only one sample - shown with a surface weigh.

Materials Hot Disk, the technique incorporates a heating element that

is supported on a backing, which provides a rectangular one­
The sample set included 4 cardboard sandwiched dimensional heat flow. The heating element relies on
polyurethane squares. These were received in 300mm2 (1 interfacial contact with a solid flat sample having
ft2) by 50 mm (2in) formats that were sized for the GHP minimum dimensions of 5 mm by 25 mm.
testing that had occurred. The cardboard was 0.475 m. The The method requires a calibration with characterised
foam samples were cut into smaller sample sizes that were standard materials. The GHP thermal conductivity values
appropriate for the sample holder used in these for three of the four samples were used as reference
experiments and to facilitate efficient testing of their materials. This is a common technique for an end user to
exposed polyurethane faces (see Figure 1). When create calibration standards for a transient technique with
performing measurements, the Hot Disk sensor was fitted an established ASTM method.
between two sample halves, with each piece having a plane The timing options used for instrument calibration in
surface facing the sensor. The TC Probe™ utilized one of evaluating the foam samples were a 25 second test duration
this sample pair, also testing on the exposed PU surface and a 4 second test start time. The calibration plot of sensor
(see Figure 2). In contrast, the GHP measurements were response for the TC Probe™ using three foams as thermal
conducted with the entire sample with, the outer cardboard conductivity standards is illustrated in Figure 3. Upon
in contact with the hot and cold plates and therefor their completion of the calibration, the TC Probe™ was used to
contribution was incorporated into the results. measure the remaining unknown sample and the three
calibration samples.
Hot Disk

This instrument measures both thermal conductivity and

diffusivity, and from these two, volumetric specific heat
can be calculated. The standard software module was used
which is appropriate for materials with thickness typically
greater than 2 mm and thermal conductivity between 0.005
to 500 W/mK. Data points 10 to 200 were analyzed to
report the final thermal property results. Optimized testing
parameters included an 80-second test duration, 0.02 Watt
output of power to the sensor; a 9.734mm radius sensor
was used in these experiments. The “bridge” option was
used during the testing for optimum sensitivity.

TC Probe
Figure 3: TC Probe™ was calibrated with three samples
The instrument is based on a transient modified hot wire of polyurethane characterized by GHP. The error bars
(MHW) heat reflectance technique[2]. In contrast to the represent RSD for 5 tests in one location on the sample.

602 / Canney, Dixon, Mathis

Table 1: Thermal Conductivity Results (W/mK)

Sam ple GHP TCP % diff HD Bridge % diff

A 0 .0 2 1 2 0 .0212 0.08% 0 .0 3 0 4 43.19%
B 0 .0 2 1 3 0 .0 213 0.01% 0.0301 40.95%
C 0 .0 2 1 8 0 .0 217 - 0.26% 0 .0 3 0 6 40.50%
D 0 .0 2 2 8 0 .0 223 -2.04% 0 .0 3 1 6 38.69%

RESULTS AND DISCUSSION contact resistance. HD measurements are not influenced

The results of the thermal conductivity testing with the by this effect as long as the contact pressure is maintained
three instruments is presented in Table 1. The percentage constant during the test time (80s) [4]. Location o f the heat
difference in the columns is taken in reference to the GHP wave in the sample moves out from the sensor during the
result. The results are plotted in Figure 4. test. The heat is therefore resident in the boundary during
It is clear from the table that a significant bias exists the initial test period (<0.5s). During the data
between the HD and the GHP. The graph shows that manipulation, the first data points are deleted and the slope
while, although a bias in the absolute value exists, the trend of the data is examined. Interfacial contact resistance is
is the same and the results correlate well. Both the TCP reflected in the data by an early time jump in the
and Hot Disk have stated precision of 2% and accuracy of temperature difference, but that does not effect the later
5% so the sample and technique variations will be slope. A firm pressure is placed on the sample during the
examined. measurement to mitigate this effect as well so that only a
The HD measurements are taken orthogonal to the GHP small amount of data from the start must be eliminated. By
measurements. It has been demonstrated previously [3] removing the contact resistance, the measured conductivity
that thermal conductivity is anisotropic based on the would be higher.
direction of foam rise. To investigate this, the cardboard A third contribution comes from the variation in the
was removed from the faces of the sample and the HD tests sample. If the core has a lower conductivity, the GHP,
were repeated with the sensor in contact with that face. In which measures across the entire profile, would average in
this way the heat was flowing in the same direction for the those contributions, where the HD which only measures at
two measurements. The cardboard to cardboard direction the surface would not. Again, this would generate a higher
was referred to as facial and the results in table 1 were value for the HD measurement.
referred to as edge. The facial measurements were 10% The TCP measurements calibrate out all of these affects
higher lower than the edge testing. This would be by relating back to characterized values obtained by GHP.
consistent with the assumption that the direction of foam The calibration, although an extra step, removes this bias.
rise would be parallel with the cardboard face. The advantages of the transient methods are two fold: test
time and sample configuration.
The second consideration regarding the bias is the

C onductivity (W /m K ) Method Com parison

Figure 4: The results from the

steady state (GHP) an
transient methods (TCP an
HD) correlate very well, bu
the HD results have a 40%
higher value. This is due t
three factors: testin
orientation, contact resistance
and density variation.

Steady State (Guarded Hot Plate)

Canney, Dixon, Mathis / 603

CONCLUSIONS BIOGRAPHIES

The thermal conductivity of four samples of Brad Canney

polyurethane have been tested with three instruments with Brad is a Research and
different test methodologies. Development Scientist at
The TC Probe, which was calibrated with GHP Mathis Instruments Ltd. He
characterized samples, showed excellent correlation and holds a B.Sc. in Chemistry from
value agreement with the materials tested. Variation from the University of New
GHP vales was under 2%; within the precision of the Brunswick, Canada. His
instrument. ongoing work will continue to
The Hot Disk, which did not require calibration, showed focus on new applications of
a correlation to the GHP results of 0.93, but had a bias of thermal conductivity testing
40%. This bias has been postulated to be due to a with transient techniques.
combination of measurement orientation, contact resistance
and material density variations.
The transient methods were evaluated to demonstrate Craig Dixon
that rapid thermal conductivity techniques are available Craig Dixon is the Applications
and provide relevant information to the polyurethane Specialist at Mathis Instruments
industry. They eliminate sample size requirements and Ltd. He holds a B.Sc. in
produce results in seconds rather than minutes. This brings Chemistry from the University
advantages for the R&D and QC areas. of New Brunswick, Canada. He
has focused on rapid transient
REFERENCES thermal conductivity testing of
polyurethane foam and VIPs
1. American Society for Testing Materials, Standard Test since 1997.
Method for Steady-State Heat Flux Measurements and
Thermal Transmission Properties by Means of the
Heat Flow Meter Apparatus, Designation: ASTM Nancy Mathis
C 518-85, Philadelphia, PA, 1985. Nancy Mathis is the President
2. Mathis, N.E. ; Measurements of Thermal Conductivity of Mathis Instruments. She
Anisotropy in Polymer Materials, Ph.D. Thesis, Chem. holds a Ph.D. in Chemical
Eng. Dept., University of New Brunswick, Engineering from the University
Fredericton, N.B., Can., 1996. of New Brunswick, Canada.
3. Mathis, N.E., Chandler, C., “Orientation and Position She is the inventor and patent
Dependant Thermal Conductivity”, Proceedings of the holder for the technology
Society of the Plastics Industry’s Polyurethanes embodied by the TC Probe™
International Conference and Exposition, September thermal conductivity instrument
1999. Also accepted by Cellular Plastics, November distributed by PerkinElmer.
1999.
4. Cahill at University of Illinois at Urbana-Champaign.
(cf. Physical Review B 50 (1994) 6077 and J. Appl.
Phys. 81 (1997) 2590).

604 / Canney, Dixon, Mathis

POSTER SESSION
RIGID FOAM
Amidic Polyols, New Polyhydroxylic Raw Materials for Rigid
Polyurethane Foams

M. IONESCU, S . MIHAI a n d E. STEPAN C. ROIBU a n d V. PREOTEASA

Institute o f Chemical Research S. C. OLTCHIMS.A.
Splaiul Independentei 202 Uzinei nr. 1 Street,
77208, Sector 6, Bucharest 1000, Romnicu Valcea
Romania Romania

ABSTRACT

The aromatic polyols (ex. Mannich polyols), due to their isophthalate, dimethylphthalate, methylbenzoate, methyl
intinsique aromaticity, lead to rigid polyurethane foams salycilate or o f the corresponding acids with
with high physico-mechanical, thermal and flame diethanolamine are obtained condensation products,
proofing properties. In the present paper has been studied under form o f glassy solids at room temperature, but
the synthesis of new aromatic polyols having the stirrable viscous liquids at the normal temperatures of
structure of totally substituted amides and the alkoxylation ( 90-130°C). By the polyaddition of
substituents at the amidic nitrogens are hydroxyalkyl propyleneoxide or ethyleneoxide to the synthesised
( hvdroxyethyl, hydroxypropyl etc.). The aromatic condensation producs are obtained new liquid amidic
am idit polyols are obtained by the reaction of polyols, having convenient viscosities and with hydroxyl
dialkanolamines ( diethanolamine, diisopropanolamine number of 300-500 mgKOH/g, exactly in the domain
etc.) with dicarboxylic or polycarboxyl ic aromatic acids frequently used in the rigid polyurethane foams. The
or their esters or anhydrides. Thus, by the condensation reactions involved are:
reaction of dimethylterephthalate (DMT) , dimethyl

Using the synthesised amidic polyols were obtained rigid groups. The paper presents the properties of the
polyurethane foams with excellent physico-mechanical , synthesised aromatic amidic polyols and of the resulted
thermal and flame retardant properties. As compared rigid polyurethane foams as compared with standard
with amino-polyols , the reactivity of the amidic polyols polyols. Due to the accessibility and convenient cost of
it is not so high , because the catalytic effect of amidic raw materials, due to the simplicity of the synthesis
groups are minor, much lower than those o f aminic which needs only conventional apparatus for

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 607

condensation and alkoxylation, the studied new amidic Veronica Preoteasa
polyols present a real applicative potential for high Veronica Preoteasa obtained
performance rigid polyurethane foams. her Master's degree in
Organic Chemistry from
REFERENCES Polytechnical Institute of
Timisoara in 1981. She
1. M.Ionescu, V.Zugravu, I.Mihalache, S.Mihai, joined S.C. OLTCHIM S.A.
Advances in Urethane Science and Technology, and in about 19 years
1998, Vol.14: 151-218 become Director of Division.
She has a very important
BIOGRAPHIES contribution in the
diversification of the types of polypropylene glycoles
Dr. Mihail Ionescu and polyether polyols for flexible foams ( block foams
Dr. Ionescu Mihail obtained and moulded foams), graft polyols, polyethers polyols
his Master's degree in for rigid PU foams (sucrose and Mannich polyols ) made
Chemistry and Tecnology in industrial plants. She had an important contribution in
of Macromolecular the development of PUR foams in S.C.OLTCHIM S.A.
Compounds in (sandwich panels and spray foams) based on the own
Polytechnical Institute of polyols.
Bucharest-Faculty of
Industrial Chemistry in Stanca Mihai
1969 and in 1986 obtained
his Ph.D. in Polymer Stanca Mihai obtained her
Science. He joined Institute Master’s degree in
of Chemical Research from Bucharest and has worked Macomolecular Chemistry
30 years in the group for the synthesis of polyether from Polytechnical Institute-
polyols for polyurethanes. He developed many research Faculty of Industrial
works ( in laboratory , pilot , industrial scale ) for the Chemistry-Bucharest in
synthesis of telechelic polyethers for polyurethanes. He 1972. Since 1972 she has
is author of 66 patents, 86 communications and papers aquired experience in
and around 220 applicative research works. In present is research works concerning
the Scientific Director of Institute of Chemical Research the synthesis of
from Bucharest-Romania. polyurethanes (flexible, semiflexible and rigid foams) in
Institute of Chemical Research - Department Polymer
Dr.Constantin Roibu Modification, from Bucharest. Author of 28 patents and
Dr. Constantin Roibu 41 communications and papers. She has elaborated many
received his Master's technologies for systems and for polyurethanes applied
degree in Institute for Oil in industry.
and Gas from Ploiesti in
1977. He obtained his Dr. Emil Stepan
Ph.D. in Organic
Chemistry in University
Dr. Stepan Emil received his
"Polytechnica'' Bucharest
Master’s degree in University
in 1999. He joined S.C.
’’Polytechnica” Bucharest in
OLTCHIM S.A. from
1972 and obtained his PhD.
Ramnicu Valcea, one of the bigest petrochemical
in Organic Chemistry at the
complexes from Romania, specialised on the
same University in 1985 .
valorification of natural sodium chloride (electrolysis,
Dr.Emil Stepan proved
soda, chlorine, chlorinated solvents, vinylchloride,
serious experience in fine
polyvinylchloride, alkylamines , propylene-oxide etc.),
organic synthesis including
oxosynthesis, pesticides etc. He had a decisive
pesticides, biologic active compounds, organic
contribution in the development at industrial scale of
inteermediates and tensides. He developed successfully
propyleneoxide, propyleneglycole, polyethers for
since 1998 the research works concerning amidic-
flexible and rigid PU foams and rigid polyurethanes
polyols, which proved to be excellent starters for new
( sandwich panels and spray PUR foams) based on the
high performance aromatic polyols. In present Dr.Emil
romanian technologies. In present Dr. C.Roibu is the
Stepan is the chief of a research group in Institute of
General Manager of S.C.OLTCHIM S.A., Ramnicu
Chemical Research, Bucharest Romania, Department
Valcea., Romania.
Organic Intermediates and Tensides.

608 / Ionescu, Mihai, Stepan, Roibu, Preoteasa

Diffusion of Gases in Rigid Polyurethane Foam at
Different Temperatures

MORGAN FROLING, SARA MANGS a n d MARIA E. OLSSON a n d ULF JARFELT

OLLE RAMNAS Department o f Building Physics
Department o f Chemical Environmental Science Chalmers University o f Technology
Chalmers University o f Technology SE-412 96 Goteborg
SE-412 96 Goteborg Sweden
Sweden

ABSTRACT also be present as a condensed liquid in the cells. This

reservoir of cyclopentane can compensate for losses of the
If the diffusion behaviour in a polyurethane (PUR) foam gas due to diffusion [5,6]. Thus it keeps up the
is known for all cell gases, it is possible to predict the concentration of cyclopentane in the gas phase in the cells
change in gas composition and consequently the change in of the foam, which is beneficial for the long-term thermal
thermal conductivity over time for any configuration or performance
dimension of insulating foam [1,2]. Carbon dioxide, which The diffusion coefficients of oxygen, nitrogen, carbon
is always present, diffuses much faster out of the foam than dioxide and cyclopentane in PUR foam were determined in
e.g. cyclopentane. A simultaneous inward diffusion of air the temperature range 20-60°C. Samples were taken from
reduces the insulating capacity of the foam [3,4]. district heating pipes: carbon dioxide-blown foam (density
Calculation of heat loss presupposes knowledge of the 71 kgm"3) and cyclopentane blown foam (density
initial partial pressures of the cell gases in the foam and 61 kg m’3). The cell gas concentrations were determined
their diffusion coefficients at operational temperatures. after grinding the foam sample and analysis of the released
One application of PUR foam is in district heating pipes, cell gases by gas chromatography [7]. The effective
transmitting hot water (80-100°C) for space heating diffusion coefficients in the PUR foam were evaluated by a
purposes. The pipes consist of a PUR insulated steel pipe curve fitting procedure from the change of the partial
with a polyethylene casing. The temperature gradient over pressures of the cell gases over time [1,6]. See Table 1.
the foam cross-section influences the cell gas pressure and Very different activation energies of the cell gases, as in
thus the diffusion properties. our study (See Table 1.), means that the insulating
Use of cyclopentane as blowing agent in rigid PUR foam performance determined at a certain time and temperature
insulation has increased during the last decade, especially cannot easily be extrapolated. Thus, the diffusion rates of
in Europe and Japan. In the foam, cyclopentane is present these gases must be affected by temperature to a different
as a gas in the cells and dissolved in the polymer matrix. extent, which is in agreement with die effective diffusion
Due to concentration and temperature, cyclopentane may showed in Table 1.

Table 1. Effective diffusion coefficients and activation energiesfo r samples o f PURfoam taken from district heating pipes
(densities 61-71 k g m 3).

Gas Effective diffusion coefficient Activation energy

D.rrdO^mV1) E d (J - m ole1)
20°C 40 °C 60°C
Nitrogen 25 220 44 -103
Oxygen 150 650 30 - 103
Carbon dioxide 500 1300 19-103
Cyclopentane 0.6* 4 7 59 - 103
* __ . - ________
*Determined at 23 °C

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 609

 PUR foams exposed to elevated temperatures get darker BIOGRAPHY
and become more brittle. This is probably due to
oxidation of the polymer. In an unpublished study we Morgan Froling
observed that the amount of oxygen in PUR foams stored Morgan FrOling is Ph.D. student at the Department of
at 100°C was less than in foams stored at 80°C. Hence, Chemical Environmental Science, Chalmers University of
oxygen must have been consumed in an oxidation Technology, Sweden. He received his MSc in 1996 in
process. This means that oxygen diffusion coefficients Chemical Engineering and currently studies
based on studies of cell gas concentrations at high environmental aspects of polyurethane foam and diffusion
temperature may be underestimated. behaviour of blowing agents.

REFERENCES Sara Mangs

Sara Mangs is Ph.D. student at the Department of
1. Olsson, Maria E. (2001): Long-term Thermal Chemical Environmental Science, Chalmers University of
Performance o f Polyurethane Insulated District Technology, Sweden. She received her MSc in 2000 in
Heating Pipes, Doctoral thesis, Department of Chemical Engineering with Engineering Physics and
Building Physics, Chalmers University of currently studies the environmental performance and
Technology, Goteborg, Sweden. diffusion of blowing agents.
2. Froling, Morgan (1999): Environmental Aspects of
Polyurethane Insulated District Heating Pipes, Olle Ramnas
Licentiate thesis, Department of Chemical Olle Ramnas is Associate Professor at the Department of
Environmental Science, Chalmers University of Chemical Environmental Science, Chalmers University of
Technology, Goteborg, Sweden. Technology, Sweden. He works in the area of
3. Svanstrom, M., Ramnas O., Olsson M. E. and U. environmental analytical chemistry, especially used as a
Jarfelt (1997): Mass Transfer o f Carbon Dioxide tool for the study of the environmental performance of
through the Polyethylene Casing o f District Heating products in a life cycle perspective.
Pipes. J. Thermal Ins. and Building Env., 21:171-184
4. Olsson, M., Jarfelt, U. and O. Ramnas (1999): M aria E. Olsson
Diffusion o f Carbon Dioxide and Cyclopentane in Maria Olsson received her Ph.D. in 2001 at the
Polyurethane-Insulated District Heating Pipes, Department of Building Physics, Chalmers University of
Proceedings of The 7th International Symposium on Technology, Sweden. The title of her Ph.D. thesis is given
District Heating and Cooling, May 18-20, 1999, in reference [1]. She received her MSc in 1993 in Civil
Lund, Sweden. Engineering.
5. Olsson, M.E., Jarfelt, U., Froling, M., Mangs, S. and
O. Ramnas (2001): Diffusion of Cyclopentane in Ulf Jarfelt
Polyurethane Foam at Different Temperatures and Ulf Jarfelt is Associate Professor at the Department of
Implications for District Heating Pipes. Submitted for Building Physics, Chalmers University of Technology,
publication. Sweden. He received his PhD in 1987 and is currently the
6. Svanstrom, Magdalena (1997): Blowing Agents in leader of an interdisciplinary research group dealing with
Rigid Polyurethane Foam. Doctoral thesis, long-term thermal performance, environmental and
Department of Chemical Environmental Science, thermo mechanical questions within the area of district
Chalmers University of Technology, Goteborg heating pipe products.
Sweden.
7. Svanstrom, M and O. Ramnas. 1995: A Method for
Analysing the Gas Phase in Polyurethane Foam. J.
Cellular Plastics, 31:375-388

6 1 0 / Froling, Mangs, Ramnas, Olsson, Jarfelt

Dimethoxymethane as Co-Blowing Agent with HFC 134a for
Polyurethane Rigid Foams

M. MODESTI a n d A. LORENZETTI G. BASILE

Padova University Fluorocarbons and Development Service
v. Marzolo 9 Ausimont s.p.a.
35131 Padova V. le Lombardia 20
Italy 20021 Bollate
Italy

ABSTRACT mgKOH/g, Tercarol G310 (Enichem): n°OH=600

mgKOH/g, Tercarol G600 (Enichem), n°OH=290
In this research, rigid polyurethane foams blown with mgKOH/g;
HFC 134a, have been synthesized by hand mixing • Polyester polyols: Isoexter 4530 (Coim): n°OH=510
techniques, dissolving the gaseous blowing agent 134a in mgKOH/g; Isoexter 3340 (Coim): n°OH=360
a co-blowing agent. This is a suitable liquid ether mgKOH/g;
compound, dimethoxymethane, which can solubilize large • Polymeric MDI (methane diphenyl diisocyanate):
amount of 134a. Tedimon 385 (Enichem): NCO% = 30.5; average
We have studied the effect of foams formulations, in functionality = 2.8;
particular the nature of polyols, on blowing agent • Catalyst: dimethylcyclohesilamine (DMCEA),
compatibility and on physical and mechanical properties pentamethyldiethylentryamine (PMDETA) (Abbot),
of the polyurethane foams. Therefore several foams have • Surface-active agent: polysiloxane-polyether
been prepared employing different polyols, that is polyols copolymer Tegostab B8465 (Goldschmidt-Italy);
both polyether and polyester, both aliphatic and aromatic. • Blowing agent: HFC 134a (Ausimont), HFC 365mfc
Foams blown with n-pentane, HCFC 141-b and HFC (Solvay), HCFC 141b (Solvay), methylal
365mfc have also been prepared for comparison. (dimethoxymethane, Ausimont), n-pentane (Carlo
All the polymers produced have been physically and Erba);
mechanically characterized through evaluation of apparent The properties of blowing agents employed are reported
density, thermal conductivity, compression strength and in Table 1. The foams formulations are reported in Table
dimensional stability. 2. The foams were prepared by means of hand mixing
technique; the foam density was normalized to about 35
EXPERIMENTAL kg/m3 by varying the dosage of blowing agent. After
preparation, the foams were placed in an oven at 70°C for
The raw materials employed were die following: 24 hrs before carrying out physical-mechanical
• Polyether polyols: Arcol 3544 (ARCO chemical characterization.
products): n°OH=500 mgKOH/g; Kitane (Enichem)
n° OH = 435 mgKOH/g; RN 490 (Dow): n°OH=490

Table 1. Properties of the blowing agents employed.

n-pentane HCFC 141b HFC 365mfc HFC 134a methylal
Molecular weight 72 117 148 102 76.1
Boiling point [°C1 36 32 40.2 -27 42

POLYURETHANES EXPO 2001 , SEPTEMBER 30-OCTOBER 3, 2001 611

 Table 2.Formulations, kinetic parameters and density of the foams produced.
1 2 3 4 5 6 7 8

Arcol 3544 [g] 70 0 0 0 0 0 0 0

Kitane [g] 30 0 0 0 0 0 0 0
RN 490 [g] 0 100 100 100 0 0 0 0
Tercarol G310 [g] 0 0 0 0 60 0 100 0
Tercarol G600 [gj 0 0 0 0 40 0 0 0
Isoexter 4530 [g] 0 0 0 0 0 60 0 60
Isoexter 3340[g] 0 0 0 0 0 40 0 40
DMCEA [g] 1.7 1.6 1.4 1.6 1.6 1.3 1.8 1.5
PMDETA [g] 0.6 0.5 0.5 0.6 0.6 0 0.2 0
Tegostab B8465 [g] 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
H2O [g] 1.6 1.5 1.5 1.5 1.5 1.5 1.6 1.5
HFC 134a [g] 0 0 0 9.2 10.3 12.9 12.4 9.9
Methylal [g] 0 0 0 6.8 7.7 3.2 10.6 6.1
HFC 365mfc [g] 0 0 23 0 0 0 0 0
HCFC 141b [g] 0 18 0 0 0 0 0 0
n-pentane [g] 10 0 0 0 0 0 0 0
Tedimon 385 [g] 153 159 159 159 154 147 188 147

Cream time [s] 16 10 15 16 18 25 35 13

Gel time [s] 55 59 75 87 83 45 75 30
Tack free time [s] 68 87 95 99 96 66 100 57

Density [kg/m3] 33 36 34 34 33 36 33 36

RESULTS material: the foam isotropy index (FII), defined as the

ratio of compression strength parallel to that
The results of physical -mechan ical characterization are perpendicular to foam rise [1], is 40% higher than in
reported in Table 3. The apparent density has been presence of polyether polyols. All the foams show good
evaluated according to ISO 845, the thermal conductivity performances on dimensional stability: the change on
according to ISO 8301, the compression strength dimensions is almost always lower than 1% for all the
according to ISO 844 and the dimensional stability testing conditions. The best performances were achieved,
according to ISO 2796. like for the compression strength, with polyester polyols.
As it can be observed, the compression strength of the About the thermal conductivity, it is possible to note that
foams blown with the mixture methylal-134a is quite the values obtained for the foams blown with the mixture
similar to that of the foams blown with n-pentane, HFC methylal-134a are comparable or even lower than that of
365mfc and HCFC 141b. In presence of polyester polyols n-pentane blown foam.
it is possible to note a fairly high anisotropy of the

Table 3. Compression strength, thermal conductivity and dimensional stability of the foams produced.
1 2 3 4 5 6 7 8
Compression strength [kPa]
parallel 208 226 175 173 150 235 170 250
perpendicular 74 150 104 100 86 76 101 89

Thermal conductivity [mW/m K] 26.9 22.9 23.1 28.3 26.7 28.9 28.5 27.3

Dimensional stability [AL%]

70°C *24 hr +0.20 -0.03 -0.15 -0.58 +0.58 +0.26 -0.28 +0.24
70°C, 95% RH * 24 hr +0.45 +0.37 -0.24 -0.17 -0.17 +0.0 -0.82 +0.05
-25°C *24 hr -0.31 -0.08 -0.15 -0.21 -0.21 +0.02 -0.35 +0.01
70°C *7 days +1.47 +0.46 +0.20 +0.18 +0.18 +0.33 +0.17 +0.27
70°C, 95% RH *7 days +1.65 +0.73 -0.31 -0.36 -0.36 +0.0 -0.39 +0.01
-25°C *7 days -0.69 -0.16 -0.35 -0.50 -0.50 +0.17 -0.48 +0.19

612 / Modesti, Lorenzetti, Basile

CONCLUSIONS Giampiero Basile
Giampiero Basile received his
The blend of HFC 134a-dimethoxymethane can B.S. degree in Industrial
advantageously be used as a more environmentally Chemistry in 1966 at the
friendly foaming agent owing to its content of University of Genua. He is
dimethoxymethane. This component, belonging to the currently working at the
acetal family, has in fact no ODP, no VOC and negligible Development & Technical
GWP compared to HFC 134a. These features, along with Service for Fluorocarbons of
its extremely low toxicity and high boiling point, allow the Ausimont S.p.A., Montedison
blend with HFC 134a to be safe and more easy to be Group, Bollate (Milano). His
handled. The compression strength, dimensional stability current position involves the
and thermal conductivity obtained with this mixture are development of CFC’s
always the same or better than those obtained with n- substitutes HCFC, HFC and HFE. Nowadays his
pentane blown foams. competence is comprehensive of the studies carried out on
Further the use of methylal, a flammable compound, in the relationship between the chemical structure of small,
blend with nonflammable HFC 134a can deeply restrict halogenated molecules and their potential applications. He
the flammable behavior in air of dimethoxymethane. is author of 15 patents and contributed to numerous papers
Therefore, the use of such mixture allows achieving the and publications in his country and abroad.
same or better results obtained with n-pentane blown
foams, and is safer than the use of n-pentane.

REFERENCES

1. Frisch K. C., Sendijarevic V., Sendijarevic A. et Al.

1991. “New heat resistant isocyanate based foams for
structural applications”, presented at Polyurethane
World Congress, September 24 -26,1991.

BIOGRAPHIES

Michele Modesti
Michele Modesti graduated in
Chemical Engineering from
Padova University in 1984, has
developed his research in the
field of polyurethanes. He
attained the Ph.D. at the
University of Padova in 1989,
with a thesis on fire behavior of
polyurethane systems. He has
been a researcher in Chem.
Eng. Dept, of Padova since
1991 in the field of polyurethane recycling.

Alessandra Lorenzetti
Alessandra Lorenzetti, received
her degree in Chemical
Engineering in 1999 from
Padova University. She is
actually a Ph. D. student. Her
research deals mainly with fire
behavior of polyurethane and
interpenetrating polymer
network (IPN), alternative
blowing agents and ageing of
polyurethane rigid foams.

Modesti, Lorenzetti, Basile / 613

Glycolysis Reactivity of Rigid Polyurethane Foam
for Refrigerator

M. MURAI, M. SANOU, T. FUJIMOTO a n d F. BABA

Mitsubishi Electric Corporation
Advanced Technology R&D Center
8-1-1, Tsukaguchi-Honmachi, Amagasaki-City
Hyogo, 661-8661
Japan

ABSTRACT BIOGRAPHIES

In Japan, The Law for Recycling of Specified Kinds of M khio M urai

Consumer Electric Goods has been enforced on April 1,
2001. And Japanese Appliance Manufacturers are obliged Michio Murai joined Mitsubishi
by this law to undergo the recycling of wasted refrigerators Electric Corporation in 1992
as well as air conditioners, televisions and washing after receiving a B.S. and M.S.
machines. degree in Applied Chemistry
Rigid polyurethane foam, about 10wt% of a refrigerator, from Osaka University in 1990
is used as the thermal insulators. The Japanese Appliance and 1992, respectively. He has
market alone has an annual production of about 25,000 been engaged in development of
tons. But the recycling processes of rigid polyurethane chemical recycling of insulation
foams have not been yet established. materials for refrigerator and
We have studied chemical degradation methods of rigid showcase currently.
polyurethane foams to raw materials (polyols) for the
future target of the recycling law (recycling ratio of M uneharu Sanou
refrigerator; 90%). In this study we examined glycolysis
reactivity of rigid polyurethane foam under various Muneharu Sanou joined Mitsubishi Electric Corporation in
reaction conditions. The degradation rate was 1973. He has been engaged in development of plastics
approximately doubled for each 10°C rise in the range of molding and the life cycle assessment method.
170 to 200°C. This result shows the glycolysis reaction of
rigid polyurethane foam obeys Arrhenius’s law. The Takamitsu Fujimoto
degradation rate was proportional to the catalyst
concentration in the case of KOH as catalyst. But Takamitsu Fujimoto joined Mitsubishi Electric Corporation
dibutyltindilaurate concentration had less influence on the in 1971. He has been engaged in development of electric
degradation rate. Faster degradation of polyurethane foam insulation materials. He is currently the manager of
was achieved with smaller particle size and initial advanced materials group.
degradation rate was proportional to the surface area of
foam particles. Polypropylene glycols had higher Fumiaki Baba
glycolysis reactivity than polyethylene glycols. Glycolysis
reactivity also depended on the molecular weight of glycol Fumiaki Baba joined Mitsubishi Electric Corporation in
and dipropylene glycol displayed the highest reactivity. 1973. He has been engaged in development of plastics and
This work was supported, in part, by the New Energy plastic molding. He is now responsible for the ecomaterials
and Industrial Technology Development Organization. and material analysis department.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 615

HFC-134a as a Co-Blowing Agent with Pentanes

JINHUANG WU a n d DOUG DILLON

ATOFINA Chemicals, Inc.
900 First Avenue
King o f Prussia, Pennsylvania 19406

ABSTRACT BIOGRAPHIES

The Montreal Protocol regulates the uses of Jinhuang W u is a Research Scientist at

ozone depletion substances (ODSs). In the Atofina Chemicals Technological Center in King
developed countries, the phaseout of high ozone of Prussia. He received his PhD in
depletion potential (ODP) CFCs in rigid foam Macromolecular Science and Engineering from
applications was made possible with the the University of Michigan, and joined Atofina
introduction of lower ODP HCFCs. HCFC -14lb in March 1997.
was chosen as an excellent replacement for CFC- Doug Dillon is a Research Chemist at Atofina
11 in many rigid foam applications. HCFC-141b Chemicals Technological Center in King of
blown rigid foams offer excellent thermal Prussia. He received his BS in Chemistry from
insulation, fire resistance and dimensional Juniata College, and joined Atofina Chemicals in
stability. May 1997.
According to the Montreal Protocol, HCFC-
141b will eventually be phased out. Lower ODP
HCFCs ( i.e., HCFC-142b and HCFC-22), zero
ODP HFCs and hydrocarbons are considered to
be alternatives. Unfortunately, neither HFCs nor
pentanes can match performance and cost of
HCFCs.
An HFC and pentane(s) coblowing system can
potentially offer some advantages over purely
pentane blown foams including lower foam
density (therefore lower cost), better thermal
insulation and better fire performance than
purely pentane(s) blown foams. The coblown
foam has lower cost than HFC blown foams.
In this study, we choose HFC -134a, a
commercially available blowing agent, as a
coblowing agent with pentane(s). We make
polyisocyanurate (PIR) foams with HFC -134a as
a coblowing agent with n-C5, i-C5, i-C5/c-C5,
and the levels o f HFC -134a are 10 and 20
mole%. We determine thermal conductivity (or
k-factor) of foams from -8 to 40 °C. The results
show that the effects of HFC -134a on the low
temperature k -factors are different, depending on
the pentane isomers. Furthermore, we screen a
series of surfactants to determine their effects on
the k-factors.

POLYURETHANES EXPO 2001 , SEPTEMBER 30 -OCTOBER 3, 2001 617

HFEs as New Generation Blowing Agents

R. TAMAI a n d S . URATA A. SEKIYA

Dept, fo r the New Refrigerant and Other National Institute o f Advanced Industrial Science and
Substances Research Technology (AIST)
RITE, c/o AIST Fluorine Center
Central 5, 1-1-1, Higashi, Tsukuba, 305-8565 Central 5, 1-1-1, Higashi, Tsukuba, 305-8565
Japan Japan

H. TAKEYASU a n d H. SATO
Asahi Glass Co., Ltd.
3-474-2, Tsukakoshi, Sawai-ku
Kanagawa-shi, 210
Japan

ABSTRACT 4. K.Tokuhashi et al., Int. J. Chem. Kinet., 1999, 31, 846.

5. K .Tokuhashi et al., J. Phys. Chem. A, 2000, 104,
The development of blowing agents that exhibit lower 1165.
thermal conductivity, shorter atmospheric lifetime, and BIOGRAPHIES
less adverse effects on the atmosphere is an urgent issue.
We investigated four fluorinated ethers as alternative
Ryouichi Tamai
blowing agents (1,1,2,2-tetrafluoroethyl methyl ether R. Tamai jointed Central Glass Co., Ltd. in 1987 after
(HFE-254pc), 1,1,2,2-tetrafluoroethyl difluoromethyl ether receiving his B.S. and M.S. in Industrial Chemistry from
(HFE-236pc) and 1,1,2,2,3,3,3-heptafluoropropyl methyl Yamaguchi University in 1985 and 1987. He has been
ether (HFE-347mcc)) in terms of physical properties engaged in R&D of developing new alternatives to CFCs
(Tablel), gaseous thermal conductivity (Table2)[2], from 1990 and was also engaged in R&D of HCFC-141b.
environmental adaptability (atmospheric lifetime) [4] [5], He also participated RITE project in 1998.
toxicity, flammability, solubility in polyols, and thermal
stability. Shingo U rata
By using these fluorinated ethers as blowing agents, we S. Urata joined Asahi Glass Company in 1997 after
experimentally produced rigid polyurethane foams, which receiving his B.S. and M.S. in Physics from Kyushu
all showed thermal conductivities lower than or equal to University in 1995 and 1997. He was engaged in R&D
foams produced with the blowing agent alternative using Computational Molecular Design system. He
candidates, HFC -245fa and cyclopentane. participated RITE project in 1999.
Included in this study are the newly discovered Hiromitsu Takeyasu
azeotrope or azeotrope-like of HFEs with HFC-245fa or H.Takeyasu is a director of Urethane Form Group at
cyclopentane. Chemicals R&D Division after Developmental work on
catalysts and new polyols.
ACKNOWLEDGMENTS Hisashi Sato
This works was supported by the New Energy and H.Sato has worked for R&D of Rigid Urethane Form after
Industrial Technology Development Organization (NEDO). receiving B.S. and M.S. degree in Chemistry from
We also thanks for the support this research by the member Touhoku University.
of this project.
A kira Sekiya
REFERENCES Dr.Akira Sekiya is the deputy director of Reserch Center
for Developing Fluorinated Greenhouse Gas Alternatives.
I. N.Takada, R.Tamai, H.Yamamoto, A.Sekiya, After getting his doctor degree in Tokyo Institute of
N.Tsukida and H.Takeyasu, J. CELLULAR PLASTICS,35 Technology in 1977, he worked as postdoctoral fellow in
(1999) 389. USA for three years and also worked at Sagami Chemical
2. N.Takada, S.Matsuo, Y.Tamnaka and A.Sekiya,. J Research Center and Tokyo Institute of Technology. He is
Fluorine Chem. 91 (1998) 81. working the research of developing new alternatives to
3. S.Matsuo, Y.Tamnaka, N.Takada, H.Yamamoto and CFCs from 1989. Now he jointed the collaborate research
A.Sekiya, J. Chem. Eng. Data, 43(3), 473. to RITE project from 1990.

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 619

 Table 1. Physical properties of blowing agent candidate ethers
HFE-254pc HFE-236pc HFE-347mcc HFC-245fa HFC-141b CFC-11
Boiling point (°C) 37.19 27.94 34.18 15.2 32.1 23.8
Vapor pressure (kPa) 63.1 90.3 71.8 123 68.9 88.3
Latent heat of vaporization (kJ/kg)*** 223.84 174.60 142.26 208.80 223.27 180.50
Specific heat (mJ/kg K)** 1.444 1.386 1.246 1.369 1.202
Viscosity (mPa.s)** 0.448 0.471 0.422 0.44 0.42
Density (kg/m3)** 1288.0 1476.5 1409.4 1344.3 1240 1490
Surface tension (mN/m)** 16.97 13.80 12.39 12.35 18.2 18
Thermal conductivity (W/m.K)** 0.1059 0.0996 0.0753 0.0893 0.0929 0.0862
Solubility in water (g/100g)** 0.24 0.25 0.01 0.56 0.090 0.011
* at 25°C , **:at 23°C , ***:at boiling point

Table 2. Gaseous thermal conductivities of hydrofluoroethers

Blowing agent Thermal conductivity (mW/m*K) Bp.
Converted
Code Formula at 50°C at 70°C
value to 50°C
(°C)
HFE -236pc c h f 2c f 2o c h f 2 13.19 - 13.19 28
HFE-245mc c f 3c f 2o c h 3 13.81* - 13.81* 6
HFE - 245mf c f 3c h 2o c h f 2 13.75* - 13.75* 29
HFE - 245pf c h f 2c h 2o c f 3 14.00 - 14.00 26
HFE-245qc c h 2f c f 2o c h f 2 - 14.82* 13.08 43
HFE - 254pc c h f 2c f 2o c h 3 13.34* - 13.34* 37
HFE-338mcf c f 3c f 2c h 2o c f 3 13.41 - 13.41 26
HFE -347mcc c f 3c f 2c f 2o c h 3 12.79* - 12.79* 34
HFE-347mcf c f 3c f 2c h 2o c h f 2 - 14.67* 12.93 46
HFE-365mcf c f 3c f 2c h 2o c h 3 - 15.66 13.92 48
HFE-347pc-f c h f 2c f 2o c h 2c f 3 - 14.11* 12.37 56
HFE-347mmy (CF3)2CFOCH3 13.01* - 13.01* 29
HFC -245fa c f 3c h 2c h f 2 13.86 - 13.86 15
HFC-365mfc c f 3c h 2c f 2c h 3 14.06 - 14.06 40
HCFC-141b c c i 2f c h 3 11.65 - 11.65 32
Cyclopentane C5H io - 15.40 14.19 50
*:[21

620 / Tamai, Urata, Takeyasu, Sato, Sekiya

Hydrocarbon Blown Rigid Foam Formulations and
Physical Properties

JOHN CANADY, CARINA MCADAMS a n d PHILIP MERCHANT

KELLY BROWN Exxon Mobil
KoSa 5200 Bayway Drive
4600 Highway 421 North Baytown, Texas 77520-2101
Wilmington, North Carolina 28402

ABSTRACT

Most rigid foam manufacturers have started their research

in preparation for the January 1, 2003 phase-out of most
HCPC blowing agents, including HCFC 141b.
Manufacturers and development labs alike are currently
being challenged with tight conversion schedules and
aggressive performance targets. Part of the conversion
process is the optimization of basic formulations to achieve
the best available physical properties and manufacturing
efficiencies. Multiple polyols, surfactants, catalyst systems
and blowing agent(s) are typically screened as part of the
optimization process.

Hydrocarbons are the blowing agents of choice for many

applications because of their balance of reasonable k-factor
performance, high blowing efficiency and favorable
economics. A variety of hydrocarbon blends are available
for Exxon, as well as other suppliers. KoSa has introduced
new Terate® polyester polyols, designed to maximize the
performance of hydrocarbon-blown polyisocyanurate foam
formulations.

A survey of available Exxsol® blowing agents and Terate

pentane polyols is presented along with formulations,
flammability data, and physical testing results from
laboratory and machine trials. Suggested starting
formulations are presented and discussed for
polyisocyanurate “boardstock” and “bunstock”
applications. The thermal conductivity, dimensional
stability, and flammability properties for the resulting
foams are presented and compared to those of conventional
HCFC 14lb -blown foams. Results for a selection of other
alternate blowing agents are also presented and discussed.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 621

Influence of Expandable Graphite on the
Physical-Mechanical Properties and Fire
Behaviour of Flame Retarded PIR-PUR Foams

M. MODESTI, A. LORENZETTI, F. SIMIONI M. GILBERT

Padova University Graphtech, Inc.
Dept, o f Chemical Process Engineering 11709 Madison Ave.
v. Marzola 9 LakewoodO hio 44107
35131 Padova
Italy

ABSTRACT • Polymeric MDI (methane diphenyl diisocyanate):

Tedimon 385 (Enichem): NCO% = 30.5; average
Rigid Polyisocyanurate-Polyurethane foams (PIR-PUR) functionality = 2.8;
have been synthesized with expandable graphite (EG) as • Polyester polyols: Glendion 9801 (Enichem): n°OH
the flame retardant halogen free additive. In order to study = 351 mgKOH/g, viscosity to 25°C = 5600 mPa-s;
possible synergistic effects, mixtures of EG and Kitane (Enichem) n° OH = 435 mgKOH/g viscosity
triethylphosphate (TEP) were used to make fire retarded, to 25°C= 1100 m Pas;
pentane-blown PIR/PUR foams with a constant NCO • Catalyst: pentamethyldiethylentryamine (Abbot),
index equal to 250. To this point, there has been no potassium octoate (Air Products);
commercial production of such rigid PIR/PUR foams • Surface-active agent: polysiloxane-polyether
using expandable graphite as the fire retardant additive. copolymer Tegostab B8469 (Goldschmidt-Italy);
This study investigated the influence of the flame • Blowing agent: n pentane (Carlo Erba);
retardant additives on the foam physical and mechanical • Flame retardant: expandable graphite: Graf Guard
properties, such as compression strength and thermal 160-80N (Ucar Graph. Tech); triethylphosphate
conductivity. In order to evaluate fire performance, the (Eingenmann & Veronelli)
foams were studied using cone calorimeter analysis and The foams formulations are reported in Table 1. We
the oxygen index test. prepared foams characterized by a constant isocyanate
indexes (250) by means of hand mixing technique. The
EXPERIMENTAL foam density was normalized to about 38 kg/m3 by varying
the dosage of n-pentane. After that, the foams were placed
The raw materials employed were the following: in an oven at 70°C for 24 hrs before carrying out physical-
mechanical and fire behavior characterization.

Tablet. Foams formulations

Formulation Ref EG1 EG2 EG3 EG4 TEG1 TEG2 TEG3 TEG4 TEG5
Glendion 9801 70 70 70 70 70 70 70 70 70 70
Kitane 30 30 30 30 30 30 30 30 30 30
PMDETA 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.17
K-15 1.8 2.5 2.5 2.2 2.2 2.5 2.5 2.5 2.2 2.0
B 6489 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5
HaO 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5 1.5
n-pentane 10 12 14.5 20 24 24.2 25.7 27.2 24 26.3
Tedimon 385 290 290 290 290 290 290 290 290 290 290
GRAF Guard 160-80N 0 45.4 72.5 104.1 140 22.9 48.5 77.5 76 75.5
Triethylphosphate 0 0 0 0 0 13.7 14.5 15.5 10.1 5

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 623

 Figure 1. Thermal conductivity of EG filled foams.

RESULTS

The results obtained show that an increase in the About the fire behaviour, we found a considerable
amount of expandable graphite or triethylphosphate does increase in the O.I. values, particularly for EG-TEP foams
not significantly degrade mechanical properties. (Fig.2). It can be seen that the O.I. increases in a linear
Increasing the amount of expandable graphite does cause way increasing E.G. amount, while in presence of a
an increase in thermal conductivity (Fig.l), probably due constant amount of TEP (3%) the rise is exponential and
to an increase in the dimensions of the foam ceils. Instead, no longer linear. Even if for E.G. filled foams the
rising the triethylphosphate content, it has observed a improvement on O.I. values are satisfactory, for TEP-E.G.
decrease on thermal conductivity of about 1 mW/m K. filled foams they are surprising: in presence of 15% of
This could be due to the plasticizing effect of TEP, that E.G. and 3% TEP die O.I. is 75% greater than for unfilled
causes a decrease of the cells size of the foam, leading foams.
therefore to a decrease on thermal conductivity.

Figure 2. Oxygen index of EG and EG-TEP filled foams.

624 / Modesti, Lorenzetti, Simioni, Gilbert

 Figure 3. RHR peak and average values in function of the expandable graphite amount

The cone calorimeter results are in agreement with those the decrease in the rate of heat release (RHR) (up to 50%).
of O.I: EG and EG -TEP filled foams showed a The C 0/C 02 weight ratio (Fig.4), that is a index of smoke
considerable decrease in the rate of heat release (Fig.3) toxicity, is quite constant varying the filler amount. Only
with respect to unfilled foams. In particular, for EG-TEP in presence of very high EG content (25%) the ratio
foams, the higher the triethylphosphate content the greater increases considerably.

Figure 4. OO/CO2 average values in function of the expandable graphite amount

Modesti, Lorenzetti, Simioni, Gilbert / 625

CONCLUSIONS Alessandra Lorenzetti
Alessandra Lorenzetti received her
In this work we analysed the influence of expandable degree in Chemical Engineering
graphite and triethylphosphate on physical-mechanical in 1999 from Padova University.
properties and fire performances of PIR-PUR foams. Actually she is a Ph. D. student.
As it is resulted from physical-mechanical characterisation, Her research deals mainly with
the presence of filler causes only slight worsening on fire behavior of polyurethane and
mechanical properties, while the presence of EG leads to a interpenetrating polymer network
considerable worsening on thermal conductivity. On the (IPN), alternative blowing agents
other hand, the employment of low amount of TEP, thank to and ageing of polyurethane rigid
its plasticizing effect, allows the decrease of about lmW/m K foams.
of the thermal conductivity of the foams.
About the fire behaviour, we have observed that for filled
foams there is an overall improvement of fire behaviour: the Francesco Simioni
higher the filler content the higher the oxygen index and the Francesco Simioni graduated in
lower the rate of heat release; in particular, the best fire Chemical Engineering from
performances are obtain in presence of both TEP and EG. Padova University in 1961. He
was an organizer and teacher for
REFERENCES the Engineering Faculty of Padova
teaching a course in Technology
1. Modesti M., Lorenzetti A., Simioni F. et Al. 2001. “ of High Polymers until 1970.
Expandable graphite as intumescent flame retardant in since then he has become a
polyisocianurate-polyurethane foams” presented at Professor of Industrial Organic
8thEuropean Conference on Fire Retardant Polymers, Chemistry and has also worked as
June 24 -27,2001. a consultant for various industries
2. Hutchings D. A., Krassowsky D. W., Qureshi S. P. 1996. producing PRFV, polyurethanes and Interpenetrating
“Expandable graphite flake as an additive for a new flame Polymer Network (IPN).
retardant resin”, presented at Fire Retardant Chemicals
Association Fall Meeting, October, 19%.
3. Schiling B. 1997. “Expandierbarer graphit“, Kunststoffe,
87(8):1004-1006.
4. Okisaki F. 1997. “Flamecut GREP series new non-
halogeneted flame- retardant system “, presented at Fire
Retardant Chemicals Association Spring Meeting, March,
1997.
5. Miller B. 1996. “Intumescent, FR efficiency pace flame
retardant gains”, Plastic World, December 1996: 40.

BIOGRAPHIES

Michele Modesti
Michele Modesti graduated in
Chemical Engineering from
Padova University in 1984, has
developed his research in the field
of polyurethanes. He attained the
Ph.D. at the University of Padova
in 1989, with a thesis on fire
behavior of polyurethane systems.
He has been a researcher in Chem.
Eng. Dept, of Padova since 1991
in the field of polyurethane
recycling.

626 / Modesti, Lorenzetti, Simioni, Gilbert

Optimization of Flame-Retardants for Rigid PIR Foams: A New
Screening Apparatus, and Correlation to Large-Scale
Flammability Tests

BERT FESKE JOHN CANADAY

Albemarle Corporation KoSa
8000 GSRI Avenue 4600 Highway 421 North
Baton Rouge, Louisiana 70898 Wilmington, North Carolina 28402

ABSTRACT

Building Codes in the United States require that insulation

materials meet defined standards of flammability
performance. For roof insulation in particular, the Factory
Mutual Calorimeter test (part of the FM 4450 test
procedure), and Underwriters Laboratories 756 procedure,
among others, are critical performance criteria.
Optimization of the flame -retardant type and use level is a
key component in the formulation of polyisocyanurate
foams to meet the applicable flammability standards.

Traditionally, the industry has used small-scale methods in

an attempt to predict performance in the full-scale
flammability tests. These methods have been convenient
and inexpensive, but have not been entirely satisfactory as
predictive tools.

The previous small-scale flammability screening tests have

now been improved upon with the development of an
inexpensive larger-scale screening test that duplicates the
exposure temperatures of the full - scale tests. Assessment
of char formation, weight and dimension retention, and
char integrity can be made as a function of the flame-
retardant type. The isocyanate, polyol, catalyst and flame-
retardant package can be evaluated identify materials that
minimize cracking and shrinkage, and maximize the
thickness retention of the char. The apparatus also provides
for the testing of up to four samples simultaneously, under
identical conditions.

Data is presented to demonstrate he correlation of the new

test procedure to the existing FM and UL tests.
Recommendations are made for flame-retardant
combinations that provide the highest degree of char
integrity. A list of materials, cost estimate, and design
drawing of the apparatus is also included.

POLYURETHANES EXPO 2001, SEPTEMBER 30 -OCTOBER 3, 2001 627

Performance of Non-ODP Rigid Foams in Picnic Coolers

JONE CIONNI a n d V. MICHAEL MAUTINO

Bayer Corporation
100 Bayer Road
Pittsburgh, Pennsylvania 15205

ABSTRACT

Manufacturers of picnic coolers face serious challenges

and choices regarding impending blowing agent
technology changes, and many are currently exploring
options for meeting environmental regulations that will
curtail the use of ozone-depleting potential (ODP) blowing
agents such as HCFC-141b. Choices made over the next
several months will determine whether these manufacturers
can remain competitive in an industry that will likely suffer
from increased raw material costs for alternative HFC
blowing-agent technologies as well as soaring prices for
stockpiled HCFC-141b after its production ceases at the
end of December 2002. Therefore, comparative
information on available replacement blowing agent
technologies is both useful and timely.
This poster will present data generated in machine trials
of several rigid insulation systems produced with approved
non-ODP blowing agents. Included are HFC-134a, HFC-
245fa, and water, as well as combinations of the same.
The data includes foam reactivity, test-panel physical
properties, and results of Ice Retention (I-R) and Reverse
Heat Leak (Bayer RHL) applications tests on finished
coolers.

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 629

Sandwich Panel Lines: Comparison between Discontinuous
and Continuous Process. Advantages Given by the
Continuous Production Method Compared to the
Manufacture of Panels by Discontinuous Means

ERALDO GRECO a n d VITTORIO MARIANI

Impianti OMS Spa
Via Sabbionetta 4
20050 Verano Brianza (Mi)
Italy

ABSTRACT
discontinuous to continuous process as these plants offer
Since ‘energy saving’ has become the key requirement the optimum quality of end -product and high productivity.
for the building industry (roof, wall, architectural panels With more than 30 years experience in the market, OMS
etc.) and food preservation (cold store panels), the market Group has a leading position in the sandwich panel
for sandwich panels with PUR or PIR cores has shown a production field, offering discontinuous and continuous
constantly increasing demand all over the world. This is production lines. The main differences between the two
due to the excellent thermal insulation properties types of technologies presently used, can be summarised in
combined with the extremely good physical-mechanical Table 1.
features, which are provided by the polyurethane foam. Figure 1 shows typical polymerisation times of the PU
The manufacturers of PU insulated panels mainly in the foam between the two different production processes.
highly industrialised Countries, are demanding evermore In our example, we assume the production of a panel 12
higher quality standards. Further to that it is the number of meters long, 1.2 m. width and different thickness 50-100-
manufacturers who prefer the use of highly automated 150-200 mm
plants is increasing. An ever greater number of Companies Figure 2 shows the different productivity of the two
are changing their production technology from the types of plants

POLYURETHANES EXPO 2001 , SEPTEMBER 30-OCTOBER 3, 2001 631

 Table 1: Main differences between Discontinuous and Continuous technology

Discontinuous Panel Line Continuous Panel Line

Limited productivity Extremely high productivity

High labour costs, because several manual Labour expenses cut down to the minimum, because
operations are necessary in preparing the panels it is matter of fully automated equipment.
and in the manufacturing cycle.
Lower starting investment if compared to a The starting investment for a continuous line is
continuous line. Though costs increase according significantly higher, but due to its high productivity is
to the final composition of the plant itself with the amortizable in a short time
supply of a set of devices to automate the
production cycle and facilitate the work of the
operators (presses with movable platens, multi­
daylight press system, carriers for the movement
and the automatic foaming of the panels, bridge
crane systems to move the sheets and the
foamed panels etc. etc.).
Compared to the continuous production, the Excellent foam quality, uniform over the whole panel
quality of the foam is not homogeneous. It surface
depends on the different foaming process which
may be used.
Higher foam density value which is not uniform Very good distribution of the density, obtained by the
over the entire panel surface “open pour” pouring and distribution technology of the
PU foam. This means that the polyurethane is
uniformly laid down in between the two rigid facings
of the insulating panel, just before they enter into the
double belt conveyor. This also means a raw material
saving.
Long polymerisation times and consequent plant Really short polymerisation times and possibility to
limited (low) productivity ‘play1 with the PU reaction times, since all the
different components of the mixture are separately
injected (one by one) (catalysts, silicones, water,
blowing agent, flame retardant etc.etc.). This
simplifies the preparation of different types of foam
consequently having different features.
The manufactured panels have a length that only As far as a continuous plant is concerned the only
occasionally exceeds 12 meters which is the limit to the length of the panels is the ability to
normal maximum dimension of the press transport them. In many countries panels up to 24m
length are produced. This is a tremendous
advantage since it is possible to cover large areas
with a single panel and without overlaps.

The comparison has been done considering the use of a 12 minutes polymerisation time
press with 4 shuttle platens (OMS 2+2 Panel System, say 3 minutes “Service Time” (time necessary for
with two platens inside the press during the curing phase opening/closing of the press, to translate the platens of
of the PU foam and 2 platens outside the press for the the press and make the pouring cycle);
assembly and preparation of the next pouring cycle) consequently considering a total cycle time of 15
(Figures 3,4 , 5). minutes, the plant can make 4 cycles per hour. That
Assuming a production of panels having following means a production of 8 panels 12 x 1,2 x 0,05 m. each
dimensions (because of a 2+2 Panel Press)
discontinuous production the total productivity of the plant on a 7 working hours
12 m. total length shift corresponds to 806.4 square meters/day
1,2 m. working width 230 working days per year give a productivity of
50 mm thickness 185.472 square meters/year.

632 / Greco, Mariani

Figure 3: Typical press for discontinuous panel Figure 4: Typical press for discontinuous panel
production production

Greco, Mariani / 633

 Figure 6: Typical double belt conveyor for continuous Figure 7: Distribution traverse
panel production

Whenever an OMS Group Continuous Panel Line plant In this Particular field OMS Group has an in-depth
(Fig. 8) with a 30 meters length of pressure area (Fig. 6) experience in continuous and discontinuous lines which are
is used, it will be possible to produce the same panel at a currently in operation all around the World (from China to
production speed of 10 meters/minute. Canada, France, United Kingdom, Italy and USA, etc..).
Many sandwich panel manufacturers have chosen OMS
* Continuous production at 10 meters/minute speed Group as their ideal partner to realise their projects
* Panel width 1.2 meters thanks to continuous and constant involvement in the
* Thickness of the panel 50 mm technological research, OMS Group has recently planned
* 3 minutes polymerisation time. and realised two important projects that will revolutionise
* No manual operation involved continuous process production of sandwich panels.
* The productivity of the plant will be 12 square meters The first project concerns the high pressure mixing of all
for each minute of production, that means 720 square components used into the PU formulation directly to a
meters/hour = 5.040 square meters/day working shift special self-cleaning “flaf* head (MDI, POLYOL,
of seven hours. CATALYST, SILICONE, WATER, BLOWING AGENT
* 230 working days per year give a productivity of etc.). The head being small and light (less weight) can be
1.159.200 square meters/year. easily fitted on a distribution reciprocating traverse (Fig.7).

Figure 8: Continuous panel line

634 / Greco, Mariani

 The joint use of this special type of new head and of a Vittorio M ariani
specific hydraulic circuit allowing the recycle phase of
the components makes it possible to obtain an optimum Vittorio Mariani, doctor of
mixture without waste material even at the beginning and industrial chemistry, is one of
end of the production cycle (OMS PATENT), as usually the founders of OMS Group
happens to other continuous line. and has been since then
The second project covers a system for the continuous involved as R&D Manager.
production of cold store panels, with a special unit fully He started his activity in the
automated to insert the ‘cam-lock’ between the two metal polyurethane field as
sheets in order to grant a perfect alignment of cam-locks continuous slabstock foams
placed on the two sides of die panel (OMS PATENT). production manager.
By this it is possible to produce cold store panels in
continuous process at high capacity.
The plants manufactured by OMS Group are the result
of experience, research and project management in tight
contact/co-operation with the end-user because that is the
formula fo r obtaining the best results.

BIOGRAPHIES

Eraldo Greco

Eraldo Greco, bom in Seregno

Italy in 1965, has an education
background in Technical
Industrial Engineering. He
joined OMS Group Technical
Assistance Dept in 1987 and is
presently working in the
Commercial Dept as Sales
Manager for North America
and Europe markets.

Greco, Mariani / 635

Solkane® 365/227 Blown Rigid Foams: Optimized Insulation
and Cost Performance

LOTHAR ZIPFEL a n d KARSTEN BOERNER PIERRE DOURNEL

SOLVAY FLUOR und DERIVATE GmbH SOLVAY Research and Technology
Hans-Boeckler-Allee 20 310 rue de Ransbeek
D-30173 Hannover B-1120 Brussels
Germany Belgium

ABSTRACT

HFC -365mfc is a key for promising blends. A new More blends of Solkane® 365 are flammable azeotropes
blend Solkane® 365/227 meets best the main requirements with hydrocarbons, having a boiling points between 23°C
as a substitute for HCFC-141b: Solkane® 365/227 is a (73°F) and 32°C (90°F)
non-flammable liquid, starts boiling around 30°C and has The objective of the technical evaluation of the several
all the advantageous properties of HFC -365mfc. Different blends was to identify rules how to optimize foam quality
to HCFC-141b it is a binary blend and must be handled and cost.
adequately to maintain the non-flammable character.
Several hundreds of tons of Solkane® 365mfc have been RESEARCH APPROACH AND KEY POINTS
produced since January 2000 in a pilot plant. This enabled
the foam industry to start the optimization of their SOLVAY has concentrated on HFC-365mfc as a key
systems. component for blends. Table 1 shows prominent blends.
A limited amount of blowing agent can give the best The blowing agent loading has been checked for the
foam insulation performance at low temperatures and can best lambda and cost performance by following the vapor
save costs. Rules assist the producers to develop quality pressure curve (figure 1) at a given temperature. The
and cost effective formulations when replacing HCFC- correlated amount of blowing agent in foams has been
141b with HFC-365mfc-blends. calculated and then checked in practice.

PURPOSE Table 1 Solkane® 365mfc and its blends

Options Solkane® 365mfc Boiling Point
The phase out of HCFC -141b calls for non flammable content [weight- %]
liquids as a substitute. The deadline for the HCFC-141b Non flammable
phase out for foams is 1.1. 2003/2004 in industrialized Liquid
countries. Solkane®365/227 9 3 -9 4 30°C/86°F
Solkane® 365 forms promising blends, and Solkane®
365/227 has been identified as a substitute for HCFC- Flammable
141b. The properties are: Azeotropes
• no flash point Solkane®365mfc /
• a low thermal conductivity n-pentane 58 27°C/81°F
• boiling starts at 30°C / 86°F Solkane®365mfc /
• non toxic, environmentally acceptable i-pentane 46 23°C/73°F
• binary blend, ~ 6 pbw. Solkane® 227ea Solkane@365mfc /
(Heptafluoropropane) c-pentane 73 32°C/90°F

POLYURETHANES EXPO 200l f SEPTEMBER 30 -OCTOBER 3, 2001 637

Optimization W ork: How to apply best in Practice

The cell gas composition for the best lambda value calculated. Figures 2, 3 & 4 show the data and how to
depends on the mean temperature where the lambda value apply best in practice for a typical PU and a typical PIR
is checked. According to the vapor pressure curves of formulation.
figure 1, the optimized cell gas composition has been

Figure 1 Vapor pressure curves of Solkane® 365 and blends

Typical PUR Formulation

O
Temperature

oo
CM
CO
o
o
wt % physical 84% 97%
blowing agent
S 365/227

wt % chemical 16% 3%
blowing agent
H20

Ratio 5.2 32.3

physical/chemical
in system

Temperature 10 °C
o
o
CM
CO

Polyol 100 100

365/227 15 29

H20 2.9 0.9

ISO index 110

Figure 2 Optimized cell gas composition for 365/227 and typical PU foam

638 / Zipfel, Boerner, Dournel

 Figure 3 Optimum cell gas composition for different blends

Figure 4 Optimized cell gas composition for typical PIR foam

Zipfel, Boemer, Dournel / 639

Rules how to optimize blowing agent loading for best Insulation perform ance in comparison
insulation and cost perform ance
The insulation performance of foams has been checked
It has been found that, in most of the cases, the optimum for laminates, blown with different new blowing agents.
amount of blowing agent is 10 30% higher than found Table 3 compares the typical initial lambda values of
by calculation. The following table gives rules for foams, blown with different blowing agents. The
optimized blowing agent loading for best cost and lambda evaluation of the lambda value confirms, that
performance. • the insulation performance of foams blown with
Solkane®365/227 is close to HCFC-141b
Table 2 Blowing agent loading • the retention of insulation value is better than for
HCFC-141b
For 10°C/50°F mean temperature limitation
Solkane®365mfc/227 2.0- 2,3 kg/m3 of foam Important application types for Solkane® 365/227 are:
Solkane®365mfc/i-pentane 3.0- 3,5 kg/m3 of foam • high performance rigid insulation foams
Solkane®365mfc/n-pentane 2,6-3,1 kg/m3 of foam • fully formulated systems
Solkane®365mfc/c-pentane 2,2-2,6 kg/m3 of foam • spray foams, Pour-In Place foams
• SME (small manufacturer equipment)
For 23°C/73°F mean temperature no real limit • Integral skin foams
Solkane®365mfc/227 best: 3,7-4,2 kg/m3 of foam
min.: 2,3 kg/m3 of foam

• A good foam insulation performance is given at a low

blowing agent loading !

Table 3 Insulation performance of laminates blown with different blowing agents

HCFC-141b Solkane S 365mfc/ S 365mfc/ S 365mfc/
365/227 c-pentane i-pentane n-pentane
Density
[kg/m3] 32 33 34 30

Therm. Cond. @ 18.6 19.5 19.5 19.8 21.2

10°C/50°F[mW/m.K]

Environm ental Aspects

A sound overall environmental profile is a prerequisite products. Results of the study are summarized in figure 4.
for acceptance and market success. Solkane®365mfc has Looking at the full picture, the compound Solkane®
an excellent insulation performance but a substance- 365mfc has the potential to replace HCFC-141b and to
bound GWP (Global Warming Potential). This is why a become the future environmentally sound high
Life Cycle Assessment (LCA) according to ISO 14040 performance blowing agent for high performance rigid
series has been performed (1) to check the environmental PU-foams.
profile of Solkane® 365mfc-blown rigid foam insulation

640 / Zipfel, Boerner, Dournel

 Figure 5 Results of a life cycle assessment

SUMMARY Key applications for Solkane®365/227 are for quality

and safety reasons:
Most developments with Solkane® 365mfc are in blends • Fully formulated systems, spray / PIP, SM E ,...
• Solkane®365/227 is a liquid with a suitable boiling • Integral skin foams
point and a viable substitute for HCFC-141b
• Azeotropes of Solkane® 365mfc and Hydrocarbons CONCLUSION
can improve hydrocarbon blown foams
• An LCA on Solkane®365mfc as a prerequisite for A cost effective replacement of HCFC-141b is still a
acceptance and market success confirms a sound high priority. Solkane®365/227 is a viable candidate to
environmental profile. substitute HCFC-141b. Quality and foaming costs can be
• Solkane®365mfs exhibits a good insulation quality optimized by using a limited blowing agent loading.
and price performance at a limited blowing agent Solkane®365/227 and Solkane® 365 are available from a
loading pilot plant, and the commercial production is scheduled
for the second half of 2002.
• The insulation performance of foams blown with
Solkane®365/227 is close to HCFC-141b

• The retention of insulation value is better than for

HCFC-141b

Zipfel, Boemer, Doumel / 641

REFERENCES

1. H. Kraehling et al., LCA on rigid polyurethane foams:

Environmental competitiveness of 365mfc blown
products, Polyurethanes Conference 2000, Proceedings p.
23

BIOGRAPHIES

Lothar Zipfel
Lothar Zipfel received a PhD in Chemistry from the
University of Hanover, Germany. He joined the
Fluorinated Products Group of SOLVAY in 1982. He is
Manager of the technical service for plastic foams of
SOLVAY FLUOR und DERIVATE GmbH. His
responsibilities include technical support for SOLKANE®
blowing agents and for the brominated polyol IXOL®.

Karsten Boemer
Karsten Boemer joined SOLVAY in 1980 and started in
the Fluorine Division in 1983. He entered the technical
service group for foam blowing agents in 1997 and is
working in the technical service laboratories.

Pierre Doumel
Pierre Doumel received a PhD in Chemistry from the
University of Bordeaux, France. He joined SOLVAY in
1991 and is now responsible for technical customer
service in the area of foam blowing agents, as well as for
the brominated polyol DCOL®.

642 / Zipfel, Boerner, Doumel

The Effective Diffusivity of C 0 2 and HCFC-141 b in PU Foam
Prepared with HCFC-141b as Insulating Blowing Agent
at 20, 25 and 40 °C

J . RICHARD BOOTH KENNETH E. WILKES

Department o f Chemical Engineering Oak Ridge National Laboratory
Tennessee Technological University P.O .Box 2008
Campus Box 5013 Oak Ridge, Tennessee 37831-6092
Cookeville, Tennessee 38505

ABSTRACT

The temperature dependent effective scaled-age uses the short-term approximation of the
diffusivities of gas components diffusing from one-dimensional diffusion equation to estimate the
experimental insulating panels are required to effective diffusivity. Using die effective diffusivity
correctly model the transient composition change of values determined at 10, 25 and 40°C, the zero
the diffusing components with temperature. The intercept of the Arrhenius plot provides the frequency
polyurethane reaction by-product C 0 2 and insulating factor while the slope provides a measure of the
blowing agent HCFC-141b have a significantly molar activation energy.
different influence on the aging of polyurethane foam
insulation because of the specific magnitudes of their BIOGRAPHIES
effective diffusivity values. Foam panels prepared
with HCFC -141b as insulating blowing agent were J. R ichard Booth
initially bank -sawed and then sliced using a precision
sheer. The precision sliced foam specimens Dr. Richard Booth received his Ph.D. in Chemical
accumulated oxygen and nitrogen from the Engineering from Clemson University in 1965. He
atmosphere while depleting the C 0 2 and HCFC-141b received his B.S. from Virginia Polytechnic and State
from the preparation of the foam panels. The University and his M.S. from Clemson, both in
gravimetric method for measuring the effective Chemical Engineering. He worked for the Dow
diffusivity o f gases and vapors in closed-cell plastic Chemical Company until 1994 and is an Adjunct
foams is well-developed and has been used to Associate Professor at Tennessee Technological
measure other rapidly and slowly diffusing blowing University. He has performed research on gas
agents in polyurethane and other polymeric foam diffusion and thermal properties of foam for 20 years.
systems. To prepare the slices for measurement of
the effective diffiisivity, a single slice of air-filled Kenneth E. Wilkes
C 0 2 and HCFC -141b depleted foam was placed into
a saturation (pressure) vessel and sufficient material, Ken Wilkes is a senior research staff member at the
gaseous C 0 2 or liquid HCFC -141b was added to Oak Ridge National Laboratory. He has performed
produce a 6 psig saturation pressure within the foam research on thermal insulation for over 20 years, first
at room temperature. After 14 days, the saturation at Owens-Coming and, since 1987, at the Oak Ridge
material is vented, the saturation vessel is opened and National Laboratory. He holds a B.S. in Mechanical
the slice moved to the pan of an electronic analytical Engineering from Duke University an M.S. in
balance. A computer data-acquisition and control Mechanical Engineering from Purdue University and
system monitors and controls the atmospheric a Ph.D. in Condensed Matter Physics from the Ohio
conditions inside the insulated environmental State University. He is a registered professional
chamber. The computer records the weight of the engineer and is active in ASTM Committee C 16 on
foam slice at specific time intervals during the Thermal Insulation.
depletion o f the single diffusing component. The
slope of the dimensionless weight ratio with the
POLYURETHANES EXPO 2001 , SEPTEMBER 30 -OCTOBER 3 , 2001 643
Unique, Environmentally Friendly Release Agents for
MDI Foam Articles

BRIDGET SMITH, SCOTT LYTLE a n d LAUREN REAMS

Franklynn Industries, Inc.
167 Commerce Blvd.
Cincinnati, Ohio 45140

ABSTRACT to be reduced by 26% when using our tool conditioner

-
in combination with our water based mold release
Franklynn Industries introduces a new water-based agent. Using the tool conditioner also drastically
mold release agent for MDI molded foam articles as reduced polyurea build up- on the tool, a characteristic
well as a water-based tool conditioner to work in that occurs when isocyanate reacts w ith residual water
conjunction with our mold release agent. Franklynn left over by the mold release. The molded foam
Industries, a privately held company established in surface also showed no signs o f cell collapse or
1991, is a leader in developing customized increase in cell size.
formulations in water -based mold release agent Franklynn Industries continues to develop new
technologies for industries involving rubber, technologies based on MDI foam chemistry and is also
composite and urethane molding. Our company’s working towards expanding their capabilities of
W orld Headquarters is located in Cincinnati, Ohio producing a water based- mold release agent for
with various sales offices and facilities located Toluene Diisocyanate (TDI) polyurethane chemistry.
strategically throughout the world, including Canada We are committed to providing value added products
and Mexico. We select only those raw m aterials that to our customers and will work with our customers to
meet the global community’s increasing concern for develop custom form ulation to meet their needs.
environm entally responsible products and systems.
Franklynn Industries is QS -9000/ISO 9001 certified BIO G RA PH IES
The polyurethane mold release industry has endured
many significant changes since the early eighties. B ridget Sm ith
Initially, release agents consisted of naphtha and other
flammable and combustible solvents but due to the Bridget Smith is currently the Technical Service
increasing awareness o f environm ental and safety Representative for Franklynn Industries Incorporated.
concerns these solvent -based products continue to be Awarded Bachelors o f Science degree in Biological
elim inated Today’s m arket consists o f both water and Sciences from Ohio University in 1998 and has four
solvent -based release agents and hybrids o f water and years of industrial experience in quality assurance and
solvent. technical service.
Franklynn Industries has developed a 100% water -
based release agent for Diphenylmethane Diisocyanate Scott Lytle
(MDI) molded foam parts under our Aqualift product
line. Notable benefits were zero voc’s, less buildup of Scott Lytle is currently the Special Project / Technical
polyurea and release agent on the tool surface and Service M anager for Franklynn Industries
excellent part removal. Testing was completed on Incorporated. Awarded Bachelors o f Arts degree in
epoxy and aluminum tools in the tem perature range of Chemistry from Cedarville College and a M asters of
-
105 115°F. The polyurethane system was based upon Science from M iami University in 1995 and has nine
MDI chemistry at 3.5 5.5 - p sf densities. In addition, a years experience in the m old release industry.
-
new water based tool conditioner has been developed
for use in combination with our current water based - L auren Ream s
topcoat system. -
This water based conditioner is
applied to cracked, patched or other damaged areas on Lauren Reams is currently the Business Development
the tool by brushing or wiping. Testing with this Specialist for Franklynn Industries Incorporated.
conditioner as a pre-coat to our water-based mold Awarded Bachelors of Arts degree in International
release agent has noticeably reduced sticking in Studies from University o f Richmond and has six
damaged areas of the tool, therefore reducing scrap years of experience in the area of marketing.
rates. Results from field tests have shown scrap rates

POLYURETHANES EXPO 2001, SEPTEMBER 30-OCTOBER 3, 2001 645

Author Index

Aguirre, F., 17 DeLeon, A., 125 Idacavage, M., 347

Aneja, A., 195 DePalma, P., 177 Ingold, K. A., 461
Apichatachutapan, W., 563 Dillon, D., 339, 617 Inohara, H., 89
Arnold, K., 171 Divjakovic, V., 557 Ionescu, M., 607
Dixon, C., 71, 599 Ishikawa, A., 33
Baba, F., 615 Doumel, R, 325, 637 Isobe, M., 267
Bailey, R. J., 285 Durairaj, R. B., 535 Ivanyi, S. I., 3
Bambrick, C. 171
Bartenstein, M. F., 381, 539 Egawa, S., 255 Jacobs, P., 537
Basile, G., 611 English, S., 133 Jarfelt, U., 609
Behrendt, G., 3 Eyrisch, O., 417 Javni, L, 557
Berthevas, R, 17 Jayakody, C., 409
Berube, A., 201 Fanget, A., 47, 551 Johnson, B., 537
Bharadwaj-Somaskandan, S., 239 Farkas, R, 143, 579 Johnson, M., 533
Bladon, J., 133 Flammang, B., 577 Jones, S. A., 61
Blair, G. R., 285 Froling, M., 609 Kageoka, M., 403
Blaszkiewicz, M., 247 Fujimoto, T., 615 Karajkov, V., 557
Boemer, K., 637 Fukami, T., 89, 255 Karas, G. S., 183
Bogdan, M., 115 Furuta, T., 443 Karoly, B., 75
Boinowitz, T., 41 King, J., 473
Boncza-Tomaszewski, Z., 565 Gabbard, W. A., 313 Kiso, H., 491
Bontinck, D., 347 Gay, K. M., 373, 491, 547 Kiuchi, K., 33
Booth, J. R., 313, 643 Gebreselassie, G., 571 Klietsch, J., 41
Borgogelli, R., 417 Geiger, B., 513 Klincke, M., 501
Bowman, J. M., 151, 483 Gerard, E.-J., 423 Kniss, J. G., 431
Bradt, D. E., 285 Ghosh, S., 211 Kometani, H., 373
Bridge, J., 409 Gilbert, M., 619 Kono, S., 443
Brown, L. J., 219 Godoy, J., 133 Koshute, M. A., 247
Brown, M., 409 Goering, H., 3 Krahling, H.,333
Brzezinski, A., 81 Goldstein, S. L., 363 Krishnamurthi, B., 239
Bunting, W., 41 Goyvaerts, H., 537 Kuczynski, E., 579
Bures, K., 409 Greco, E., 631 Kumaki, T., 403
Burkhart, G., 41, 417, 501 Grimminger, J., 431 Kurita, M., 89
Gross, D. J., 589 Kurtycz, E., 133
Cageao, R., 177 Gross, S., 389
Canney, B., 71, 599 Guamieri, W., 177 Langenstrassen, R., 3
Casati, F. M., 47, 551 Larre, A., 133
Cenens, J., 577 Hanne, D. R., 513 Latham, I., 219, 473
Chandalia, K. B., 363 Hayashi, M., 255 Lawrey, B. D., 191
Chao, H. S., 587 Herrington, R., 47 Laycock, D. E., 227
Ciborowski, S., 389, 521 Hofmann, B. H. W., 589 Lee, B., 177
Cionni, J., 629 Hogg, A., 133 LeFever, A., 275
Crocker, M., 409 Horiguchi, H., 443 Little, K., 551
Crooker, R., 107 Huang, C. C. T., 519 Liu, S. C. P., 519
Huhtasaari, M. S., 431 Lockwood, R. J., 451
D’Errico, D., 177 Hurley, M. F., 161 Lorenzetti, A., 611, 623
de la Cruz, M., 537 Huspek, M. R., 97 Lupini, M., 183
DeAlmeida, J., 535 Huth, H., 3 Lytle, S., 645

Author Index / 647

Maas, W., 423 Pellacani, L., 353 Thiede, V. M. T., 353
Mangs, S., 609 Penczek, P., 565 Tokumoto, K., 491
Mariani, V., 631 Petrovic, Z. S., 557, 595 Tomosada, T., 443
Martin, C., 473 Pieper, K., 513 Tu, J., 17, 47
Mathis, N., 71, 599 Plaumann, R., 431
Mautino, V. M., 461, 629 Pocol-Savage, M. A., 211 Udeshi, S. V., 565
May, D., 389, 521 Pohl, M., 3 Urata, S., 619
McAdon, M. H., 227 Preoteasa, V., 607 Usaka, K., 267
McClusky, J., 211 Utsumi, H., 267
McEvoy, J. T., 275, 281 Ramazzotti, D., 593
Mendoza, L., 143 Ramnas, O., 609 Van Eetvelde, E., 577
Messina, A. J., 571 Reams, L., 645 Van Maris, R., 373, 491, 547
Meyer-Ahrens, S., 183 Righi, C., 97 Vandichel, J. C., 577
Mihai, S., 607 Roibu, C., 607 Vanmeulder, G., 347
Milliren, C. M., 183 Rossio, R., 569 Verstraete, H., 423
Mispreuve, H., 47, 551 Russ, A., 285 Villwock, R., 11
Modesti, M., 611, 623
Morgan, M. J., 363 Saiki, K., 255 Wei, X., 211
Morii, M., 33 Sakai, M., 33 Wheeler, I., 107
Murai, M., 615 Sanou, M., 615 Wickwire, D., 107
Murakami, S., 255 Sato, H., 619 Wieczorek, D., 565
Myers, D., 409 Sato, Y., 403 Wilkes, G. L., 191, 195
Schlenter, B., 423 Wilkes, K. E., 313, 643
Nanno, H., 89 Schmidt, K.-H., 3 Williams, D. J., 115, 151,483
Nakamura, H., 547 Schroder, T., 577 Willison, J., 513
Narayan, S., 201 Scott, K. W., 61 Willoughby, B. G., 61
Naruse, A., 89 Sekiya, A., 619 Willoughby, K., 501
Neal, B. L., 247 Sheard, E. A., 61 Wojcik, R. T., 363
Neff., R., 563 Shutov, F., 239 Womack, F. D., 431
Nelson, G. E., 313 Simioni, F., 623 Wu, J., 339, 617
Skowronski, M. J., 97
O’Connor, J. M., 227 Smith, B., 645 Xue-Wei, X., 573
O’Sickey, M. J., 191 Sonney, J. M., 47
Obi, B. E., 219 Stanciu, R., 143, 579 Yamasaki, R., 281
Ohkubo, K., 267 Stepan, E., 607 Yarbrough, D. W., 313
Ohta, H., 443 Stovall, T., 81 Yoshimura, H., 491
Olsson, M. E., 609 Su, C., 573 Young, G. R, 389, 521
Ootsuka, H., 403 Swedo, R. J., 551 Yourd, R. A., 461
Orizondo, R, 537 Symosko, G., 177
Ostrysz, R., 565 Zhang, W., 557
Oswald, L., 389, 521 Takeyasu, H., 619 Zhong, B., 151
Tamai, R., 619 Zipfel, L., 325, 333, 637
Parekh, P. P., 75 Tamano, Y., 373, 491, 547 Zlatanic, A., 595

648 / Author Index

Company Index

Air Products and Chemicals, Inc., 431 Great Lakes Chemical Corporation, 537 Oak Ridge National Laboratory, 81, 313
Air Products GmbH, 431 Great Lakes Technology Belgium, 537 Orbseal LLC, Inc., 183
ANGUS Chemical Company, 551 Gusmer-AdmiraL Inc., 381, 539
Asahi Glass Co., Ltd., 617 Padova University, 609, 619
ATOFINA Chemicals, Inc., 107, 339, Hennecke GmbH, 513 Pittsburg State University, 557, 595
615 Hennecke Machinery, 513 Pou Chien Chemical Co., Ltd., 519
Honeywell, 115, 151, 483
BASF Corporation, 563 Huntsman Polyurethanes, 75, 451 Rapra Technology Ltd., 61
Bayer AG Leverkusen, 177, 623 Rogers Corporation, 201
Bayer Corporation, 161, 177, 183, 191, Impianti OMS Spa, 625
247, 461 INDSPEC Chemical Corporation, 535 Sanyo Chemical Industries, Ltd., 443
Bundesanstalt fur Materialforschung Industrial Chemistry Research Sartomer Co., Inc., 587
und -priifung, 3 Institute, 565 S. C. Oltchim, S.A., 605
Institut fur Kreislaufwirtschaft Shell Chemicals, 423, 577
Chalmers University of Technology, der Technischen Fachhochschule SOLVAY FLUOR und DERIVATE
607 Wildau, 3 GmbH, 325, 333, 631
Chem-Trend, Inc., 533, 569 Institute of Chemical Research, 605 SOLVAY Management Support GmbH,
Chestnut Ridge Foam, Inc., 409 333
Collins and Aikman Corporation, 571 Jayant Oil Mills, 565 SOLVAY Research and Technology,
Johnson Controls, Inc., 275, 281 325,631
SynUthane International Inc., 227
Dow Chemical Canada Inc., 133, 171,
Kaneka Texas Corporation, 211
227
Kao Corporation, 33 Tennessee Technological University,
The Dow Chemical Company, 17, 47,
97, 133, 171, 219, 227, 473 239,313, 637
LaserComp, 81 TOSOH Corporation, 373, 491, 547
Dow Deutschland GmbH and Co.,
Lescon, Inc., 61 TOSOH EUROPE B.V., 373, 491, 547
OHG, 353
Lyondell Chemical Company, 363 TOSOH USA, Inc., 373, 491, 547
Dow Europe S.A., 17, 47, 133, 551
Dow Italia S.p.A., 97
Mathis Instruments Ltd., 71, 599 UCB Chemicals, 347
Meridian Automotive Systems, 177 UCB Chemicals, Inc., 347
Edge Sweets Company, 593 Mitsubishi Electric Corporation, 613 University of Chemical Technology and
Electrotech Industrial Inc., 589 Mitsui Takeda Chemicals, Inc., 267, Metallurgy, 3
403 University of Novi Sad, 557
Fluorocarbons and Development Mobius Technologies, Inc., 11 University of Texas at San Antonio, 211
Service, 609 U.S. Army Soldier Chemical-Biological
Format Messtechnik GmbH, 589 Nanjing University of Chemical Command, 389, 521
Franklynn Industries, Inc., 639 Technology, 573
National Institute of Advanced Virginia Polytechnic Institute and State
Goldschmidt AG, 41, 417, 501 Industrial Science and Technology University, 191, 195
Goldschmidt Chemical Corp., 41, (AIST), 617
417, 501 Nippon Polyurethane Industry Woodbridge Foam Corporation, 143,
Graphtech, Inc., 619 Co., Ltd., 89, 255 285, 579

Company Index / 649